Get Complete Project Material File(s) Now! »
Importance of grasslands and perennial ryegrass
Grasslands are essential to ruminant feeding for ages. Depending on the soil and the climate, they could provide a good yield of a balanced ration (energy and protein) when exploited correctly (right physiological stage, fertilized and good species composition). Moreover, they are the cheapest source of feed when provided to ruminants directly through grazing pasture land (Huyghe et al., 2014). For these reasons, they are still largely used either grazed or mechanically harvested to produce conserved forage such as hay, wrapping forage or silage. Grasslands have other benefits than biomass production. They are providers of ecological services such as the preservation of water quality, biodiversity, landscape biodiversity and carbon sequestration.
Grasslands cover about 40 % of the agricultural area in Europe in 2013 (60 million ha) (van den Pol-van Dasselaar et al., 2019). Permanent grasslands (maintained more than five years) represent 85 % of these areas. The rest is composed of temporary grasslands, which are sown regularly with different bred forage species including legumes such as alfalfa or white and red clovers, and grasses such as Italian and perennial ryegrasses (Lolium multiflorum L. and Lolium perenne L.), timothy, fescues or cocksfoot. Temporary grasslands sown with elite varieties of Italian and/or hybrid ryegrass produce generally a better yield of a better quality than permanent grasslands and are mainly used in intensive systems in particular for milk production (Søegaard et al., 2007).
The area for seed production of forage plants in Europe represented 515 700 ha in 2020 which was at the first position just before soft wheat (GNIS, 2021). Forage grasses dominate the forage seed market and, in 2020, represented 262 963 ha with perennial ryegrass being the first one with 88 140 ha. In 2019, the exchange of forage grass seeds represented 325 Million € (SEMAE personal communication). Among grass forage species, perennial ryegrass is the main species used in temperate climate with about 70-90 000 tons of seeds used per year in Europe for forage and turf between 2007 and 2015 (Bruins, 2016). It is often sown in association with white clover and is well known for its high quality, its good persistency under temperate climate, its good response to nitrogen fertilizers and its very good behaviour under grazing (McDonagh et al., 2016; Sampoux et al., 2011).
General presentation of perennial ryegrass
Perennial ryegrass: Lolium perenne L., belongs to the genus Lolium, the sub-tribe Loliinae, the tribe Poeae, the super tribe Poodae, the sub family Pooideae and the family Poaceae (http://lifemap-ncbi.univ-lyon1.fr/?tid=640630). Perennial ryegrass is phylogenetically close to major crops such as oat, wheat and barley and to the grass model species: Brachipodium distachyon (Figure 1).
Figure 1: A phylogenetic ranking of the Poaceae evolving clockwise from Joinvillea and Ecdeiocolea. BOP: Bambusoideae, Oryzoideae, and Pooideae; PACMAD: Panicoideae, Aristidoideae, Chloridoideae, Micrairoideae, Arundinoideae, and Danthonioideae. Modified from Soreng et al. (2017).
Perennial ryegrass is a perennial herbaceous plant composed of roots, a crown with non elongated internodes during the vegetative stage (Figures 2 and 3) and tillers: main tiller and daughter tillers. The absence of internode elongation during the vegetative stage is a powerful way to protect apical meristems from defoliation by herbivores and to provide a good persistency under grazing. Each tiller consists of an apical meristem (apex), nodes with a leaf and a lateral meristem at the origin of a daughter tiller and internodes. A leaf is composed of a sheath and a lamina separated by a ligule (Figure 4). Each leaf grows within the sheath of its previous leaf. The sheaths and growing leaves within the sheaths during the vegetative stage are often called pseudo-stem by contrast with true stems corresponding to elongated internodes and nodes during the generative or reproductive stage. During this generative stage, the apical meristem produces lateral meristem which will produce spikelets instead of vegetative tillers and the internodes elongate. The increase of length of the successive internodes during the lifespan of a tiller from vegetative to generative stages is presented in Figure 5. The beginning of the generative stage is triggered by both a vernalisation and an elongation of day length (Rouet, 2021).
