Some sexual consequences of being a plant

Abstract

Plants have characteristic features that affect the expression of sexual function, notably the existence of a haploid organism in the life cycle, and in their development, which is modular, iterative and environmentally reactive. For instance, primary selection (the first filtering of the products of meiosis) is via gametes in diplontic animals, but via gametophyte organisms in plants. Intragametophytic selfing produces double haploid sporophytes which is in effect a form of clonal reproduction mediated by sexual mechanisms. In homosporous plants, the diploid sporophyte is sexless, sex being only expressed in the haploid gametophyte. However, in seed plants, the timing and location of gamete production is determined by the sporophyte, which therefore has a sexual role, and in dioecious plants has genetic sex, while the seed plant gametophyte has lost genetic sex. This evolutionary transition is one that E.J.H. Corner called ‘the transference of sexuality’. The iterative development characteristic of plants can lead to a wide variety of patterns in the distribution of sexual function, and in dioecious plants poor canalization of reproductive development can lead to intrasexual mating and the production of YY supermales or WW superfemales. Finally, plant modes of asexual reproduction (agamospermy/apogamy) are also distinctive by subverting gametophytic processes.

This article is part of the theme issue ‘Sex determination and sex chromosome evolution in land plants’.

1. Introduction

Much of our knowledge of the evolution and diversity of sex comes from zoological or fungal systems, for instance from much studied or model systems such as Homo, Mus [1], Drosophila [2], Caenorhabitis [3] and Neurospora [4]. Nearly all these insights can be applied directly to plants. However, plants have some unique features, and characteristic natural histories, that give a distinctive cast to plant sex. In some cases, this uniqueness provides specific scientific opportunities for the study of sex; in others, they lead to caveats where zoocentric thinking is inappropriate or misleading. It is these issues that will be explored here. A survey of sex over the whole tree of life is beyond my scope here. My plant examples will come mainly from land plants (embryophytes) and related algal groups, but some could also be relevant to other algal groups in the plants (Archaeplastida) as a whole. However, my comments on plant development apply specifically to embryophytes, with the usual emphasis on the important terrestrial clade, angiosperms (flowering plants). Wider comparisons between the main eukaryotic clades would be valuable although is currently data-limited [5].

Land plants (embryophytes) are especially rich in plant-specific biology. Many of these are obvious, such as the transport of plant sperm by animals, as in many bryophytes [6] and flowering plants [7,8]. In flowering plants, the sperm nuclei are packaged within sperm transport organisms called pollen grains. Animal-mediated sperm transport is discussed elsewhere, e.g. [8,9] and will not be further reviewed here. Another well-known feature of many plants is clonal reproduction. While clonal reproduction itself is not unique to plants, the scale of it is (although massive genets are known in fungi too). In extreme cases, it produces single organisms weighing thousands of tonnes (by contrast, the Antarctic blue whale, the largest mammal, weighs in at around 180 tonnes). When asexual reproduction interacts with dioecy (possession of only unisexual individuals), this can result in massive investment in certain single sex individuals. One example is Pando, the nickname given to the 5800 tonne male aspen (Populus tremuloides) that alone makes a sizeable (43.6 ha) forest [10]. Another is the female clone of Japanese knotweed (Fallopia japonica) that is distributed across almost the entirety of the United Kingdom [11]. Such large unisexual individuals may have reduced sexual function, as access to the appropriate opposite sex over the whole clone may be limited. Plant clonal reproduction has been well discussed elsewhere, e.g. [12] and will not be treated further here.

Instead, I have chosen a selection of topics I think have especial evolutionary interest. It should be noted that, in what follows, saying that a trait is characteristic of plants but not of animals does not mean to imply that that trait is necessarily ubiquitous in plants or entirely absent in animals. Exceptions abound in biology but should not be allowed to diminish the importance of general patterns. Also, this article is not intended as an exhaustive review but as a short perspective, so the literature cited is selective rather than comprehensive. I apologize to authors of important papers whose work is omitted here, for reasons of space.

