Defining a Phenotypic Variability and Productivity in Wild Type Red Clover Germplasm

Abiotic and biotic factors can cause great damage to crops. So, a key approach is to investigate whether the crops’ wild relatives are more flexible to withstand abiotic and biotic stress. As well as to evaluate their phenotypic variability and productivity in response to changing climatic conditions. In this study, red clover germplasm was collected from natural red clover habitats and a field trial was arranged ex situ. Twelve phenotypic traits and their effects on final harvest were analysed in 2018–2019. Principal component analysis (PCA) demonstrated that the most important trait for biomass yield was the height of the plant during the first season of harvest (2018). Interestingly, that the most significant trait in the second year of harvest (2019) was growth habit. Meantime, two way-joining analysis was performed to extent of phenotypic variation within and among red clover accessions, based on the most important trait for biomass yield. We found three main groups based on variation in plant height: “cultivars”, “wilds” and “mediators”. This analysis leads to identify typical populations of wild type red clover, which has not been done yet. Finally, the feed value of each red clover accession was analysed. It was found that “cultivars” have a higher level of crude proteins, while “wilds” contains higher levels of crude fibre. This indicates that there is a relationship between plant structure elements and forage value which is particularly important to select a breeding material.


Introduction
Red clover (Trifolium pratense L.) is quite a common species, well-recognized by the public. Historically red clover is native to the Mediterranean basin. It spread around the world from the 16 th century as a consequence of expeditions of exploration and the settlement of new lands, projects which were at their most intense up to the 18 th century (Annicchiarico, Barrett, Brummer, Julier, & Marshall, 2015;Taylor & Quesenberry, 1996;Taylor, 2008).
Red clover became favoured due to its beneficial health properties. Its chemical composition was identified as equipment and laboratory methods improved at the turn of the 20 th century. It was found that red clover is rich in secondary metabolites: isoflavones, flavonoids, coumarin derivatives, cyanogenic glycosides, volatile oils, etc. (Vlaisavljević et al., 2017;Butkutė et al., 2018). Even before this scientific validation, the value of red clover was recognized in organic farming, due to its ability convert and absorb atmospheric nitrogen in symbiosis with Rhizobium (Thilakarathna et al., 2016). Moreover, red clover is highly valued due to its high level of crude proteins and other feed qualities, therefore this species is widely used to implement greening programs and to feed cattle (Cassida et al., 2000). There are also beekeepers, who grown red clover for its high provision of nectar, although researchers report that red clover is a more suitable source of nectar for bumble bees than honey bees, who cannot reach the nectar from its long corolla tubes (Sands & Rowntree, 2016;Vanommeslaeghe et al., 2018). While other studies have shown that bees pollinate red clover quite readily, especially the wild forms that have a shorter corolla tube than the cultivars (Rao & Stephen, 2009;Vleugels et al., 2019). Schmidt, Pacher, Houben, & Puchta, 2020). However, new and original genetic material is needed for breeding programmes. We are in agreement with researchers, that this need can be met by constant monitoring and by collecting germplasm from wild communities (Morris & Greene, 2001;Solberg et al., 2017;Jones et al., 2020).
Unfortunately, information on the morphology of wild type red clover is not abundant and phenotyping of such crops wild relative is limited by lack of efficient techniques in situ (Moreira, Oliveira, Volenec, Rainey, & Brito, 2020). Despite this challenge, it is important to assess and understand the potential of crops wild relative for adaptation to a changing climate and new habitats. Longitudinal monitoring is an effective method of assessing the most important traits and a basis on which to construct a model of future requirements under climate change (Moreira et al., 2020). The model would be useful if harvests were assessed and the most promising populations were chosen for breeding programs according to the selected traits. Inter alia, longitudinal monitoring can detect and demonstrate phenotypic variation within and between populations or cultivars. Moreover, a good model would promote reliable renewal of breeding material or complete elimination old cultivars from the market.
Statistically the model could be implemented, because expressed variation determines the power of variability which describes the potential for variation in population (Willmore et al., 2007). Solberg et al. (2017) have noted that variation in wild red clover may be a response to natural or artificial selection or even genetic drift, while the effects of climate change have forced the flora to adapt to abiotic and biotic factors (Gratani, 2014). Meantime, the most plastic individuals are adapted to survive, while these factors are acting in plant communities (Gratani, 2014;Pagnotta, Annicchiarico, Farina, & Proietti, 2011). We believe this is strongly expressed in red clover populations because this species has different phenotypic types.
In this study, the phenotypic variability of wild red clover and its biological expression by integrating germplasm ex situ was analysed. In addition, the most important trait for biomass yield was distinguished and forage value was evaluated. Ultimately, a prototype for breeding was identified.