Naturally, perennial ryegrass is diploid with 2n = 2x = 14 chromosomes but tetraploid varieties have been created. Its haploid genome size corresponding to 1X is 2.68 pg = 2.62 Gb (Pellicer and Leitch, 2019). The genome of perennial ryegrass has been recently sequenced (Byrne et al., 2015) and assembled (Frei et al., 2021). A limited number of different molecular markers have been produced (microsatellite, AFLP, amplicon sequencing) (Forster et al., 2004; Roldan-Ruiz et al., 2019) but a major step was the production of several hundreds of thousands markers by a simple genotyping-by-sequencing (GBS) method based on restriction enzymes (Elshire et al., 2011). This method allows to genotype individual plants but also populations for which the allelic frequencies could be estimated with a single DNA pool per population (Byrne et al., 2013). This method allows to obtain markers all over the genome but not at selected positions. For this reason, new markers are developed based on previous knowledge (candidate genes, QTLs) to sequence selected loci by either high throughput amplicon sequencing or by sequencing capture (Roldan-Ruiz et al., 2019).
Perennial ryegrass is an obligate outcrossing species due to a self-incompatibility system and presents a strong inbreeding effect (Slatter et al., 2021; Thorogood et al., 2002). The absence of a simple method for inter-crossing selected plants at a large scale like in maize leads to the creation of synthetic varieties originating from the crossing of a few founders (between 4 to 20) followed by two to three generations of panmictic multiplication in order to produce enough seeds for the market need and to create a stable off spring from the synthetic progeny. These biological constraints with the practical and financial constraints lead to the definition of breeding schemes. A classical breeding scheme is proposed on Figure 6. Starting from a panel of diverse material (the construction of this panel is essential for the success of the breeding programme), plants are first selected only on their phenotype observed in spaced plant field trail (mass selection) during three years for highly heritable traits such as disease resistance, vigour, heading date, persistency, growth habit. The selected ones are crossed in polycrosses (inter-crossing of a set of plants) to obtain half-sib families (each mother plant is harvested separately) or alternatively in pair-crosses (inter-crossing of two plants) to obtain full-sib families. Then the families are selected based on their genetic value estimated in forage yield plots with replicates and possibly several locations during three years for low heritable traits such as yield and quality. Then randomly selected plants within the selected families are crossed in polycrosses to obtain synthetics which will be multiply for further phenotyping evaluations in forage plots. The best ones are selected to be evaluated by national institutions like GEVES (Groupe d’Etude et de contrôle des Variétés Et des Semences) in France for three years in order to be registered by CTPS (Comité Technique Permanent de la Sélection des plantes cultivées) if they pass the tests : DUS (distinctness, uniformity and stability) and VCU (value for cultivation and use). This breeding scheme could release a variety after about 15 years. This long process could be shortened be using genomic selection which is currently being implemented in some breeding companies (Barre et al., in press; Cericola et al., 2018; Fè et al., 2016, 2015; Hayes et al., 2013; Lin et al., 2016, 2014; Pembleton et al., 2018). Moreover when selecting on half-sib families, the genetic gain could be increased by using molecular markers for paternity tests in order to select siblings within a family having the best paternal parents instead of selecting randomly.
The main targets for selection are forage yield, quality, disease resistance, heading date (since for registration, varieties are classified within precocity groups) and persistency. The importance of forage quality in breeding programmes increased particularly when in 2013 it has been included in the variety testing for registration in France (including proteins, water soluble carbohydrates and lingo-cellulose (ADF) contents). One of the main difficulties to select for forage quality is that it depends on many factors including the stage of development when harvested which could hide genetic variation for quality behind variation for heading date.
Figure 6: A classical example of breeding scheme for perennial ryegrass (Barre et al., in press).