The topics below almost all result from two developmental traits of overarching importance that are characteristic of plants: firstly that they have the presence of a haploid organism in the life cycle and secondly that they often have open, plastic and iterative development through the production of repeated modules, rather than the closed and unitary developmental trajectories of many animals. These two features, as I will argue below, have important consequences for sexual function. Putting the emphasis on the evolution of development suggests the power of an evolutionary developmental genetic approach, both to understanding differences between plants and animals in their sex expression and in understanding the molecular basis of sexual variation between plants.

2. Some features of plants and their consequences

(a) . Haploid selection on organisms not gametes

In typical animals, the life cycle is diplontic, i.e. all in the diploid phase, except for the gametes [13]. In such a system, meiosis gives rise to gametes. The primary natural selection (i.e. immediate post-meiotic selection, the first filter of the new genotypes produced by meiosis) is thus on the gametes. In plants, gametes are not produced by meiosis; instead mitosis gives rise to gametes (life cycle haplontic or diplohaplontic). Meiosis by contrast gives rise to organisms. In plants, the primary selection is therefore on organisms, as these are the first new genotypes produced by meiosis.

We can conveniently represent the difference in life cycle of a typical (diplontic) animal as follows: meiosis > gamete > zygote > mitosis > meiosis. By contrast, a typical plant is either haplontic, as in some algae: meiosis > mitosis > gamete > zygote > meiosis or diplohaplontic as in land plants (embryophytes): meiosis > mitosis > gamete > zygote > mitosis > meiosis. Haploid selection is enormously important as it is selection on the exposed genome without the masking of diploidy and probably plays an important role in the evolutionary trajectory of sex determining systems [14]. Does it matter whether haploid selection is on organisms or gametes? It does, as mature sperm (in both animals and plants) have condensed chromatin and are transcriptionally inactive, and the proteome of Drosophila sperm consists of 952 proteins only [15]. By contrast, the plant gametophyte, such as the pollen grain, is highly transcriptionally active [16], and much of the haploid genome is thus exposed to selection. However, it should be noted that animal sperm carries a very large repertoire of untranslated RNAs [17], for instance, 12 989 separate autosomal transcripts in haploid spermatids of mouse [18]. This repertoire is transcribed during spermatogenesis and, at least in part, passed to the zygote. Its role is debated [19], and whether it could have a phenotype or be selected on is unclear. There is some evidence that plant sperm too may have an RNA repertoire [20]. What is certain is that different genes contribute to the phenotypes of gametes and organisms. Whereas gametes are single cells with very precise and limited function, organisms (such as prothalli or pollen grains) are multicellular with a range of different functions, often including for instance photosynthesis, nutrient uptake and gametogenesis. As such, they have a more complex fitness landscape.

A peculiar form of ‘sperm competition' may come into play via a type of polyembryony [21]. Where female gametophytes with many archegonia have access to sperm from multiple individuals, as in ferns receiving sperm from numerous neighbours or in gymnosperms receiving numerous pollen grains, multiple embryos may form with identical maternal haploid genotypes, but different sires. As only one sporophyte usually survives, the sperm genotypes compete (as diploids) in an identical maternal background [22]. In bryophytes, the possibility of female choice has recently been raised and that the male genotype may influence the amount of nutrients a female allocates to a particular sporophyte [6]. In flowering plants however, which have a single archegonium, polyembryony is a quite different: it is often a sign of adventitious embryony, a form of agamospermy (seed production without sex), and the embryos are genetically identical.

(b) . Double haploids and asexual sex in monoicous bryophytes

It follows from the above that in animals gametes are generally all different, as they are products of meiosis. In plants, this is not necessarily so: as organisms are the products of meiosis, all gametes produced by the same haploid organism (such as the free-living gametophyte of a bryophyte) will be identical. Bryophytes, for example, may be ‘monoicous’ which means they produce eggs and sperm on the same haploid individual. In such cases, mitoses after meiosis allow that a single product of meiosis (haploid individual) can produce both sperm and eggs, all of which will be genetically identical. If such an organism selfs, it will produce double haploids. This has been described as ‘the strongest possible form of inbreeding' [23, p. 201], but arguably it is not really inbreeding at all but is instead effectively akin to asexual reproduction (despite sperm and egg being involved). Intragametophytic selfing appears to be quite common in monoicous bryophytes, as sperm tends not to be dispersed great distances. It also occurs in fungi where it is referred to as ‘haploid selfing’ [24].

