Chemometric Characterization of Almond Germplasm: Compositional Aspects Involved in Quality and Breeding

The oil content and the percentage of the main fatty acids were determined in a set of 73 almond (Prunus amygdalus Batsch) cultivars from 10 different countries present at the almond germplasm collection of the Centro de Investigación y Tecnologı́a Agroalimentaria de Aragón, Spain (CITA). Wide variability was observed for oil content, ranging from 51.5% to 66.8% on a dry weight (DW) basis. For the main fatty acids in the lipid fraction, the variability ranged from 62.9% to 77.3% for oleic acid, from 14.0% to 26.8% for linoleic acid, from 4.9% to 7.0% for palmitic acid, from 1.5% to 3.4% for stearic acid, and from 0.3% to 0.6% for palmitoleic acid. No correlations were found between the oil content and the percentages of the different fatty acids, but a significant negative correlation was found between the percentages of oleic and linoleic acids. Principal component (PC) analysis showed that palmitic, oleic, and linoleic acids and the oleic acid/linoleic acid ratio were primarily responsible for the separation on principal component 1. The content of each component was not related to the country of origin of the different cultivars, indicating that almond fatty acid composition is genotype-dependent. Cultivars with high and stable oil content and low linoleic acid should be selected as parents in a breeding program to increase kernel oil stability and nutritional value. Almond is the most important tree nut crop in terms of commercial production. This production is limited to areas characterized by a Mediterranean climate (Kester and Asay, 1975), including regions in the Mediterranean countries, the Central Valley of California, the Middle East, and some equivalent areas in the Southern Hemisphere. Traditional almond culture used open-pollinated seedlings (Grasselly, 1972; Rikhter, 1972), which, together with self-incompatibility, produced very high heterozygosity in this species (Kester et al., 1990; Socias i Company and Felipe, 1992). This large variability has provided a useful genetic pool for almond evolution, allowing in each growing region the selection of almond cultivars well-adapted to this area (Kester et al., 1990). However, some cultivars have shown high plasticity, being adapted to different growing conditions (Felipe, 2000). Several characteristics (fruit traits, blooming date, productivity, resistance to pests and diseases, etc.) were taken into account by human selection pressure when looking for the adequate genotypes adapted to local conditions and to the preference of the consumers. The genotypes having incorporated these highly selected traits represent very valuable germplasm for addressing future challenges in almond breeding. As a consequence, these genotypes are preserved, characterized in several almond collections and incorporated into advanced breeding programs (Kester et al., 1990). One of the most important collections was initially assembled by A.J. Felipe at CITA with 250 accessions introduced from all over the world (Espiau et al., 2002). This collection shows very large variability reflecting the wide genetic diversity of almond (Socias i Company and Felipe, 1992). Despite the large amount of information on these accessions, the chemical composition of their kernels has only been partially studied. This information would be crucial to increase the knowledge of their diversity, the nutritional and healthy value of the kernels, and the possibility of selecting the most adequate parents in a breeding program for increasing kernel quality (Socias i Company et al., 2008). The modern almond industry requires commercial cultivars characterized by kernels with high-quality attributes, because the best end use for each cultivar is a function of its chemical composition (Berger, 1969) and of the consumers’ trend for foods without synthetic additives (Krings and Berger, 2001). Thus, the CITA almond breeding program has incorporated the chemical quality criteria in the evaluation of the new cultivars (Socias i Company et al., 2009). The high nutritive value of almond kernels arises mainly from their high lipid content, which constitutes an important caloric source but does not contribute to cholesterol formation in humans. This is the result of their high level of unsaturated fatty acids, mainly monounsaturated fatty acids (MUFA), because MUFAs are inversely correlated with serum cholesterol Received for publication 26 Apr. 2011. Accepted for publication 8 June 2011. This work was supported by Spanish grants AGL2010-22197-C02-01 and INIA RF2008-00027-00-00, European grant AGRI GEN RES 870/2004 068 (SAFENUT), and Research Group A12 of Aragón. Technical assistance by J. Búbal and O. Frontera is highly appreciated. Corresponding author. E-mail: rsocias@aragon.es. J. AMER. SOC. HORT. SCI. 136(4):273–281. 2011. 