1. Introduction
Orchardgrass (
Dactylis glomerata) is an auto-tetraploid, out-crossing forage grass, and is one of the main perennial grasses used for grazing and hay production in temperate climates [
1]. Although
D. glomerata is the sole species in the
Dactylis genus, the species includes both Mediterranean and Continental ecogeographic subspecies that are diploid and tetraploid [
2]. Orchardgrass is highly preferred by livestock, exhibits early season growth, and is one of the most compatible perennial forage grasses when sown with perennial legumes, e.g., alfalfa (
Medicago sativa L.) [
1,
3,
4]. It is less tolerant to abiotic stresses than tall fescue and, for winter injury, than timothy, but more tolerant than perennial ryegrass. The forage quality of orchardgrass is intermediate, less than perennial ryegrass (
Lolium perenne), equal to timothy (
Phleum pratense), but greater than tall fescue (
Lolium arundinaceum). In the US orchardgrass is primarily utilized under rainfed conditions of the central and eastern portions, and under irrigated conditions in the semi-arid western portions [
5].
Acute and chronic drought can result in substantial decreases of forage grass biomass production [
6]. Winter damage caused by freezing includes delayed spring growth, uneven growth, and mortality. Available orchardgrass germplasm has substantial genetic variation for tolerance to drought and winter injury [
7,
8,
9], and cultivars have been developed with drought tolerance in mind (e.g., cv. Paiute). However, cultivar performance under different or stressed conditions has been inconsistent [
9], possibly due to a lack of uniformity in orchardgrass cultivars [
10,
11] or the unpredictability of appropriate environmental conditions in field settings [
12]. Improving forage production and quality under abiotic stresses are objectives in orchardgrass breeding and genetics programs throughout the world (e.g., [
13]).
In previous breeding efforts to improve perennial forage grass quality, it was noticed that water-soluble carbohydrate (WSC) accumulation in forage grasses was correlated with abiotic stress tolerance [
14,
15,
16,
17,
18,
19]. The WSC in crown tissues were found to stabilize lipids in cellular membranes [
19], prevent membrane leakage, and maintain osmotic potential of the cells under cold and drought stress [
17,
18,
20,
21,
22]. In addition to their role in abiotic stress tolerance, increased WSC in leaf and stem tissue can result in improved protein digestion, decreased nitrogen waste through feces and urine, and increased meat and milk productivity in feeding animals [
23]. In orchardgrass, the overwhelming majority of all WSC are fructans [
14,
24], which bridge the gap between resource availability and demands [
17]. Increasing WSC (i.e., fructans) is a tangible method of increasing both abiotic stress tolerance and forage quality in orchardgrass.
Although published reports on variation for orchardgrass WSC concentrations are limited, more information is coming forth in this arena. Robins et al. [
25] identified substantial genetic variation for WSC within orchardgrass germplasm. Sanada et al. found significant differences in crown [
13] and stubble WSC concentrations, and that WSC concentrations were related to winter hardiness and spring growth. Additionally they assessed European, Asian, and North American germplasm and found that European germplasm generally contained less WSC than other location sources [
26]. Casler et al. [
10] also detected differences in forage productivity and quality traits between North American and European orchardgrass varieties, although WSC was not included in that study. Other evaluations reported significant variation for herbage mass (HM) and forage quality under drought conditions, and found the WSC concentration to be correlated with forage productivity [
27,
28]. These reports indicate substantial variation for forage production and quality traits between different breeding programs, and suggested that increases in WSC can improve both traits. An understanding of the variation within a breeding program for WSC, and the relationship with forage productivity and quality, could result in great gains in this forage grass species.
In the study described herein, we examine the effect of differing environments in North America (US) and Europe (EU) on the expression of WSC, and the relationship between WSC and associated agronomic and nutritive value traits in populations of orchardgrass from the North American United States Department of Agriculture (USDA) Forage and Range (FRR) and European Deutsche Saatveredelung Aktiengesellschaft (DSV) breeding programs. The objective was to determine the utility of these breeding pools for the improvement of WSC and to determine how to capitalize on the genotype by environment interaction (GEI) effects underlying WSC, HM, and nutritive value expression in orchardgrass.