We should thus not assume that sexual reproduction is always capable of generating recombinant organisms. Nevertheless, double haploids may differ from true asexuals by the resetting of epigenetic marks during gametogenesis. Another thing to bear in mind, as a peculiarity of this system, is that consequently haploid selection (of a sort) occurs in the diploid sporophyte also, when it is double haploid. Furthermore, meiotic drive (often included under the umbrella ‘haploid selection') is absent when these double haploid sporophytes go into meioisis, as meiotic drive depends on heterozygotes to provide competing alleles.

(c) . The sexless diploid in homosporous plants

Most animals only consist of the diploid organism: as they have no alternation of generations and there is no haploid organism to consider. Sex expression in the diploid animal is often genetically determined (for instance, XY or ZW sex determination systems) and is correlated with production of the gametes appropriate to the sex of the individual. As we have seen above, in plants, it is the haploid organism that produces gametes, and therefore gametic sex is determined at the haploid stage. Thus, in plants with unisexual gametophytes (termed ‘dioicous’), sex is determined by a single copy of a polymorphic chromosome which exists in U and V forms, as in some bryophytes [25,26]. A haploid with a single U chromosome is female and will produce sex-appropriate gametes (i.e. eggs in archegonia), whereas haploid organisms with a V chromosome will be male and produce sperm in antheridia. This is the distinctive haploid UV sex determination system unique to plants and outside plants to certain brown algae [27]. The importance of haploid sex determination is that it creates a significant opportunity for those who study the developmental genetics of sex characteristics: the UV system allows sex determination to be studied in a haploid background, which is not generally feasible in animal systems.

(d) . Transference of sexuality, epigender and the genetically sexless haploid in heterosporous plants

In animals, diploid organisms produce gametes appropriate to their sex chromosome complement. By contrast, diploid plants do not produce gametes. That does not mean they are necessarily sexless, as in seed plants the diploid sporophyte controls when and where the gametes are produced, and they are the organisms in which the sex chromosomes actively determine sex. Rather than produce gametes themselves they enforce epigenetic sex (epigender) on a genetically sexless haploid organism (sexless in the sense of the absence of fixed genetic differences), which is thereby manipulated to produce gametes appropriate to the sex (i.e. sex chromosome complement) of the parental diploid. Genetic sex therefore results from carrying a particular gene or allele, whereas epigender results from developmental context and the resulting epigenetic programming.

When an organism determines features outside that of the organism itself, the ‘extended phenotype' concept can be employed [28]. This ‘extended phenotype' of the diploid plant (determining the sex of derivative haploid organisms) means that direct production of specific gametes cannot be used as a marker of sex. Rather, it is the sex chromosome complement, the ability to produce mega versus microspores and the ‘gamete production at a distance' (i.e. gamete determination by, but not production by, the sporophyte) that are markers of sex. This is an important difference between animals and dioecious plants and means that zoocentric thinking in plant biology can be highly misleading.

In mosses, sex determination resides in the haploid gametophyte, whereas in seed plants, it resides in the sporophytes. Somewhere between the two, sex was transferred in evolution from the haploid to the diploid. The tropical botanist E.J.H. Corner put this well when he wrote in ‘The Life of Plants’ [29, p. 171]: ‘By transferring maleness and femaleness, that is sexuality, from the gametophyte to the sporophyte, the gametophyte part of the life cycle is eliminated as a free-living state of the plant.' This process of ‘transference of sexuality' from gametophyte to sporophyte starts, in a small way, with heterospory (homosporous sporophytes, e.g. filicine ferns, being quite sexless).

