Comparison of DNA analysis, targeted metabolite profiling, and non-targeted NMR fingerprinting for differentiating cultivars of processed olives

Abstract

Genetic and metabolite analyses have been frequently employed in food traceability studies, although their effectiveness in heavily-processed products such as canned olives is uncertain. A robust method for identifying cultivars of processed table olives is necessary to ensure product authenticity and economic fairness. This study compared microsatellite analysis, fatty acid profiling, phenolic profiling, and NMR fingerprinting for differentiating cultivars of California-style olives. Methods were tested on domestic Manzanilla, domestic Sevillano, imported Hojiblanca, and imported Gordal olives. Microsatellites were successfully amplified in all cultivars except Hojiblanca, and Sevillano/Gordal were found to be genetically-related, possibly synonyms for the same genetic cultivar. Coupled with multivariate statistics, fatty acid profiling differentiated the three genetically-distinct cultivars, and NMR fingerprinting differentiated all four cultivars. Phenolic profile showed too much variability for effective discrimination. Future work includes expanding current models by size, cultivars, and type of table olive.

Introduction

Table olives are becoming an increasingly globalized commodity. Worldwide production of table olives has grown steadily over the past few decades, reaching over 2.5 million tons in the 2018/2019 season (International Olive Council Newsletter, 2019). While Spain, Greece, and the United States were traditionally responsible for the majority of table olive production, new countries in the Mediterranean Basin including Egypt, Algeria, and Turkey are quickly becoming important producers as well.

To meet growing consumer demand, table olive producers often import processed olives from other regions to pack or label. There are suspected cases of cultivar mislabeling or fraud in the supply chain, which undermines traceability and economic fairness. Selected cultivars are preferable for table olive processing based on physical properties of the fruit such as size, shape, oil content, and susceptibility to bruising (Jiménez-Jiménez et al., 2013; Pinheiro & Esteves da Silva, 2005), as well as the composition of flavor- and health-contributing compounds like phenolics, volatiles, fatty acids, and sugars (Charoenprasert & Mitchell, 2012; Kalua et al., 2007; Marsilio, Campestre, Lanza, & De Angelis, 2001). Producers currently do not have a way to determine whether they are actually receiving the cultivars they purchased. Additionally, many table olive cultivars across Greece, Italy, Spain, France, and Portugal hold a protected designation of origin (PDO) by the EU, and it is important to be able to prevent adulteration of these products (Concepcion, García, Medina, & Brenes, 2019).

One approach to cultivar authentication is through analysis of genetic markers. Microsatellites, or simple sequence repeats (SSRs), have been reliably used to differentiate cultivars of olive trees, olive oil, and fresh and brine-stored olive fruits (Bracci & Sebastiani, 2011; Doveri, O'Sullivan, & Lee, 2006; Pasqualone et al., 2013, 2016). In our previous work, we demonstrated for the first time that highly-processed California-style olives could be genotyped using a select panel of microsatellites (Crawford, Carrasquilla-Garcia, Cook, & Wang, 2020). The study had limited sampling, and the effectiveness of the developed method for olives sourced from different processing lots and production facilities has not been explored.

Specific classes of metabolites have also shown potential for discriminating olive cultivars. Previous studies were able to differentiate varieties of olive oil and fresh olives by analyzing fatty acid profile (Casale et al., 2010; Mannina et al., 2003; Matos et al., 2007) or phenolic profile (Bajoub et al., 2017; Gómez-Rico, Fregapane, & Salvador, 2008; Kalua, Allen, Bedgood, Bishop, & Prenzler, 2005) data with multivariate statistics. Yet very few works have focused on processed olives. Lopez et al. (2006) found that chemometric analysis of fatty acid data could effectively discriminate eight cultivars of brined, Spanish-style, and California-style olives (López, Montaño, García, & Garrido, 2006), and Malheiro, Sousa, Casal, Bento, and Pereira (2011) achieved good discrimination of five Portuguese cultivars of brine-stored olives using principle component analysis (Malheiro et al., 2011). However, many metabolites, particularly phenolic compounds, can be significantly affected by processing conditions such as sodium hydroxide treatments, water rinses, oxidation, and heating (Charoenprasert & Mitchell, 2012). The ability of targeted metabolite profiling to reliably differentiate between cultivars of highly-processed olives is not well understood.

The most recent approach for cultivar discrimination is non-targeted analysis, in which chromatographs or spectra with or without assignment of individual compounds are used as a type of “fingerprint” for each sample. Methods such as near infrared spectroscopy (NIR), mid infrared spectroscopy (MIR), and nuclear magnetic resonance (NMR) have been used for cultivar fingerprinting of olive oil and fresh olives (Casale et al., 2010; Dupuy et al., 2010; Mannina et al., 2003; Piccinonna et al., 2016). Because the spectrum is analyzed as a whole, there is potentially greater ability to capture variability between cultivars than by analyzing a single class of compounds. A previous study found that NIR achieved better discrimination between cultivars of fresh olives than gas chromatographic analysis focused only on the fatty acid profile (Casale et al., 2010). The use of a non-targeted method for differentiating cultivars of processed olives has not been explored in the literature. Similar to phenolics, it is not known whether the spectrum would show sufficient differences between cultivars, or be robust to variability caused by processing.

In the current work, the effectiveness of microsatellite analysis, fatty acid profiling, phenolic profiling, and untargeted ¹H NMR fingerprinting for discriminating cultivars of processed olives was compared. California-style olives were selected for this study because they are subjected to the highest amount of post-harvest treatment including brine-storage, repeated lye and water rinses (during which air is bubbled throughout the solution to promote oxidation), and sterilization. Olives of this style are expected to undergo the most significant chemical changes compared to fresh olives. California producers pack four major cultivars of olives, which were sourced for this study: domestically-grown Manzanilla, Hojiblanca imported from Spain, domestically-grown Sevillano, and Gordal imported from Spain. Individual olives were used as replicates under the assumption that each olive originated from a different tree. However, for each cultivar, cans from different production lots or facilities were sampled in order to study the variability introduced by processing.