Evaluation of forage quality
In order to build a ration to cover animal needs, it is necessary to evaluate the quality (equivalent to the digestibility in this thesis not considering the micro-nutriments) of the forage i.e. the energy and the protein content brought by the forage usable by the ruminant. For that purpose, many methods have been developed through years with the in vivo apparent digestibility being the reference method. In this part, we will go through all currently used protocols. Starting with the in vivo methods, we will continue with in situ (or in sacco) method, then we will see the large diversity of in vitro methods.
In vivo methods
The principle of in vivo methods consists in measuring the difference between the feed given to an animal and the produced feces: total collection. This method is performed since the beginning of forage digestibility measurement in the XIX century (Schneider and Flatt, 1975). It allows to measure “apparent digestibility” following one assumption: the proportion of feed intake not excreted is totally digested. We talk about “apparent digestibility” because the feces contain, in addition of undigested food, some materials produced by the animal body (Schneider and Flatt, 1975) and a part of carbohydrates, fermented into methane, is lost as eructation and is not absorbed by the animal (McDonald et al., 2011).
The main protocol can be divided in multiple steps:
– The feed is analysed for its chemical composition (% of dry matter, crude protein, ash, ether extract, crude fibre, nitrogen-free extract and gross energy);
– Animals are individually locked in a metabolism crate, a sort of cage used to facilitate feces collection, and fed with only the analysed feed. The feed intake is calculated as the weight of proposed feed minus the weight of refused feed;
– After a period of time necessary to rid the digestive system of the precedent feeds, feces are started to be collected and weighted separately of urines;
– Samples of feces follow the same chemical analysis path than feed;
– Differences between results from chemical analysis of feeds and feces allow to obtain the digested percentage of each component of the feed and the feed digestibility itself: (feed intake – feces) / feed intake x 100 in percent.
In situ method
Also called in sacco method, the in situ technique requires the use of fistulated ruminants. The method allows the estimation of small samples of 2 to 5 g, which are milled dry feed (to pass a 3 mm sieve) or minced wet feed (Ørskov, 2000). The samples are placed in nylon bags with a pore size around 40-60 µm, which are inserted inside the rumen for a maximum of 48 h. After that, bags are washed and weighted. The difference of sample weight gives an estimation of the dry matter digestibility. It is also possible to compare the amount of nitrogen or fibre between undigested samples and digested ones to determine their digestibility. This technique is particularly used for determining degradability kinetics of samples.
In vitro methods
Tilley and Terry (1963) proposed a two-stages in vitro technique that has become the reference. The principle consists in measuring the difference of weights of a sample before and after two incubations: a first incubation of the sample in rumen liquor from fistulated animals with synthetic saliva followed by a second incubation in an acid aqueous solution of pepsin. More precisely, the samples are dried, milled (to pass a 1 mm sieve), weighted and incubated with synthetic saliva (McDougall, 1948) and strained rumen liquor into 80-90 mL at 38°C for 48 h in a carbon dioxide atmosphere. The samples are agitated and the pH checked to be maintained between 6.7 and 6.9 to optimize conditions for microbial growth and development. After 48 h of incubation, mercuric chloride (HgCl2) and sodium carbonate (Na2CO3) are added to inhibit further microbial activity and aid sedimentation. The supernatant is discarded following filtration or centrifugation. Acid pepsin is added to the residue and the mixture incubated again for 48 h at 38°C. The sample is centrifuged again, the supernatant discarded and the residue washed. This undigested residue is dried. Digestibility is calculated by dividing the weight loss of the sample by its initial weight, after correction for the blank weight of the diluted rumen liquor (Omed et al., 2000). This method was adapted to use a filter bag system instead of digestion tubes allowing the analysis of several samples in a single jar (Wilman and Adesogan, 2000).
To avoid the necessity of fistulated cattle, a method was investigated by Balfe (1985) with the use of a liquor obtained from sheep feces. This method was modified by El Shaer et al. (1987) by the use of McCartney bottles fitted with wine fermentation locks. Next, Omed et al. (1989a, 1989b) found that the procedure could be carried out satisfactorily in sealed bottles, dispensing with the fermentation locks. The sample size was reduced to 180 mg and the faecal liquor volume scaled down correspondingly to 18 ml.