How is the appropriate gamete production set by a male or female sporophyte? Or for that matter, how is positionally appropriate gamete production set by an individual with both male and female sexual function (termed a cosexual individual)? The genotype of the haploid gametophyte is not involved, as in, for instance, dioecious XY systems, gametophytes carrying only an X chromosome produce sperm as readily as gametophytes carrying a Y. Instead, sporophyte development sets up developmental contexts (carpel or anther) to which gametophyte development, and ultimately gametogenesis, captively respond.

An important part of this is the setting up, by the sporophyte, of sex-specific epigenotypes in the gametophyte. Sex-specific epigenotypes are sometimes called ‘epigender' [30] and involve the methylation, or lack of it, of certain genes. Epigender is normally thought of in terms of the egg, sperm and central cell, as all these have sex-specific epigenotypes that are essential for normal embryo and endosperm development [30]. However, all the gametophyte nuclei have epigender, as epigenetic repatterning is set up as early as the megaspore and microspore mother cells. For instance, the chromomethylases dmt102 and dmt103 in maize are expressed in the megaspore mother cells and through meiosis into the gametophyte [31]. Therefore epigender, as set by the developmental context imposed by the sporophyte, applies to the whole seed plant gametophyte. Proper sexual functioning is as much controlled by the sporophyte as it is a product of the gametophyte. In Corner's words, sexuality has been ‘transferred' from the gametophyte to the sporophyte, starting with heterospory. In this regard, it would be of great interest to examine differential methylation (for example) during mega- and microsporogenesis in heterosporous lycophytes and ferns.

(e) . Sexual consequences of iterative development

In many (although by no means all) animals, development is unitary. In unitary development, a fertilized egg divides, and cell divisions continue following a set and highly conserved pattern until a single adult unit is reached, after which there is little further development. In modular plants (including embryophytes and modular algae such as stoneworts), development is very different. In flowering plants, for example, normal development proceeds by iterating increasing numbers of four basic modules: roots, stems, vegetative leaves and flowers (collectives of distinctive sterile and fertile leaves). Some or all of these modules are produced throughout the active life of the plant, and their rate of production, position and turnover is controlled by continuous feedback from the environment. Plants can be seen as populations of modules and their growth and form described by birth-death demographic equations [32]. This has considerable consequences for sexual reproduction. The most obvious is in the plasticity of reproductive output; the same or similar genotype in a favourable environment may produce orders of magnitude more flowers and seed [33]. This reproductive plasticity is a potentially a feature of any clonal growth and thus applies widely in plants.

More significantly perhaps is the fact that reproductive modules are produced iteratively with their production in time and space controlled by an interaction between genotype and ‘environment' (considered broadly as the external environment and the physiological status of the plant). This allows for a large number of reproductive strategies involving the specific placing of sexual function in time and space. These strategies are well explored in flowering plants but could also profitably be further explored in other groups. Flowers may be bisexual and can be produced in large dense inflorescences which are highly attractive to insects but risk geitonogamy (selfing by transfer of pollen between different flowers of the same individual), or they may be spread out on the plant. This leads, inter alia, to the flowering strategies Gentry noted in neotropical trees, including cornucopia and steady state [34]. Cornucopia is mass flowering with huge pollinator attractiveness but little incentive for pollinators to move from tree to tree. Steady state is the constant production of a few flowers at a time, forming a reliable reward for specialized traplining pollinators and optimizing interplant gene flow. Alternatively, unisexual flowers may be produced which may then be separated on the plant, in time or space, as a form of dichogamy (temporal separation of male and female function) or herkogamy (spatial separation of male and female function). Monoecy (the presence of male and female unisexual flowers on a cosexual individual) is an interesting and widespread sexual state. It provides for a large amount of flexibility in sexual strategy. Between species, the location and timing of male and female flowers can vary widely in patterns that could be highly adaptive (although conceivably they might result merely from developmental constraint). Monoecy deserves greater comparative study, and its evolutionary and developmental origin (from the assumed ancestral hermaphroditism of the angiosperm crown group) is an outstanding biological problem. Furthermore, monoecy has been argued on comparative evidence to be the most common intermediate step in the origin of dioecy [35], whereas a transition through gynodioecy is rarer. A reasonable estimate for the frequency of monoecy (including gyno- and andromonoecy) in the world flora is 7% (estimated as ‘slightly higher' than the frequency of dioecy, i.e. 5–6% [36]), contrasting with a frequency of less than 1% for gynodioecy [37]. When it is further observed that there is considerable generic overlap between monoecy and dioecy, but very little generic overlap between gynodioecy and dioecy [36,38], it is evident that gynodioecy is likely to be only a minor pathway to dioecy, while monoecy is the major one [38,39]. Transitions between the main sexual systems in flowering plants, i.e. monocliny (bisexual flowers) to monoecy (unisexual flowers on the same individual) and dioecy (unisexual flowers on different individuals) are a rich field for theoretical investigation [40–42], as also should be the transitions between monoicous and dioicous systems, as recently discussed elsewhere [43,44].