Materials and Methods
The research was carried out from 2016 to 2019. In the first year (2016) seeds were collected from wild populations inventoried as NATURA 2000 (codes 6510, 6530, 6270 and 6210) across Lithuania. Full ripened flower heads were threshed, and the seeds were sorted from admixture in each population. The seeds were dried for three months at 20 °C temperature during 15% relative humidity and then stored at +4 °C until sowing began on 18 April 2017. Totally 49 accession of populations and cultivars were sown. Either, seeds of wild red clover which had been stored in the Lithuanian Plant Gene Bank (PGB) were used in this study. The seeds from PGB had been dried to 3-6% relative humidity and stored at -18°C temperature for at least 10 years. Also, diploid cultivars of red clover bred in Lithuania ('Arimaičiai', 'Kamaniai' and 'Vytis') were used as accessions along with two diploid cultivars (2887 and 2889) that were received from other countries' PGBs. One tetraploid cultivar (2908) was used as control in relation to the diploid cultivars. Finally, cultivar 'Liepsna' was selected as a standard; this cultivar is one of the oldest and has been grown in Lithuania over long time period, therefore, it is likely that germplasm of 'Liepsna' have been spread to wild populations.
In the second year, on 18 April 2017 wild red clover seeds were sown in growing trays filled with a peat and soil mixture (30:70). While plants were transplanted ex situ on 10 July 2017. The field trial was arranged in a randomized complete block design with four replications, located at 55°23′35.0″N, 23°52′38.6″E. Endocalcari-Epihypogleyic Cambisol soil with a moderately heavy loam texture predominates in the field. The soil pH KCl 7.5, humus content 2.64%, available P 2 O 5 220 mg kg -1 and K 2 O 156 mg kg -1 . Fertilizers of N 6 P 18 K 34 (24 kg N, 31 kg P 2 O 5 and 113 kg K 2 O) at a dose of 400 kg ha -1 were spread before planting, while no fertilizer was used during the experiment. However, weeds were controlled mechanically and using active agent (a.i. bentazone 480 g l -1 ) at dose of 960 g ha -1 twice per season.
Twelve different agro-morphological parameters were evaluated during the first (2018) and second (2019) year of harvest. Abbreviations of measured and scored characteristics are shown in Table 1. Visual assessment of plant traits was performed based on the guidelines of the International Union for the Protection of New Varieties of Plants (2016).  Vol. 12, No. 9; 'Liepsna' produced (Table 2). Meanwhile, four populations had a significantly higher DMY than 'Liepsna' but did not differ significantly in GMY. There was no significant different in GMY and DMY between all the remaining populations and the standard. The lowest GMY and DMY were found in population 2886 (0.334 and 0.094 g, respectively), while the highest difference between GMY and DMY was found in the cultivar 'Kamaniai' (0.996 g). Thus, it can be stated that wild forms of red clover tend to produce a lower biomass yield; however, a significant difference depends on the relationship to that standard cultivar. Nevertheless, wild forms and landraces can be used to breed highly productive cultivars. Likewise, Hoekstra et al. (2018) reported that Switzerland locally adapted red clover landrace called "Mattenklees" produced a higher biomass yield than traditional cultivars of "Ackerklees" type. Note. *: significantly higher than standard cultivar 'Liepsna' P ≤ 0.05 (Dunnett's test).
In terms of feed quality indicators, the lowest amount of CP was found in the population 2902 (10.6%), and the highest in cultivar 2887 (19.7%), while 'Liepsna' tended to produce 17.6% CP which is higher than that found in other Lithuanian bred cultivars. The lowest content of CF was found in cultivar 2887, while the highest quantity was produced by population 2885. Meantime 'Liepsna' had 28.6% CF, which was a lower amount than 'Kamaniai', 'Arimaičiai' and the remaining 27 populations produced.
It was found that population 2876 had the lowest water-soluble carbohydrates (WSC) content (5.71%), while 'Vytis' tended to produce 10.2%, which was the highest amount of WSC among all accessions, including the standard cultivar 'Liepsna' (9.06%). Meanwhile, two populations (2893 and 2898) were characterized by higher WSC content than the standard. Thus, we may conclude that red clover tends to produce higher levels of WSC in response to G × E interaction.
Results of our experiment showed that those populations tend to form more stems with smaller leaves than cultivars, and those stems contain a higher percentage of CF than leaves, whereas cultivars have a higher percent of CP due to their higher leaf to stem ratio (Hoekstra et al., 2018;Tucak, Popović, Čupić, Španić, & Meglič, 2013). Meantime, WSC content is more dependent on the environmental factors such as drought and even depends on the time of day (Kagan et al., 2020;Ruckle et al., 2017). Thus, the highest amounts of CP are accumulated in the leaves (Taylor, Quesenberry, 1996). CF strongly negatively correlates with CP, so populations with higher CP level tend to have lower CF levels and vice versa (Hai-xia et al., 2013). Similar trends of CP were noted by Hoekstra et al. (2018), who reported that CP concentration differed slightly between cultivars, while there was a significant difference between cultivars and landrace type.
However, there is a lack of information about the feed quality of wild type red clover, while there is more data about breeding lines which are bred for several generations ex situ and seemed to differ from their ancestors.

Conclusions
Natural plant height (NPH) in the first year of harvest (2018) was the most important trait for biomass yield. Whereas grow habit (GRH) became the most important indicator in the second year of harvest (2019), with NPH and all other traits becoming less influential.
High variation of NPH was found within populations, while the low variation was identified among those populations in terms of NPH. This shows that NPH lost relevance in the second year of harvest (2019) and all accessions became more resemble.
GRH was constrained by the biological potential of plants to regrow after being cut. Results of our experiment confirms the claim that cultivars' regrowth ratio is higher than that of wild populations.
Two way joining analysis based on the most important traits for biomass yield have clustered red clover accessions in three groups "cultivars", "wilds" and "mediators". The largest group between them all was "mediators", who stood out with a high phenotypic variability within each population.
The vast majority of the populations in the field trial produced a lower green (GMY) and dry (DMY) matter yields than the standard cultivar 'Liepsna' but did not differ significantly.
The cultivars attended to produce higher rate of foliage; as a result, plants had higher crude protein (CP) content. Meantime, wild red clover populations were found to have a high stem-leaf ratio, thus producing a higher proportion of crude fibre (CF) than that of cultivars.