Jan 20, 2017

Determination of Sex Expression in Cycads

Cycads, be it extinct or extant are unique among gymnosperms and plants in general due to their absolute dioecious nature. Apart from the living cycads, no bisexual cycad cones or individuals have ever been found in the fossil record. This dioecy can be correlated with occurrence of sex chromosome having unequal length but occasional sex change as encountered in cycads compounds the problem of explaining the sexuality. Whatever factor behind the sex expression, distinguishing the maleness or femaleness at an early vegetative stage is virtually impossible for cycads. This poses serious problems in raising a population prior undertaking a proper conservation strategy either through in situ or ex situ means. The present review, hence aims to address the classical and state-of-art molecular methods of sex determination in cycads at pre sporangial stage. Since genome information of cycads is scanty, most of the PCR (Polymerase Chain Reaction) based approaches adopt anonymous markers, which later on are usually converted to more informative SCAR (Sequence Characterized Amplified Region) markers. Even more reliable marker for sex determination in cycads at pre sporangial stage will plausibly emanate from a technique like Methylation—Sensitive Amplification Polymorphism (MSAP), which considers the epigenetic mechanism, especially DNA methylation since perturbation of environmental cue leads to occasional sex change.   

Gymnosperms include the earliest living lineages with innovations of a greatly reduced male gametophyte (pollen), pollination as well as seeds, but with extremely large genome. Recent molecular phylogenetic studies  indicate that the five major lineages of extant gymnosperms (cycads, Ginkgo,  Gnetales, Pinaceae, and all other conifers) form a monophyletic group that is  sister to angiosperms, although an alternate placement of Gnetales as sister  to the angiosperms cannot be unambiguously rejected. Cycads represent the likely sister lineage to all other extant gymnosperms and are consistently included in comparative developmental and molecular systematic studies. As a basal lineage, cycads provide exemplars to help ascertain the generalized gymnosperm reproductive features from which flowering plant morphology and genetic controls were likely modified (http://www.  greenbac.org/tree.html: The green plant BAC library project).

Charles Darwin hypothesized that ancestral lineages were hermaphroditic (each individual produces both functional female and male gametes) and that dioecy evolved as a derived condition (Darwin 1873). In plants and animals, evolution from hermaphroditism to strict dioecy almost certainly occurred via an intermediate stage that involved individuals who were both hermaphroditic and either functional females or males (Darwin 1873; Charlesworth and Guttman 1999; Gorelick 2003). These intermediate stages are referred to as gynodioecy (functional females) or androdioecy (functional males). However, in spite of being an ancient lineage (existing for 275 to 300 million years in near modern form), hermaphrodite sexual morph was never found in cycads and dioecy appears to be quite a primitive trait among cycads. No bisexual cycad cones or individuals have ever been found in the fossil record (Gorelick and Osborne 2007) though occurrence of bisporangiate cones has been reported in conifers (Flores-Renteria et al. 2011). The only instances of anything other than strict dioecy that have ever been seen in cycads are those examples of sex change (Osborne and Gorelick 2007).

The living cycads can be divided into three families; Cycadaceae, Stangeriaceae and Zamiaceae which consisted of 11 genera, 297 species and sub-species. Since all cycads are dioecious, they appear to possess little if any sexual dimorphism as far as vegetative structures is concerned. Therefore, one can hardly distinguish the sexes of a species by visual inspection unless their reproductive organs are present. This poses serious problems in raising a population prior undertaking a proper conservation strategy either through in situ or ex situ means since many of the cycads, including Zamia, have already been placed in the endangered red list category (Roy et al. 2012). Attempt towards artificial pollination leading to recovery of large number of viable seeds is, hence, the only way to date, for scaling up the already diminishing population of these threatened plants, for which substantial number of donor plants of both sexes is a prerequisite.

In the backdrop of this, the present review is aimed to address the classical and state-of-art methods of sex determination in cycads, which are a popular area of research among experts in various fields, including plant molecular biology, agriculture, horticulture, ecology and environmental protection.

Ecological Approach: Sex and Population Differences
Four natural populations of Cycas micronesica growing under differing ecological conditions were surveyed over a 4 year period to assess the response of juveniles before and after the introduction of the cycad aulacaspis scale insect (Aulacaspis yasumatsui Takagi) and to test the hypothesis that monopodial ovulate plants are taller than pollen-producing plants with equivalent diameter (D). Height (H), D, and leaf and stem tip numbers were recorded for 297 ovulate and 186 pollen-producing (“male” and “female”, respectively) plants and a total of 493 juveniles (n=976). Among the 483 adults, mean female plant H and D did not differ significantly from those of male plants. However, population-specific differences in mean plant size were observed; i.e., female plants achieve greater H and have significantly smaller leaf and stem tip numbers than do their male counterparts with equivalent increases in D. Population differences, however, were not statistically significant for juvenile plants (Niklas and Marler 2008).

