Journal of Ecology ( IF 5.3 ) Pub Date : 2024-07-31 , DOI: 10.1111/1365-2745.14378 Kai‐Hsiu Chen 1 , John R. Pannell 1
|
1 INTRODUCTION
The rate of self-fertilization in self-compatible plants and the level of inbreeding depression experienced by progeny derived from self-fertilization (hereafter ‘selfed progeny’) are major determinants of female reproductive success. Traits that affect the selfing rate are thus expected to be under strong selection. One such trait is the spatial separation between the sexual parts (herkogamy) (Barrett et al., 2003; Webb & Lloyd, 1986), with a negative correlation between the degree of herkogamy and the selfing rate having been found in at least 20 species (reviewed in Opedal, 2018). Another such trait is the relative timing of stigma receptivity and pollen release from anthers in the same flower (dichogamy) (Lloyd & Webb, 1986), for example, reduced temporal separation between the two sex functions was found to be positively correlated with the selfing rate in Campanula americana (Koski et al., 2018) and Gilia achilleifolia (Schoen, 1982). Almost all evidence for the effects of herkogamy and dichogamy is drawn from comparisons among populations that vary in their mean levels of spatial or temporal separation between the sexes (Koski et al., 2018; Opedal, 2018; Schoen, 1982), and there is a surprising dearth of evidence for variation in the selfing rate as a function of variation in these factors among individuals within populations, where ultimately natural selection will be strongest.
In addition to herkogamy and dichogamy, a flower's sex allocation, that is, the absolute and/or relative numbers of pistils and stamens, should also be expected to affect the rate of within-flower selfing (including both autonomous and pollinator-facilitated components of selfing). Although species in most of the monocots and core eudicots tend to have fixed numbers of pistils and stamens, many species from basal clades (e.g. basal angiosperms and Magnoliids) do have variable numbers of floral parts (Ronse de Craene et al., 2003), and this variation could affect the selfing rate in self-compatible species (Allen & Hiscock, 2008; Barrett, 2013; Igic et al., 2006). Moreover, many species present single or few flowers daily or throughout a flowering season, and within-flower selfing should be the dominant source of selfing in these species (e.g. Karron et al., 2004; Méndez & Traveset, 2003; Sun et al., 2018). In these species, we should expect flowers with more stamens per pistil to have a higher selfing rate, just as selfing due to pollen transfer among flowers of the same individuals (geitonogamy) has been shown to depend positively on the number of flowers in the floral display (Harder & Barrett, 1995; Karron et al., 2004; Williams, 2007). The idea that sex allocation should affect the selfing rate is central to a number of models of mating system or sex allocation evolution (Brunet & Charlesworth, 1995; de Jong et al., 1999; Gregorius et al., 1987; Holsinger, 1991; Lloyd, 1992; Sakai, 1995). However, although there has been substantial empirical work on the effects of geitonogamy and sex allocation on the selfing rate at the individual level (Harder & Barrett, 1995; Karron et al., 2004; Williams, 2007), we are not aware of any study that has quantified the relationship between within-flower sex allocation and the selfing rate, despite its likely importance for the evolution of floral morphology and sex allocation.
The within-flower selfing rate should depend not only on the numbers of stamens and pistils in flowers of species that vary for these traits, but also on the timing of their presentation to pollinators and of the presentation of the other flowers in a population. If mating follows a ‘mass-action’ process, in which the mating system depends on the relative availability of self versus outcross mates at a given point in time (Gregorius et al., 1987; Holsinger, 1991), we may expect a decrease in the selfing rate in flowers that produce fewer stamens and that are presented to pollinators when the density of other flowering individuals is high. For example, in the semi-desert plant Incarvillea sinensis and the alpine plant Phyllodoce aleutica, an elevated selfing rate was found in the late and early seasons, respectively, when pollinator visitation rates were low (Kameyama & Kudo, 2009; Yin et al., 2016). However, we are largely ignorant of how temporal variation in plant mating systems relates to quantitative variation in the timing of male and female functions over the course of a flowering season.
