Comparison of the grain composition in resveratrol-enriched and glufosinate-tolerant rice (Oryza sativa) to conventional rice using univariate and multivariate analysis

Highlights

• We analyzed 55 components in resveratrol-enriched transgenic and non-transgenic rice.

• The effects of transgenesis or glufosinate on the content were insignificant.

• Growing location was the greatest contributor to compositional variability in grains.

Abstract

Resveratrol-enriched rice (RR) contains genes that express the recombinant Arachis hypogaea stilbene synthase (AhSTS1) and phosphinothricin-N-acetyltrasferase (PAT) for resveratrol production and glufosinate tolerance, respectively. To satisfy regulatory safety evaluations, herein, the content of 55 analytes in the RR (non-sprayed and sprayed with glufosinate) and conventional non-transgenic rice grown at three different sites in the Republic of Korea was determined. Data evaluated using univariate and multivariate analyses indicated compositional equivalence between the RR and conventional rice. Principal components analysis (PCA) revealed significantly higher differences among plants from different locations than between the transgenic versus non-transgenic plants. Pearson correlation and hierarchical clustering (HCA) analyses indicated significant correlations among the contents of most minerals. Furthermore, PCA could not distinguish among glufosinate-sprayed RR, non-sprayed RR, and its control, supporting the conclusion that glufosinate-ammonium treatment caused insignificant changes in the RR grain composition.

Introduction

Resveratrol (3,5,4′-trihydroxy-trans-stilbene) is a non-flavonoid polyphenol-type compound found in peanuts and grapes (Creasy and Coffee, 1988, Jeandet et al., 1991) shown to possess a number of human health benefits including anti-cancer, anti-hyperlipidemia, and anti-aging properties (Goldberg, 1996, Marchal et al., 2013, Min and Song, 2010, Osman et al., 2013). In addition, a previous study demonstrated the skin-whitening effect of resveratrol (Lee et al., 2014). To increase its dietary intake, many groups have recently attempted to create transgenic plants that accumulate resveratrol. For example, Baek et al. (2013) introduced Arachis hypogaea stilbene synthase gene, AhSTS1 (RS3, GenBank accession no. DQ124938), into a commercial rice variety, Dongjin (DJ), to develop a resveratrol transgenic rice line containing a single copy of the transgene. The resveratrol-enriched rice (RR) showed significant suppression of the expression of tyrosinase, the predominant enzyme in melanogenesis, in mouse melanocytes (Melan-a cells; Lee et al., 2014). In addition, the RR significantly suppressed the expression of melanogenic proteins such as tyrosinase, tyrosinase related protein-1 (TRP-1), and microphthalmia-associated transcription factor (MITF) in guinea pig skin. Histological studies suggested that melanin production also decreased in the epidermis. Although rice is known to possess skin-whitening effect, RR was even more effective than the untransformed rice. Furthermore, in recent studies, feeding of mice with the RR revealed excellent anti-obesity effects (Baek et al., 2014). Similarly, stilbene synthase genes have been transferred to a number of crops to improve their nutritional value with respect to their resveratrol content (Giovinazzo et al., 2012).

Extensive safety assessment is required before genetically modified (GM) crops are commercialized for use as food (Kok et al., 2008, Kuiper et al., 2001). Compositional equivalence is considered a cornerstone of the case-by-case safety assessment of GM foods (Herman et al., 2009). Compositional safety, assessed in terms of substantial equivalence, is investigated by comparing the analyte levels of the transgenic food with those of the non-transgenic counterpart crop that has a history of safe use. Levels of the analytes of concern from the transgenic crop are compared to those reported in other varieties of the same crop. A new transgenic variety is considered safe if its analyte levels are similar to other varieties of the same crop (Herman et al., 2010). No compositionally unsafe varieties have been identified in crops (e.g., maize, soybean, cotton, rice, etc.). However, to the best of our knowledge, such assessment of the resveratrol-enriched crops has not been reported. As a preliminary step towards the commercialization of RR, the present study was aimed at comparing the composition of RR with that of non-transgenic comparators.

In the case of herbicide tolerant transgenic plants, data from plants with and without the recommended herbicide applications should be included to allow for the effects of the herbicide on crop composition to be evaluated (Oberdoerfer et al., 2005). The substantial equivalence assessment to determine the compositional safety of RR, which expressed the phosphinothricin-N-acetyltransferase (PAT) enzymes, in addition to the stilbene synthase, was employed. Although, bar gene, encoding the PAT enzyme, was used as a selection marker, its incidental benefit of herbicide tolerance conferred on the RR plants mandated comparison of the compositional data from the RR plants grown with or without the application of glufosinate-ammonium herbicide.

In addition to confirming the equivalence of GM and non-GM crops, compositional studies have consistently demonstrated that other factors such as germplasm, environment, and management practices contribute more to the variability in the levels of key nutrients and anti-nutrients than transgenic breeding (Drury et al., 2008, Harrigan et al., 2010, Harrigan et al., 2009, Herman and Price, 2013, McCann et al., 2007, Ridley et al., 2002, Zhou et al., 2011). Statistical approaches that can provide meaningful evaluation of the relative impact of different factors to compositional variability may offer advantages over traditional frequentist testing. In this regard, multivariate approaches have been considered. For example, principal component analysis (PCA) provided summary graphics of the impact of multiple experimental factors on compositional variability in a study on GM maize grown at multiple locations in Chile (Xu et al., 2014). In addition, the genotypic diversity of analytes in rice was characterized using PCA and Pearson correlation analysis (Jiang et al., 2007). Bergman and Xu (2003) reported genotypic and environmental effects on the content of tocopherol, tocotrienol, and γ-oryzanol. Jiang et al. (2007) reported the correlation of the content of mineral elements and quality traits in rice.

The objective of this research was to assess the composition of the RR rice (with and without glufosinate-ammonium spray) using univariate and multivariate analysis by applying the principle of substantial equivalence. Compositional analyses included measurement of proximates, fibers, amino acids, fatty acids, minerals, vitamins, and phytic acid. PCA, hierarchical clustering (HCA), and Pearson correlation analysis were used to examine the distribution of information in the compositional data.