A pan-European citizen science study shows population size, climate and land use are related to biased morph ratios in the heterostylous plant Primula veris

Abstract

1 INTRODUCTION

The last century has witnessed unprecedented habitat loss and fragmentation as a result of global land use change (Haddad et al., 2015). Along with climate change, this trend adversely affects several aspects of biodiversity, causing the loss of genetic diversity (Des Roches et al., 2021; Laikre et al., 2020; Schlaepfer et al., 2018), species richness (Tilman et al., 2017), and related ecosystem services (Cardinale et al., 2012). However, not all species respond in the same way to these factors, with some species being more vulnerable to the mentioned threats than the others. In plants, the response of certain species may depend on their life history, functional traits, phenotypic plasticity, and biogeographic origin (De Kort et al., 2021; Hamrick & Godt, 1996). In addition, the effects of fragmentation and climate change on plant mutualistic partners, such as pollinators (Bennett et al., 2020; Rodger et al., 2021), seed dispersers (Donoso et al., 2022) or mycorrhizal fungi (Kiesewetter et al., 2023; Outhwaite et al., 2022; Senapathi et al., 2017), may affect the relative vulnerability of plant species to the factors of global change. Outcrossing, animal-pollinated plants may be more susceptible to climate change and habitat fragmentation than clonally reproducing, selfing, or anemophilous plants due to potential negative impacts of habitat loss and climate change on pollinators (Aguilar et al., 2008; Bennett et al., 2020; Rodger et al., 2021). Furthermore, reduced pollinator abundance and diversity may ultimately cause shifts in plant–pollinator networks (Zoller et al., 2023), potentially triggering selection of phenotypes with reduced herkogamy or self-incompatibility (Bodbyl Roels & Kelly, 2011; Cheptou, 2021; Jacquemyn et al., 2012; Opedal, 2019). However, our understanding of how reproductive plant traits respond to climate change and land use shifts in contemporary landscapes is still limited (Pontarp et al., 2023).

Insufficient pollination poses a particular threat to plant species with floral traits preventing self-pollination, such as heterostyly. Heterostyly is a genetically determined floral polymorphism expressed in the reciprocal positioning of female and male reproductive organs (Barrett, 2019). It has evolved independently across at least 28 plant families (Barrett, 2019). Populations of heterostylous plants comprise two (distylous species) or three (tristylous species) morphs with reciprocal lengths of style and anthers. Differences between morphs may also be expressed in the size and morphology of stigmatic papillae and pollen grains (Costa, Castro, et al., 2017). The study of floral polymorphism has a long history with Charles Darwin being the first to suggest that such reciprocal floral polymorphism helps to ensure outcrossing among plant individuals (Barrett & Shore, 2008; Darwin, 1862; Simón-Porcar et al., 2022). Recent research has shown that heterostylous plant species are often characterised by a genetically determined incompatibility system, promoting disassortative mating (Costa, Ferrero, et al., 2017; Huu et al., 2016, 2022; Keller et al., 2014). Self-incompatibility and associated traits in heterostylous plants, including the position of anthers, the height of style, pollen size differences between morphs, and biochemical self-incompatibility are controlled by a set of tightly linked genes located in the S locus, which is hemizygous in one morph type (S-morph) but absent in the other morph (L-morph) in distylous plant species (Huu et al., 2016).

Different morphs usually have balanced morph frequencies in populations at equilibrium (isoplethy). However, drastic landscape changes and a decline in plant population size can cause stochastic deviations from isoplethy (Endels et al., 2002; Kéry et al., 2003), or even lead to a complete loss of one (in distylous species) or two morphs (tristylous plants; Ferrero et al., 2020; Heuch, 1980). Unbalanced morph ratios can reduce the number of suitable mating partners, reproductive output and genetic diversity in populations of heterostylous plants (Kaldra et al., 2023; Kéry et al., 2003; Meeus et al., 2012). Due to disassortative mating and a strong dependence on pollinators, heterostylous plants are thus particularly vulnerable to the negative consequences of habitat loss (Brys et al., 2004), which may eventually lead to increased self-compatibility and a breakdown of heterostyly under conditions of severe pollen scarcity (Wang et al., 2020). Furthermore, shifts in climatic conditions may have significant effects on plant–pollinator interactions. Changes in temperature and precipitation, but also an increasing frequency of extreme weather events may cause phenological decoupling of plant–pollinator networks, altered pollinator foraging patterns, increased pollen degradation and stronger nectar dilution (Lawson & Rands, 2019; Settele et al., 2016). In heterostylous species, higher humidity caused by shifts in precipitation patterns can lead to distinct viability of pollen of different morphs (Aronne et al., 2020). The combined effects of land use and climate change may thus impose an increased threat to plants with complex mating systems, such as heterostyly (Aronne et al., 2020; Stefanaki et al., 2015; Thomann et al., 2013).