In animals, such a complex descriptive language for the distribution of sexual function is lacking. That is not to say that there is not vast reproductive variability in animals. The terms monoecy and dioecy, as borrowings from botany, are used to describe this (and occasionally the various polygamous terms too, but these are rare). In the sponges, for example, both ‘monoecious’ and ‘dioecious’ species occur. Some coelenterates approach plants in forming gonads in response to weather and in releasing gametes according to environmental cues, as in some corals. Corals may be dioecious or monoecious and may form same sex or mixed sex colonies. There may be imbalances in the numbers of sexual organs: the New Zealand flatworm Anonymus multivirilis has over 200 penises (male complexes) arranged around its periphery and a single female copulatory organ [45]. In molluscs, there may be strict dioecy, as in cephalopods. Other molluscs are monoecious, and some species change sex (oysters are a well-known example). Monoecious molluscs vary in the independence of their sexual organs, which are free in some but combined into an ovotestis in others. The ovotestis, a single gonad producing sperm and eggs, is possibly the nearest analogue to the angiosperm hermaphrodite flower which also brings the sexes close together. Arthropods are largely dioecious although hermaphrodites do occur. The same is true of vertebrates, where hermaphrodites are rare but found for instance in some fishes. Yet despite this variation, no general terminological expansion has been considered as needed in animals and dioecy versus monoecy (hermaphroditism)—terms borrowed loosely from plants—generally encompass all.

(f) . Pleogamy and supersexes

Male × male mating in XY males is rare in animals, but because of pleogamy, there are several examples known in plants. Pleogamy refers to the ‘rare occurrence of unexpected floral forms in plants of a given mating system' [42] and is sometimes referred to as ‘sexual inconstancy'. Thus, the occasional production of flowers with female organs on an XY male will alliow XY selfing and the production of YY individuals (or vice versa for ZW systems). An example of economic importance is found in Asparagus officinalis. Males are agriculturally preferred [46], as they do not put resources into making fruit, and there are no seedlings to weed out. For this reason, seed companies produce pure XY seed by using YY ‘supermales’, as YY × XX mating will produce only XY progeny. In this case, rare hermaphrodite flowers on XY males (female pleogamy) allow YY supermales to be generated for breeding purposes. Alternatively, male pleogamy in dioecious ZW systems can readily allow formation of WW superfemales. In the asparagus case, the supermales were deliberately bred. However, the frequent occurrence of pleogamy in chromosomally dioecious species potentially allows supersexes to form in nature. When supersexes occur in nature, they will probably skew the sex ratio. They will also potentially have an effect on the evolutionary trajectory of the Y chromosome by allowing recombination between Y chromosomes [47]. Recombination is an important means of deleting inserted genetic material such as transposons, by unequal crossing over. It will also introduce Y chromosome variation, which is raw material on which Y-selection may act and reduce the use of the Y chromosome in ancestry tracking.