3. Discussion
Previous evaluations of North American and European orchardgrass breeding programs [
10,
30] noted substantial differences between the two germplasm sources for HM and some forage quality traits. The current study focused on the characterization of the WSC concentration in the germplasm pools and the relationship between WSC and the other measured phenotypes. Additionally, this study focused on the GEI effects on orchardgrass from the DSV and FRR breeding programs when grown at sites in the EU and US. The DSV populations exhibited comparatively high HM and RR, but lower nutritive value, and late maturity. In contrast, the FRR populations exhibited higher nutritive value, including WSC, but lower HM and earlier maturity. Maturity, in particular, reflects the strong preference for much later material in the EU and the preference for seed production of early flowering cultivars in the US. Because rust infection was only rated in September at the EU locations, this study did not allow the further elucidation of this relationship by correlating the response at the individual harvests.
The success of breeding perennial ryegrass for increased WSC [
31] has created a new objective for many forage breeding programs, including the DSV and FRR programs. Using near infrared spectroscopy approaches, WSC is easy to measure and characterize in targeted population improvement programs; however, its inheritance is strongly affected by location and GEI effects. High WSC perennial ryegrass cultivars may fail to express the phenotype in vastly differing environments, such as New Zealand [
32] and the US [
33]. The prevailing hypothesis for the GEI effect is that the higher summer temperatures limit the photosynthesis of the C
3 grasses and result in lower WSC concentrations in the forage [
32]. This hypothesis is consistent with the results of the current analysis. While annual maximum temperatures did not differ substantially among the locations, the mean maximum temperatures and growing degree days during the growing season were higher at the US locations. Interestingly, the EU locations are at higher latitudes and receive more sunlight for increased photosynthesis during the growing season, such that the EU locations resulted in a higher WSC concentration while the FRR populations exhibited a higher WSC than the DSV populations at those EU locations. However, there was no difference between the DSV and FRR populations for WSC at the US locations. These results are consistent with the report of Sanada et al. [
26], which also found that European orchardgrass sources generally exhibited a lower WSC than orchardgrass from other regions. Further studies of the genetics and GEI of WSC will be required to determine appropriate selection targets for increased WSC in different regions and to determine limits of WSC expression in orchardgrass germplasm.
The GEI effects detected herein were also complicated by the relative RR of the two germplasm sources included in this evaluation. The relationship between the WSC and disease is not well defined, but Sanada et al. [
26] found a negative relationship between rust infection and WSC. Thus, our results for WSC contradict this finding. The potential for increasing the WSC concentrations under US conditions is unclear. However, breeding for this trait is still in its infancy and substantial genetic variation exists in orchardgrass [
25]. Thus, increases in WSC are likely, but overall expression levels comparable to EU conditions may not be attainable.
The higher HM of the DSV populations compared to the FRR populations, at both US and EU locations, is seemingly inconsistent with the higher HM of US cultivars at US locations in previous work [
10]. At EU locations this difference may partially be due to disease pressure, where the effect of rust on the FRR populations was high and limited their productivity compared to the DSV populations. At the US locations rust was not a factor and the HM differences were not as striking, although the DSV populations as a whole had higher DM. Thus, the strong overall increased HM of the DSV populations was driven by their substantially higher HM at the EU locations.
Forage quality phenotypes included CP, cell wall components, and digestibility. At each individual location, the FRR populations had higher CP and DDM than DSV populations, but differences for NDF, ADF, NDFD were less or non-significant. In general, relationships between these phenotypes and the WSC were consistent with those identified previously in orchardgrass [
25], including the negative relationship between HM and cell wall components (ADF and NDF) and the positive relationship between HM and digestibility (DDM and NDFD). The non-significant correlation between CP and WSC in this study was unexpected. Crude protein and WSC usually exhibit a negative correlation because photosynthate diverted to the production of WSC cannot be used for CP, and vice versa [
23]. The non-significance in this evaluation is intriguing.
The genotype by environment (location by population) interaction was a key characteristic of this evaluation. The genotype by environment interaction is the rule rather than the exception in forage grass breeding [
25,
33]. However, a main objective of this study was the characterization of this interaction to allow future use of the GEI to maximize breeding goals in orchardgrass. The AMMI analysis suggested that a consistent relationship occurred between the two EU locations, but that the two US locations did not exhibit consistent phenotypic performance. Previously, the Logan and Rexburg locations exhibited substantial GEI for orchardgrass phenotypes, including WSC and HM [
25], such that there were no specific populations that would perform uniformly well for the measured phenotypes across those two locations. However, the consistent performance of the DSV and FRR populations for phenotypes suggests that combining these populations through recurrent selection can result in germplasm improved for WSC, HM, and RR. These populations could then be specifically selected for improved performance in the EU and US, respectively.