Natural selection should act on floral morphology and on their sex allocation and phenology as a function of the selfing rate and the level of inbreeding depression suffered by selfed progeny. Numerous studies have considered the effect of different floral traits on the female component of fitness (reviewed in Ashman & Morgan, 2004; Caruso et al., 2019; Munguía-Rosas et al., 2011), but these studies are typically based on numbers of seeds produced, without accounting for the possibility that many seeds might be self-fertilized and could suffer inbreeding depression (Lloyd, 1992). Nor have any of these studies considered the effects of within-flower selfing as opposed to selfing that includes geitonogamy. Although within-flower selfing and inter-flower selfing are genetically identical, they have different implications for the evolution of floral traits. For instance, a reduction in floral display by presenting fewer flowers in an inflorescence simultaneously might be an effective mechanism to avoid geitonogamous selfing (Harder & Barrett, 1995; Karron et al., 2004; Williams, 2007), while such a mechanism would have very little impact on within-flower selfing. If we are to gain a balanced understanding of how selection shapes the timing and amount of allocation to male and female functions within flowers, we need to consider how these traits affect not only the number of seeds produced but also their quality (as well as the male components of reproductive success). To the best of our knowledge, nearly all the studies of phenotypic selection have not accounted for the potential consequence of inbreeding, and those few that did were conducted in artificial experimental arrays with an unknown level of inbreeding depression (e.g. Briscoe Runquist et al., 2017; Hou et al., 2024). How mating-system variation and the consequence of inbreeding depression alters selection on different traits in wild populations thus remains an open question.
Here, we ask how the amount and timing of sex allocation at the flower level affects within-flower selfing and, consequently, the female component of seasonal reproductive success. We estimated the selfing rate of progeny within flowers and the level of inbreeding depression suffered by selfed progeny in a population of the insect-pollinated and self-compatible species Pulsatilla alpina to determine how the selfing rate and female reproductive success depends on floral sex allocation, phenology and subsidiary floral traits. Because the study population comprised mainly individuals with only one flower, we could assign paternity at the flower level and thus relate contributions to fitness to the traits of individual flowers. Although flowers vary substantially in their pistil and stamen number, we further enhanced this variation by removing stamens from a random subset of flowers in the population, a treatment that also generated purely female flowers that do not otherwise occur in P. alpina. We estimated the female component of seasonal reproductive success by counting seeds produced by individual flowers through both outcrossing and within-flower selfing, accounting for the negative effects on fitness of inbreeding depression. We then quantified phenotypic selection on the traits under two scenarios of inbreeding depression. We specifically addressed the following questions. (1) How does the within-flower selfing rate depend on the number of pistils and stamens within flowers, the timing of flowering and two subsidiary floral traits? (2) What is the level of inbreeding depression? (3) How does the inbreeding depression affect the pattern of selection on the traits as a result of the dependency of within-flower selfing on those traits?
