Proanthocyanidin diversity in the EU ‘HealthyHay’ sainfoin (Onobrychis viciifolia) germplasm collection

Abstract

This study investigated 37 diverse sainfoin (Onobrychis viciifolia Scop.) accessions from the EU ‘HealthyHay’ germplasm collection for proanthocyanidin (PA) content and composition. Accessions displayed a wide range of differences: PA contents varied from 0.57 to 2.80 g/100 g sainfoin; the mean degree of polymerisation from 12 to 84; the proportion of prodelphinidin tannins from 53% to 95%, and the proportion of trans-flavanol units from 12% to 34%. A positive correlation was found between PA contents (thiolytic versus acid–butanol degradation; P < 0.001; R² = 0.49). A negative correlation existed between PA content (thiolysis) and mDP (P < 0.05; R² = −0.30), which suggested that accessions with high PA contents had smaller PA polymers. Cluster analysis revealed that European accessions clustered into two main groups: Western Europe and Eastern Europe/Asia. In addition, accessions from USA, Canada and Armenia tended to cluster together. Overall, there was broad agreement between tannin clusters and clusters that were based on morphological and agronomic characteristics.

Introduction

Plant breeding is facing a paradigm shift and needs to develop new cultivars that can deliver lower environmental impact from agricultural production (Pilippot and Hallin, 2011). Recent evaluations of rice germplasm have already identified 500% differences in methane emissions, whilst a survey of wheat and other germplasm material discovered an enormous potential for plants with much higher biological nitrification inhibition activity. Sainfoin (Onobrychis viciifolia) is receiving renewed interest because of its many positive attributes (Hayot Carbonero et al., 2011, Website: http://sain, 0000). Consumption of sainfoin shows promise for reducing environmental emissions of nitrogen and methane from ruminant animals (Hatew et al., 2011, Theodoridou et al., 2010) and unlike many other forage legumes, sainfoin generates high voluntary intakes without causing bloat (Mueller-Harvey, 2009). In addition, its flowers are a valuable source of pollen and nectar for bees (Clement et al., 2006, Kells, 2001, McGregor, 1976) and could thus contribute to alleviating problems associated with pollinator decline. Sainfoin has also been shown to control parasitic worms, which are posing a serious threat to animal welfare and farming (Hoste et al., 2006, Novobilský et al., 2011).

Content and composition of tannins, also called proanthocyanidins (PAs), are heritable traits that are amenable to plant breeding (Orians et al., 2000, Scioneaux et al., 2011). PAs are also the active ingredients in sainfoin and their variation is, therefore, worth exploring amongst currently available accessions. Although several studies have reported better absorption of nitrogen and essential amino acids by ruminants from tannin-containing sainfoin than from iso-nitrogenous but tannin-free forages, contradictory findings also exist (Hayot Carbonero et al., 2011). Here, we set out to investigate whether the contradictory reports could be due to variation in tannins among different sainfoin lines as indicated previously (Marais et al., 2000). It was for these reasons that the EU ‘HealthyHay’ project established a 300+ sainfoin germplasm collection (Website), which was evaluated for nutritional, anti-parasitic, chemical, biochemical, genetic and agronomic characteristics. The present study focussed on variation in PA content and composition within a subset of 37 diverse sainfoin accessions.

Many analytical methods have been used to quantify and characterise PAs in plants (Fig. 1). The widely used HCl–butanol method is an acid-catalysed oxidative depolymerisation of the PA interflavanyl linkages and yields anthocyanidins (Fig. 2), which can be determined spectrophotometrically (Porter et al., 1985). However, PA measurements based on anthocyanidin yields have several limitations: (a) water in the reaction medium can suppress the anthocyanidin yield (Waterman, 1994); (b) the number of hydroxy groups in the A and B rings affect the wavelength of the absorbance maximum of the anthocyanidin products (Fig. 2) (Hemingway, 1989); (c) colour yield is not always linearly correlated with tannin concentration (Waterman, 1994); (d) the ratio of reagent to sample, the temperature and the length of the reaction time affect the colour yield (Scalbert, 1992, Waterman, 1994); (e) the lack of appropriate standards remains a problem for quantifying complex mixtures of PAs (Giner-Chavez et al., 1997, Hagerman and Butler, 1994); and (f) the acid–butanol assay is generally used to quantify soluble PAs, however, some PAs are insoluble in common solvents and this can lead to an underestimation (Makkar et al., 1999, Reed, 1986). Although the HCl–butanol reaction has also been applied directly to plant samples (Kayani et al., 2007, Makkar et al., 1999, Yu and Dahlgren, 2000) the above limitations still apply and, in addition, the presence of chlorophyll interferes with anthocyanidin measurements (Watterson and Butler, 1983). Taken together these limitations mean that the HCl–butanol method is not suitable for quantitative PA analysis in extracts or whole plants.

Therefore, Gea et al. (2011) recently developed a thiolysis method for the in situ analysis of complex PA mixtures in sainfoin (O. viciifolia). This method has now been used to screen representative samples from the EU ‘HealthyHay’ sainfoin germplasm collection at NIAB (Cambridge, UK) in order to:

• to characterise the extent of PA diversity in this sainfoin collection

• to investigate correlations amongst PA structural traits within this collection

• to identify chemovars with contrasting PA composition for subsequent investigations into their biological effects and for supporting a future sainfoin breeding programme, and

• to explore any linkages between PA composition, sainfoin accessions and geographical origin.