273 levels (Sabate and Hook, 1996). Kernel tendency to rancidity during storage and transport is a quality loss and is related to oxidation of the kernel fatty acids (Senessi et al., 1996). Thus, oil stability and fatty acid composition, essentially the oleic acid/ linoleic acid (O/L) ratio (Kester et al., 1993), is considered an important criterion to evaluate kernel quality. The determination of kernel oil quality, taking into account both its content and composition, is an imperative step in the evaluation of new cultivars before being released (Socias i Company et al., 2008). Recent studies on the transmission and heritability of oil content, fatty acid composition, and the different tocopherol homologs (Font i Forcada et al., 2011) have shown that oil content presents high heritability values (57%) and linoleic acid a medium value (25%). When heritability estimates are moderate, selection of parents based on their phenotypes should also be effective, as it happens when heritability is high (Hansche et al., 1972). Kernels with a high percentage of oil could be used to produce nougat or to extract oil, which is used in the cosmetic and pharmaceutical industries (Socias i Company et al., 2008). In addition, high oil content is desirable because higher oil contents result in less water absorption by the almond paste (Alessandroni, 1980). Low content of linoleic acid is correlated with high oil stability (Zacheo et al., 2000). Thus, selection of parents for low linoleic acid and high oil content might be undertaken in a breeding program for increased kernel quality. Thus, our main objective was the study of the genetic diversity for oil content and fatty acid composition of the almond cultivars included in the CITA collection to ascertain the best genotypes from the point of view of oil quality. These choice genotypes should be included as parents in breeding programs aiming at improving almond kernel quality. Material and Methods PLANT MATERIAL. The list of the 73 almond cultivars studied is shown in Table 1. The trees are maintained as living plants grafted on the almond · peach [Prunus persica (L.) Batsch] hybrid clonal rootstock INRA GF-677 using standard management practices (Espiau et al., 2002). Nuts from open pollination were harvested in 2008 and 2009 at the mature stage, when fruit mesocarp was fully dried and split along the fruit suture and peduncle abscission was complete (Felipe, 1977). Two samples of 20 fruit were collected for each treatment. OIL AND FATTY ACID DETERMINATION. After blanching, the kernels were ground in an electrical grinder. Oil was extracted from 4 to 5 g of ground almond kernel in the commercial fat extractor Soxtec Avanti 2055 (Tecator, Barcelona, Spain) for 2 h using petroleum ether as a solvent and keeping the heating source at 135 C because previous checks showed that extraction is practically completed after 2 h with no differences after 4 h (Kodad and Socias i Company, 2008). The oil content was expressed as the difference in weight of the dried kernel sample before and after extraction. The oil sample was used to prepare methyl esters of the corresponding fatty acids (FAMEs) by transetherification with KOH according to the official method UNE-EN ISO 5509:2000 (International Organization for Standardization, 2000). These FAMEs were separated using a flame ionization detector gas chromatograph HP-6890 equipped with a HP-Innowax column of 30 m · 0.25 mm i.d. and 0.25-mm film thickness (Agilent Technologies, Waldbronn, Germany). The carrier gas was helium at a flow rate of 1 mL min. The temperature of the inlet and detector was maintained at 220 and 275 C, respectively. The initial column temperature was 100 C for 3 min. The oven temperature was increased from 100 to 150 C at 20 C min ramp rate for 1 min, from 150 to 200 C at 15 C min ramp rate for 3 min, and from 200 to 240 C at 3 C min ramp rate. The temperature was maintained at 240 C for 4 min. Injection volume was 1.0 mL. The identification of the FAMEs was achieved by comparing with relative chromatographic retention times in a reference sample that contained standard methyl esters (Sigma-Aldrich, Madrid, Spain). The results were expressed as percentages of each fatty acid in the total oil amount. STATISTICAL ANALYSIS. All statistical analyses were performed with SAS (Version 9.1; SAS Institute, Cary, NC). The analysis of variance with the PROC GLM procedure was applied to evaluate the genotype and year effect on the studied variable. The mean separation was done with the least significant difference test at a probability of 0.05. The LSMEANS option of the GLM procedure was used to calculate least-square means for genotypes in each year to observe rank changes of genotypes from year to year. Principal component analysis (PCA) was performed to study correlation among fruit quality measurements and to interpret relationships between genotypes as a tool for germplasm characterization (Gurrieri et al., 2001; Iezzoni and Pritts, 1991). PCA analyses were performed on the correlation matrix of the synthetic variable based on the mean population value (Prosperi et al., 2006) using the PRINCOMP proced