中文翻译：
忽视花内自施和近亲繁殖抑制偏向于多年生高山草本植物花性状选择的估计
1 引言
自交植物的自交受精率和自交后代（以下简称“自交后代”）所经历的近亲繁殖抑制水平是雌性生殖成功的主要决定因素。因此，影响自交率的性状预计会受到强选择。其中一个特征是性部分之间的空间分离 （herkogamy） （Barrett et al.， 2003;Webb & Lloyd，1986），至少在20个物种中发现了雌雄同体的程度与自交率之间的负相关（在Opedal上进行了审查，2018）。另一个这样的特性是在同一朵花（二分花序）（Lloyd & Webb，1986）中，柱头接受性和花粉释放的相对时间，例如，发现两种性功能之间的时间分离减少与Campanula americana（Koski等人，2018）和Gilia achilleifolia（Schoen，1982的自交率呈正相关。).几乎所有关于一夫一妻制和二分制影响的证据都来自两性之间空间或时间分离平均水平不同的人群之间的比较（Koski 等人，2018 年;Opedal，2018 年;Schoen， 1982），并且令人惊讶地缺乏证据表明，自交率的变化是种群内个体之间这些因素变化的函数，最终自然选择将最强。
除了雌雄同体和二分雄同体之外，花的性别分配，即雌蕊和雄蕊的绝对和/或相对数量，也应预期会影响花内自交的速率（包括自交的自主和传粉媒介促进的成分）。尽管大多数单子叶植物和核心双子叶植物中的物种往往具有固定数量的雌蕊和雄蕊，但许多来自基部分支的物种（例如基部被子植物和木兰科植物）确实具有可变数量的花部分（Ronse de Craene等人，2003），这种变化可能会影响自相容物种的自交率（Allen & Hiscock，2008; Barrett，2013 年;Igic et al.， 2006）。此外，许多物种每天或整个开花季节出现单朵或很少的花，花内自交应该是这些物种自交的主要来源（例如 Karron 等人，2004 年;Méndez & Traveset， 2003;Sun等人，2018 年）。在这些物种中，我们应该预期每个雌蕊雄蕊更多的花具有更高的自交率，就像由于花粉在同一个体的花之间转移（geitonogamy）而产生的自交已经被证明对花展示中的花数量产生积极影响（Harder & Barrett， 1995;Karron 等人，2004 年;Williams，2007 年）。性别分配应该影响自交率的想法是许多交配系统或性别分配进化模型的核心（Brunet & Charlesworth， 1995; de Jong et al.， 1999;Gregorius 等人，1987 年;Holsinger， 1991;Lloyd， 1992 年;Sakai， 1995）。 然而，尽管已经有大量的实证工作研究了 geitonogamy 和性别分配对个体水平自交率的影响（Harder & Barrett， 1995;Karron 等人，2004 年;Williams， 2007），我们不知道有任何研究量化了花内性别分配与自交率之间的关系，尽管它可能对花形态和性别分配的进化很重要。
花内自交率不仅应取决于因这些性状而变化的物种的花中的雄蕊和雌蕊的数量，还应取决于它们呈现给传粉者的时间以及种群中其他花的呈现时间。如果交配遵循“质量作用”过程，其中交配系统取决于在给定时间点自身与异交配偶的相对可用性（Gregorius 等人，1987 年;Holsinger， 1991），我们可以预期，当其他开花个体的密度较高时，产生较少雄蕊的花的自交率会降低，并且这些花会呈现给传粉者。例如，在半沙漠植物Incarvillea sinensis和高山植物Phyllodoce aleutica中，分别在后期和早期发现了较高的自交率，此时授粉者访问率较低（Kameyama & Kudo，2009; Yin et al.， 2016）。然而，我们在很大程度上不知道植物交配系统的时间变化与开花季节中雄性和雌性功能时间的数量变化有何关系。
自然选择应作用于花的形态、它们的性别分配和物候，作为自交率和自交后代遭受的近亲繁殖抑制水平的函数。许多研究已经考虑了不同花性状对女性适应性成分的影响（在Ashman和Morgan，2004年;Caruso等人，2019 年;Munguía-Rosas et al.， 2011），但这些研究通常基于产生的种子数量，而没有考虑到许多种子可能经过自我受精并可能遭受近亲繁殖抑制的可能性（Lloyd， 1992）。这些研究也没有考虑过花内自交的影响，而不是包括 geitonogamy 在内的自交。尽管花内自交和花间自交在遗传上是相同的，但它们对花性状的进化具有不同的意义。例如，通过在花序中同时呈现较少的花朵来减少花的展示可能是避免geitonogamous自交的有效机制（Harder & Barrett，1995;Karron 等人，2004 年;Williams， 2007），而这种机制对花内自交的影响很小。如果我们要平衡地理解选择如何影响花中雄性和雌性功能的分配时间和数量，我们需要考虑这些性状如何影响产生的种子数量及其质量（以及繁殖成功的雄性成分）。 据我们所知，几乎所有关于表型选择的研究都没有考虑近亲繁殖的潜在后果，而那些少数研究是在近亲繁殖抑制水平未知的人工实验阵列中进行的（例如 Briscoe Runquist 等人，2017 年;Hou et al.， 2024）。因此，交配系统变异和近亲繁殖抑制的后果如何改变野生种群中不同性状的选择仍然是一个悬而未决的问题。
在这里，我们询问了花水平性别分配的数量和时间如何影响花内自交，从而影响季节性繁殖成功的雌性成分。我们估计了昆虫授粉和自交种白头翁种群中花内后代的自交率和自交后代遭受的近亲繁殖抑制水平，以确定自交率和雌性繁殖成功如何取决于花的性别分配、物候和附属花性状。因为研究人群主要由只有一朵花的个体组成，所以我们可以在花水平上分配亲子关系，从而将对适应度的贡献与单个花的性状联系起来。尽管花的雌蕊和雄蕊数量差异很大，但我们通过从种群中的随机花子集中去除雄蕊来进一步增强这种变化，这种处理也产生了纯雌花，否则不会出现在 P. alpina 中。我们通过计算单个花通过异交和花内自交产生的种子来估计季节性繁殖成功的雌性成分，并考虑了近亲繁殖抑制对适应性的负面影响。然后，我们量化了近亲繁殖抑制两种情景下性状的表型选择。我们专门解决了以下问题。（1） 花内自交率如何取决于花内雌蕊和雄蕊的数量、开花的时间和两个次要的花性状？（2） 近亲繁殖抑郁症的程度如何？（3） 近亲繁殖抑制如何影响由于花内自交对这些性状的依赖性状的选择模式？
京公网安备 11010802027423号