Primula veris L. is a distylous commonly used model species in heterostyly research (e.g. Kéry et al., 2003; Nowak et al., 2015; Potente et al., 2022). This herbaceous plant occurs in Eurasian rural and mountainous (particularly in Southern Europe) landscapes with semi-natural grasslands and semi-open forests among their preferred habitats. However, 90% of European semi-natural grassland areas are lost (Dengler et al., 2020) with adverse effects on the occurrence and genetic diversity of many grassland species (Kiviniemi, 2008; Lienert, 2004; Lindborg et al., 2005), including P. veris (Brys & Jacquemyn, 2009; Kery et al., 2000; Van Rossum et al., 2004). In populations of P. veris at equilibrium, an equal ratio of the two morphs can be expected due to negative frequency-dependent balancing selection (Heuch, 1979). Deviations from isoplethy have been associated with declining P. veris population sizes (Aavik et al., 2020; Kaldra et al., 2023; Kéry et al., 2003). Such stochastic deviations of morphs from equal frequencies are not expected to affect the general proportion of morphs across populations because the relative loss of one or the other morph is completely random. A recent study exploring the morph ratio of P. veris in more than a thousand populations at its northern distribution limits (Estonia) has demonstrated an overall excess of short-styled S-morphs over long-styled L-morphs as well as an excess of populations where S-morphs dominated (Aavik et al., 2020). This finding suggests that other deterministic processes shape the deviation of morphs, leading to non-random prevalence of one specific morph type over the other. Previous research has found considerable variation of between- and within-morph compatibility and self-fertilisation in the populations of other Primula species near their distribution range margin (Shao et al., 2019; Van Daele et al., 2024; Zhang et al., 2021) or along elevation gradients (Yuan et al., 2017). Such disruptions of mating systems can occur when colonisation events lead to small and isolated populations, where self-compatibility is advantageous. Such shifts may further be facilitated by the scarcity of pollinators, favouring the adoption of selfing and homostyly as strategies to ensure successful reproduction (Yuan et al., 2017). Recent evidence shows that these shifts can occur as a result of mutations or rearrangements in genes associated with S locus (Mora-Carrera et al., 2023). Besides the natural causes of pollinator shortage, human-induced land use changes decrease pollinator abundance and may cause shifts towards decreased pollinator dependence (Brys & Jacquemyn, 2012; Cheptou, 2021). Despite the demonstrated effect of pollen limitation on plant reproductive strategies and its adverse fitness consequences (Rodger et al., 2021), the shifts in plant mating system have rarely been examined in landscapes strongly altered by human activities (Cheptou, 2021).

To fill the gap in knowledge about the large-scale patterns of morph ratios in heterostylous plants, we explored within-population morph frequencies of P. veris across most of its European distribution range. We launched a citizen science campaign in 28 European countries, enabling a broad-scale data collection of morph ratios in European P. veris populations. In particular, we aimed to test the following hypotheses:

1. Small population size could lead to significant deviations of morphs from balanced frequencies (Kéry et al., 2003). We hypothesise that these deviations will be stochastic, that is not in favour of one or the other morph.
2. In landscapes with limited grassland and woodland habitats, that is homogeneous landscapes with a higher proportion of arable land and built-up areas, it is expected that there is a stronger deviation from isoplethy as a result of habitat fragmentation. These effects may be both stochastic (independent of one specific morph) or deterministic, that is in favour of one particular morph (Aavik et al., 2020), mediated through the effects of landscape composition on pollinators as shown in research on other distylous species (Shao et al., 2019; Yuan et al., 2017; Zhang et al., 2021). Alternatively, partial intra-morph compatibility in L-morphs in P. veris, as shown by limited evidence (Brys & Jacquemyn, 2015; Wedderburn & Richards, 1990), may lead to the dominance of L-morphs over S-morphs in landscapes with low habitat availability. In such conditions, a relatively higher number of suitable mating partners for L-morphs would give an advantage for this morph type compared to S-morphs with almost complete intra-morph compatibility. Another possibility, demonstrated, however, only in P. elatior (Van Daele et al., 2024), is that S-morphs may become overrepresented in fragmented populations due to combined effects of evolving lower herkogamy, reduced physiological self-incompatibility and more autonomous self-pollination by pollen falling onto the low stigma from the high anthers.
3. Relative morph frequencies, either stochastic or deterministic, can be affected by climatic factors, such as precipitation and temperature, through the possible impact of these factors on plant–pollinator interactions as pollinators are less active in high temperatures and rainy conditions (Ganuza et al., 2022; Jiao et al., 2023; Kammerer et al., 2021; Lawson & Rands, 2019; Teixido et al., 2022). In addition, plant–pollinator interactions can be disrupted by increasing speed of wind due to its negative effect on pollinator activities (Hennessy et al., 2021). We hypothesise that populations in regions with higher wind speed will show greater deviations from equal morph frequencies (isoplethy), as wind-mediated pollen dispersal or disrupted pollination service due to limited pollinators may differentially affect the reproductive success of the two morphs.
4. Finally, we expect that, due to increased human pressure on European rural landscapes and resulting decline in pollinator richness and abundance, we will find homostylous individuals of P. veris indicating disruption or breakdown of heterostyly.