To avoid the use of animals, several enzymatic solubility protocols have been proposed on the basis of pepsin-cellulase technique (Aufrère, 1982; Donefer et al., 1963; Jarrige et al., 1969; Jones and Hayward, 1975; Kowalski et al., 2014). The large amount of protocols is due to each laboratory using its own buffer recipe, material (crucible, Erlenmeyer, filter bag for exemple) and incubation duration. Also, the provider of cellulase (mainly obtained from fungus Trichoderma viride) change following availability for each laboratory, giving the obligation to adapt the protocol to the enzymatic activity of cellulase and its optimum of pH and temperature. This technique is the most repeatable method with small amount of dry matter, as rumen liquor (or faecal liquor) is highly dependent on the animal (diet, sex, age). This technology contribute to facilitate researches performing digestibility evaluation. In France, the Aufrère protocol (1982) act as reference for pepsin-cellulase technique. Its adaptation for the use of Ankom filter bags was recently made by Méchin V. (not published).
In practice, even if the in vivo method stays the reference for all digestibility evaluations, in vitro methods are the most commonly used due to their rapidity, facility, low cost and reapetability. In situ methods are mainly used to observe digestibility kinetics instead of measuring general quality. But the future of digestibility measurement is in the use of spectrometry.
Near infrared reflectance spectroscopy (NIRS) is a technology that can predict various feed components without destruction (Deaville and Givens, 1998). NIRS is an analytical technique based on the absorption of infrared radiations by the chemical bonds in organic matter (Bastianelli, 2013). This absorption is linked to the chemical composition and can be measured with a spectrometer. In order to predict a component proportion, it is necessary to calibrate the NIRS with the development of mathematical models, serving as a translator between infrared spectrum absorbed and the sample composition. A model is specific for a component and more or less specific depending on the type of sample. To develop a robust model, it is necessary to integrate a large and varied dataset of samples. This model could then be updated with the addition of dataset (chemical composition or in vitro or in vivo digestibility) from new samples.
In conclusion, techniques to evaluate digestibility have been simplified with time but the in vivo technique stays the reference from which all other methods have to be calibrated. These techniques could be applied directly on forage or on fibres to evaluate their digestibility.
Origins of the quality variability in forage grasses
Since the 70s, thanks to in vitro methods and later to the use of NIRS to evaluate digestibility, many forage samples have been surveyed for their quality. Rapidly, it came out that many factors affect the forage quality including the plant maturity which is in close relationship with the type of organs (leaf, stem and inflorescence) and their morphology, the chemical composition of cells, the crop management (fertilisation in particular nitrogen, frequency of defoliation, use of fungicide, watering), the growth environment (temperature, drought, freezing…) and the genetic variability between and within species (reviewed in Jung et al., 1993). We propose here to give the main conclusions with an update.
Plant maturity, type of organs and their morphology
One of the major factors influencing forage quality is plant maturity (Nelson and Moser, 1994). This is mainly due to the evolution in the proportion of plant organs: leaves, sheaths and stems, which have different level of quality: leaves > sheaths > stems > dead leaves (Beecher et al., 2013; Chaves et al., 2006). For example, in big bluestem (Andropogon gerardii, Vitman) and switchgrass (Panicum virgatum, L.), at spike emergence, the dry matter digestibility (DMD), neutral detergent fibre content (NDF) and lignin contents of leaves were 60, 66 and 4.7
% DM whereas the ones of stems were 50, 75 and 7.2 % DM (Griffin and Jung, 1983). During the reproductive stage, the proportion of stems in percent of plant weight increases leading to a decrease of the quality of the harvested forage (Buxton and Redfearn, 1997). This difference of quality between leaf and stem evolves with the organ maturity i.e. its age (Wilman and Altimimi, 1982). The leaf quality evolves slightly during its lifespan with a cell wall mass almost constant but a decrease in cell wall digestibility (Groot et al., 1999). Young stems are generally of high quality (digestibility could be better than the leaves at the same stage) (Minson, 1990; Wilman and Altimimi, 1982) but their quality drops rapidly with internode elongation which is easily noticeable after spike emergence (Nelson and Moser, 1994). This could be attributed to anatomy differences, leaves being composed mainly of mesophyll cells with thin cell walls whereas stems containing many highly lignified xylem cells. These lignified cells are the major factor providing structural strength but it is also the greatest limitation to the breakdown of stems in the rumen (Akin, 1989). By comparison, stem parenchyma is easily digested.
