Clinal variation in quantitative traits but not in evolutionary potential along elevational and latitudinal gradients in the widespread Anthyllis vulneraria

Premise of the study

Strong elevational and latitudinal gradients allow the study of genetic differentiation in response to similar environmental changes. However, it is uncertain whether the environmental changes along the two types of gradients result in similar genetically based changes in quantitative traits. Peripheral arctic and alpine populations are thought to have a lower evolutionary potential than more central ones. 

Methods 

We studied quantitative traits of the widespread Anthyllis vulneraria in a common garden. Plants originated from 20 populations along a 2000 m elevational gradient from the lowlands to the elevational limit of the species in the Alps, and from 20 populations along a 2400 km latitudinal gradient from the centre of the distribution of the species in Central Europe to its northern distributional margin.

Key results

Several traits showed similar clinal variation with elevation and latitude of origin. Higher QST-values than FST-values in some traits indicated divergent selection. The same traits were subject to strongly diversifying selection among populations (high QST) and strong stabilising selection within populations (low evolvability). Genetic diversity of most quantitative traits and neutral molecular markers was only weakly correlated. Plasticity in response to benign conditions declined with both increasing elevation and latitude of origin, but the evolvability of most traits did not.

Conclusions 

The clinal variation suggests adaptive differentiation of quantitative traits along the two gradients. Our results indicate that the evolutionary potential of peripheral populations is not necessarily reduced. However, lower plasticity may threaten their survival under rapidly changing climatic conditions.

Methods
Sampling —

We studied 20 populations of Anthyllis vulneraria along a 2400 km latitudinal gradient from the center of its distribution in Central Europe (46.4 °N) to its northern distribution limit in Scandinavia (68.1 °N) and 20 populations along three elevational gradients in the French, Swiss and Austrian Alps from 500 m to the elevational limit at 2500 m a.s.l. (Fig. 1; Daco et al., 2021; Appendix S1, Table S1; see Supplemental Data with this article). The length of the two gradients was chosen to correspond to a change of 11.5 °C in annual mean temperature.

In summer 2015, towards the end of the flowering period, we recorded at each site the elevation above sea level, latitude and longitude with a GPS (eTrex 20, Garmin Ltd.). We collected fruitheads from 20 plants/population along a 20 m transect and placed them in separate paper bags. To compare trait values in the field and in the common garden, for each mother plant we determined the height of the tallest flowering stem, the diameter of the rosette, the width of the terminal leaflet of the longest basal leaf, the number of stems with flowers (stems), and the number of flowerheads. In the laboratory, we extracted all healthy seeds (i.e. green and large) from the fruitheads of each mother plant.

Cultivation in the common garden —

In April 2016, ten seeds from each mother plant were scarified by rubbing them between sheets of sand paper, placed on moist filter paper in Petri dishes and kept at 20 °C in a greenhouse for germination. After five days, five seedlings (if available) per family (Appendix S1, Table S1) were planted into square pots of 11 cm x 11 cm x 12 cm filled with a 3:1 mixture of low-nutrient soil (Substrat 1, Klasmann-Deilmann GmbH, Geeste, Germany) and sand. The plants were randomly placed outdoors in a common garden of the municipal park service of the city of Luxembourg. Plants were watered when necessary and re-randomized several times.

Measurements of quantitative traits —

In July 2016, we recorded which of the initially 3207 plants had survived and recorded the following traits for each plant: number of leaves, diameter of the rosettes, and the width of the terminal leaflet of the longest basal leaf. We measured leaf chlorophyll content with a chlorophyll meter (SPAD-502 Plus, Minolta, Osaka, Japan) and transformed the values into chlorophyll concentrations using the formula for total chlorophyll content given by Richardson et al. (2002). In June 2017, we recorded the following traits for the 1043 surviving plants: height of the tallest flowering stem, date of opening of the first flower (flowering onset), number of stems, total number of flowerheads and the number of flowerheads with open flowers. We collected the highest cauline leaf of each plant, placed those leaves between wet paper towels in labelled envelopes in plastic bags and stored them at 5 °C. On the next day, the leaves were weighed to determine their fresh weight, placed in separate paper envelopes, pressed, and dried with silica gel.

We weighed the dried leaves and scanned them at a resolution of 300 x 300 dpi together with a length standard. With the program ImageJ v. 1.51j8 (Schneider et al. 2012) we measured the area of the cauline leaves and calculated specific leaf area (SLA) as the ratio between leaf area and dry mass. Leaf dry-matter content (LDMC) was calculated as the ratio between dry and fresh weight. As a proxy for flowering phenology we calculated the proportion of heads flowering per population as the ratio between the sum of flowerheads with open flowers and the total number of flowerheads. Survival was calculated as the number of plants that survived per population divided by the total number of seedlings planted per population.

Pollination experiments —

In June 2017, selfing-ability was tested on a subset of 223 plants from 27 populations. On each plant, an immature flowerhead was selected. One flower per flowerhead was marked with a permanent marker and the flowerhead protected by a bag of fine nylon mesh (mesh size ca. 0.1 mm) against pollinators. Once the flowers had opened, each flower was either left as a control for autonomous self-pollination or hand-pollinated with pollen from the same flowerhead by using a toothpick to gently transfer pollen to the receptive stigma. In August of the same year, the marked flowers were collected and the presence of developed seeds was determined.