Selection for peanut varieties with low aflatoxin risk

• Acharya N.G. Ranga Agricultural University, India
• Queensland Department of Primary Industries and Fisheries, Australia

Aflatoxins are a grouping of carcinogenic compounds produced in peanut seed, particularly under end-of-season drought conditions coinciding with infection by soil fungi Aspergillus flavus and A. parasiticus. This has been recognised as a major food safety issue worldwide, and is particularly severe in developing countries in Africa and South and Southeast Asia.
Although management practices such as irrigation, precleaning, inverted windrows and controlled drying are effective in controlling aflatoxin contamination in peanut, the implementation of these practices is not considered to provide economically viable and sufficiently related solutions to resolve the aflatoxin problem faced by resource-poor farmers. Hence, genetic enhancement for low aflatoxin risk in peanut has been identified as a high-priority area in ICRISAT's research agenda in India and other peanut-growing countries in the semi-arid tropics.
There has been a huge worldwide research effort to identify aflatoxin tolerance/resistance (including work at ICRISAT during the mid 1980s to the 1990s), but little progress has been made in developing screening tools or identifying sources of aflatoxin resistance, owing to the high genotype x environment interaction for this complex problem.
In Australia, although aflatoxin has been a problem for the peanut industry for nearly 20 years, but a newly introduced standard threatens viability in the rainfed peanut production systems of Queensland. Therefore the urgent need to find solutions to the aflatoxin problem provides a common objective to Australian, ICRISAT and Indian peanut improvement programs.

This project sought to identify peanut germplasm with traits that contribute to low aflatoxin risk (LAR) through an improved understanding of the underlying mechanisms associated with the soil-plant-fungus interaction, which lead to invasion by Aspergillus flavus and subsequent production of aflatoxin in the peanut pods. The ultimate aim was to develop novel selection tools that could be combined into a 'selection index' that would help identify LAR peanut genotypes.

There have only been limited attempts to explore/exploit peanut germplasm for aflatoxin resistance traits. Due to the known complexity of the aflatoxin problem in peanut, it is unlikely there is a single genotype possessing complete resistance to seed colonisation or aflatoxin production. Scientists in India and Australia screened a range of germplasm, as well as genotypes from drought-resistant populations that were developed in two previous ACIAR peanut projects.
The research built on the current knowledge base pertaining to the aflatoxin problem. In India, in addition to expertise and knowledge on peanut (wild and cultivated germplasm), ICRISAT provided its excellent facilities to conduct pathological and physiological research on seed infection and aflatoxin production. Portable rainout shelters (developed in the earlier research), growth cabinets and recently developed A. flavus sick plots at ICRISAT were used to impose the desired aflatoxin risk environments, in order to effectively study genotypic variation for LAR.
Research at ICRISAT focused on identifying mechanisms (at crop, pod, seed testa and cotyledon level) contributing to seed infection by A. flavus and aflatoxin production, and screening selected peanut germplasm for LAR traits. Genotype x environment (G x E) interaction for the LAR traits was investigated in the field and in growth chambers at ICRISAT, as well as at QDPI in Kingaroy, Queensland. QDPI also assisted ICRISAT scientists in analysis of G x E results and in modelling aflatoxin risk.

This project set itself a very ambitious set of goals in addressing this subject, one that proved much more complex and difficult to research than previously realised. The three-year time frame was enough to develop some of the knowledge needed but insufficient to arrive at the desired solutions.
Scientists found that there were significant differences in biochemical composition among genotypes, including differing aflatoxin levels in kernels depending on their stage of maturity. Some of the biochemical composition characteristics showed significant variation between genotypes. Under high aflatoxin risk conditions aflatoxin levels were generally higher in immature pods (R4-R5 stages) compared to mature pods (R7-R8) stages. Amongst the biochemical changes, the most significant was an increase in sugars, more specifically, sucrose levels in R4-R5 stage kernels, which had a positive correlation with aflatoxin levels.
In India trials were conducted at both ANGRAU (Tirupati Centre) and ICRISAT Centre in conditions likely to maximise the risk of AF contamination in the pre-harvest context. In these trials a range of genotypes with reputations for different AF-contamination responses, and those lines thought to have contrasting attributes for traits relevant to the AF risk (AFR) were grown and subjected to the conditions most likely to provoke contamination. These initial trials showed that drought avoidance traits were related to AFR. The presence of phenolic compounds was related to low aflatoxin in some of the trials conducted. Major differences in the AF response of lines were observed over the sites used. A start was made to define the strain differences between locations (as measured by AF production on a standard medium).
In Australia the initial studies established an understanding of the basis for the observed differences between varieties NC7 (high AFR) and Streeton (low AFR). This contrast showed that differences in AF contamination were associated with water relations; pod maturity related differences in sucrose, the presence of methyl hydroxyl proline, and differential pod retention in the harvest, pod shell integrity factors and drying patterns.
Both Indian and Australian sites provided little consistent AF contamination across genotypes despite natural or induced high environmental risk index for AF. In the case of the ICRISAT site contamination by AF was consistently low and scientists are now working to increase the suitability of the site for this research.
Studies of lines with low AFR in diverse locations were undertaken to examine the consistency of the various parameters. These trials showed that in Australia consistent AFR performance could be obtained within a site, but that there was little consistency across sites, although one line was identified that always had low AF contamination in all the combinations of site and year tested. Studies to examine the cause of this G x E were undertaken and discounted the possibility that there were differences due to inadequate fungal challenge, or that differences were due to non-toxigenic fungi.
In India, similar G x E interactions for AF contamination were observed, with even those lines historically associated with different levels of AFR being unpredictable in their response in some test environments. Some lines with usually low AFR across all the test sites were identified and some of these can be exploited for production, since they had already been identified as having high productivity.
These results provided the project team with a considerable setback - large unexplained G x E makes the development of a selection index of little value. Therefore Objective 3, which would develop and validate selection index for LAR as a means of selecting winners from a diverse array of peanut germplasm, did not advance.