Green leaf tissue is often the highest quality part of the forage (Hides et al., 1983). Nevertheless, the structure of leaves should be taken into account. An increase of the sheath/blade ratio decreases the quality since the quality of sheath is lower than the quality of blade with for example, after spike emergence, a blade DMD of about 80 % and 70 % in perennial ryegrass and cocksfoot, respectively, compare to 70 % and 65 % for sheath (Terry and Tilley, 1964). Moreover, leaves’ quality decreases with their lengths; this effect is often confounded with the rank effect since leaf length increases with its rank (Agnusdei et al., 2011; Duru and Ducrocq, 2002; Gastal and Lemaire, 2015; Groot et al., 1999). This length or rank effect is due to an increase of the sheath/blade ratio and to an increase of structural tissue content (Insua et al., 2018, 2017). It has been shown that leaf length is more important than leaf age to explain the variation of leaf digestibility (Avila et al., 2010).
Table of contents :
1.1. Importance of grasslands and perennial ryegrass
1.2. General presentation of perennial ryegrass
1.3. Evaluation of forage quality
1.3.1. In vivo methods
1.3.2. In situ method
1.3.3. In vitro methods
1.3.4. Spectrometry method
1.4. Origins of the quality variability in forage grasses
1.4.1. Plant maturity, type of organs and their morphology
1.4.2. Cell-Wall structure
1.4.3. Environmental factors
1.5. Plant modelling: individual based model, the example of L-grass
1.6. Objectives of the thesis
2. Seasonal differences in structural and genetic control of digestibility in perennial ryegrass
2.2. Material and Methods
2.2.1. Plant material
2.2.2. Trial design
22.214.171.124. Wet chemical analysis and NIRS prediction
126.96.36.199. Derived biochemical data
2.2.4. Statistical analysis of the phenotypic data
188.8.131.52. General models and heritability
184.108.40.206. Drivers of OMD and NDFD
2.2.5. Genetic markers
2.2.6. Correction for heading date in spring
2.2.7. Population structure
2.2.8. MLMM method
2.3.1. Impact of season on OMD and NDFD
2.3.2. Effect of cell wall content and cell wall digestibility on spring and autumn digestibility
2.3.3. The effect of cell wall composition on NDFD
2.4.1. Seasonal differences in OMD and NDFD among genotypes
2.4.2. NDFD and NDF determine OMD in both spring and autumn
2.4.3. Genetic control of digestibility
2.5. Key points for chapter 2
3. Dynamics of the quality during the lifespan of a leaf and a stem of perennial ryegrass: a survey of genotypes contrasting for their precocity and their quality
3.2. Material and methods
3.2.1. Plant material
3.2.2. Trial design
3.2.4. Wet chemical analysis and NIRS prediction
3.2.5. Statistical analysis of phenotypic data
3.3.1. General description of the four genotypes
220.127.116.11. Description of leaves
18.104.22.168. Description of stems
3.3.2. Origin of DMD variability
22.214.171.124. In leaves: Senescence effect and leaf length
126.96.36.199. In stems: structural tissues and their digestibility
3.3.3. Consequences of DMD variation at the tiller level
3.4.1. Leaves morphogenesis and quality
3.4.2. Stems morphogenesis and quality
3.4.3. Tillers morphogenesis and quality
3.5. Key points for chapter 3
4. General discussion