Before genomics, it would have been enormously difficult to survey natural populations for supersexes. Cytological surveys could be carried out, but this would be laborious and impossible in chromosomally homomorphic systems. Alternatively, progeny trials could be carried out, but this would be even more laborious. However, now the genomic basis of sex is worked out in several plant species, it is possible, in principle at least, to design molecular screens for supersexes. In cases of populations of dioecious plants with skewed sex ratios, it might be useful to screen for supersexes, to rule it out as a potential cause (and similarly for agamospermy, below).

(g) . Agamospermy and the superfluous male

In animals, parthenogenesis and gynogenesis occur widely [48,49]. Rotifers can be, famously, clonally reproducing with no sex for extended periods of time. But, parthenogenesis is usually found as a mechanism to produce many identical females during a phase in the population cycle of an otherwise sexual population, as in Homoptera or Hymenoptera. However, plants, notably flowering plants, are unusual for subverting the normal sexual process so frequently and in so many lineages. It does this by altering the normal process of gametophytic development, in various ways, and hence is plant specific. The phenomenon is commonly known as agamospermy in seed plants [50] and apogamy in ferns [51]. Agamospermy can be seen as a consequence of the sporophyte epigenetically controlling gametophytic processes, discussed above, as epigenetic (mis-)regulation is important [52,53]. Hybridization and polyploidy (both of which occur in plants) might be a source of genomic shock and epigenetic instability and thus of potential relevance here [54].

Agamospermy has a minimal effect on other parts of the life cycle, and so seed reproduction and germination appear quite normal. As such, it can be difficult to detect although it is usually obvious from microscopical study of the ovule, from genetic studies or from pollinator exclusion studies. The latter is somewhat problematic as in some variants (pseudogamy), pollination may be required to start development of the triploid endosperm.

Agamospermy is sometimes found in dioecious species, where it leads to female-skewed sex ratios as the progeny are identical copies of the mother. Indeed, agamospermy was discovered in a dioecious species (at Kew in 1841), in consequence of the single female tree in cultivation setting seed despite having no access to a male [55]. This was Alchornea ilicifolia in the family Euphorbiaceae. Alchornea ilicifolia is not wholly agamospermous (if it was, males would be extinct), but the sex ratio in its native habitat in Australia of >10 : 1 F : M [56] indicates that agamospermous reproduction is commoner than sexual. A more striking example is the tropical fruit crop, the mangosteen (Garcinia mangostana). This is a wholly agamospermous dioecious tree in which males are not only superfluous but now apparently extinct.

In species that are facultatively agamospermous at a low level, agamospermy may be especially problematic to reveal. More examples in dioecious plants are probably still to be discovered or confirmed. In willows (Salix sp.), facultative agamospermy has been repeatedly suggested [57,58] but never rigorously demonstrated (although results are suggestive). In investigating cases of female biased sex ratio in dioecious species, agamospermy is one possible cause that needs to be ruled out.

3. Concluding remarks

It is obvious that plants (at least most multicellular ones) are different from most other organisms by their combination of haploid life cycle and iterative development. For this reason, zoocentric thinking can be misleading. No more so than in the commonly made, but erroneous, assertion that the sporophytes of dioecious species cannot be male or female because they do not produce gametes. On the contrary, they are male and female because of the gamete production they control, and because of the sex chromosomes they carry that make them so. While it is important to gain insights from zoological studies, it is sometimes important to stand aside from the firehose of zoological studies and consider planta qua planta. Plants offer many special natural histories of sex. Many are connected with characteristic aspects of plant development, and it is highly desirable to understand more about the molecular control of sexual development specific to plants, and its evolution, not just in flowering plants but widely in land plants and their algal relatives.

Data accessibility

This article has no additional data.

Authors' contributions

Q.C.: conceptualization, investigation, methodology, writing—original draft, writing—review and editing.

Competing interests

I declare I have no competing interests.

Funding

Research in my laboratory is funded by the Discovery Grants program of the Natural Sciences and Engineering Research Council (NSERC), Canada, which is gratefully acknowledged.