Cytological Techniques towards Identification of Sex Chromosome
In many animals (but only in a handful of plants), sex chromosome of different lengths can be identified. For example, in humans, the Y chromosome is much shorter than the X chromosome. Evolution of sex chromosomes is usually explained by a population genetic model known as Muller’s ratchet (Griffin et al. 2002; Gorelick 2003). One of the prerequisites for Muller’s ratchet for sexual organisms is that the haploid stage of the life cycle should be largely secluded from selection. This is the case in most animals, which generally have small and short-lived gametes. Plants, on the other hand, have large multi-celled and long-lived haploid stages (gametophytes), with cycads being extreme in this regard. Cycad gametophytes are enormous; occupying much of the volume of what becomes a seed following fertilization. Female cycad gametophytes are also long-lived; sometimes more than a year passes between pollination and fertilization (Norstog and Nicholls 1997). Male gametophytes are also large and complex compared with other plants. Although there are no data concerning whether cycad haploid stages express most of the genes expressed by their diploid stages, the structural complexity, size and age of cycad gametophytes indicate that they should be largely immune from the Muller’s ratchet. Therefore, the chance of cycads (and most other plants) developing unequal length of sex chromosome is remote (Gorelick 2005). Another argument for cycads not having unequal length sex chromosome is that occasional sex change does occur in cycads and it could only be possible if large chromosomal rearrangements had taken place, which is virtually impossible (Gorelick and Osborne 2007).

However, contrary to the aforesaid rationale, the study of Sangduen et al. (2007), encompassing three species of Cycas and Zamia of Nong Nooch Tropical Botanical Garden, Thailand, revealed that all the three species of both Cycas and Zamia have an equal chromosome number in each species, namely 2n = 2x = 22 in Cycas and 2n = 2x = 16 in Zamia. The karyotype formulae of Cycas varied into 3 groups: 12 M + 8 SM + 2 A, 10 M + 8 SM + 2 A and 12 M + 8 SM + 2 T and of Zamia were 12 M + 4 SM. Nevertheless, of all the three Cycas species studied, the karyotype of female and male plants could be distinguished. The Zamia evidence, however, was further complicated as all three species had the same chromosome number and karyotype pattern as well. Only Z. pumila could be differentiated between male and female plants (Sangduen et al. 2007). Hence, concepts and/or reports on identification of sex at pre-sporangial stage through cytology or karyomorphological study seems to be plausibly conflicting.

The recent elegant review on sex chromosomes in land plants (Ming et al. 2011); however, is a proponent of unequal length of sex chromosomes in cycads. The hypothesis of six stages of plant sex chromosome evolution as proposed by them have placed the candidature of Cycas revoluta in the fifth stage where severe degeneration of the Y chromosome has caused the loss of function for most genes, and loss of nonfunctioning Y chromosome sequences resulting in a shrinking of the Y chromosome. They have hypothesized that some sex chromosome systems might not have undergone this phase of shrinking but instead kept expanding and degenerating until a complete loss of the Y chromosome had taken place. In either case, a small portion of the Y chromosome has continued to meiotically pair with the X chromosome allowing proper disjunction. Since there were no known angiosperm sex chromosomes at this stage, they (Ming et al. 2011) are of opinion that the gymnosperm species Cycas revoluta having heteromorphic sex chromosomes with a reduced Y chromosome has played the intermediary stage (Fig. 1).

More refined and state-of-art technique like FISH (Fluorescence in situ Hybridization) as attempted by Tagashira and Kondo (2001), who have studied the chromosome phylogeny of Zamia and Ceratozamia by rDNA analysis through FISH may throw some light in this direction.  Theoretically, at least, it can be postulated that there must have an epigenetic control behind this conflict regarding the existence of sex chromosome having unequal length. Cycads appear to have retained an ancestral form of dioecy, with the sex of an individual being determined by cytosine methylation down regulating genes responsible for production of gametes or sex chromosomes. Cycads have thereby retained the phenotypic plasticity to change sex via removing methylation in the face of large environmental perturbations. It is not obvious how many other plants have retained this plasticity in sex determination. It does not appear that any genetic assimilation of the epigenetic mechanism of sex determination has occurred in cycads. However, such canalization of dioecy may have been unnecessary because cycads cannot revert to a hermaphrodite condition via allopolyploidy. Finally, it appears that cycads are immune from Muller’s ratchet because they have haploid stages that express most of the genes expressed in their diploid stages. Micro array studies could be used to test this inference of immunity from Muller’s ratchet on the large size, complexity and longevity of cycad gametophytes (Gorelick and Osborne 2007).

Molecular Marker Techniques for Sex Determination in Pre Sporangial Stage
If the epigenetic basis of sex expression is conceived in case of cycads, then it is the outcome of environmental perturbation mediated phenotypic plasticity, which is defined as the ability of an organism to change its phenotype in response to changes in the environment. However, while looking for an ideal marker for sex determination in pre sporangial stage, the markers should be completely independent of environmental conditions and should be detected at virtually any stage of plant development. The DNA based molecular markers per se probably satisfy all the essential criteria of ‘true and full proof’ markers. These markers are based on naturally occurring polymorphisms in DNA sequences (i.e., base pair deletions, substitutions, additions or patterns). There are various methods to detect and amplify these polymorphisms so that they can be used for genetic analysis. The DNA based molecular markers are superior to other forms of markers because they are relatively simple to detect, abundant throughout the genome, completely independent of environmental conditions and can be detected at virtually any stage of plant development. There are five conditions that characterize a suitable molecular marker (Gupta and Varshney 1999): 1) Must be polymorphic 2) Co-dominant inheritance 3) Randomly and frequent distribution throughout the genome 4) Easy and cheap to detect 5) Reproducible. Consequently, these markers have been used for several different applications including: germplasms characterization, characterization of transformants, study of genome organization, phylogenic analysis and genetic vis-à-vis sex diagnostics. Almost all of the molecular markers are either based on DNA-DNA hybridization or polymerase chain reaction principle or sometimes a combination of both. Some of these important markers, which have been used in sex determination, particularly in case of gymnosperms are as follows.

The discovery of RAPD or Random Amplified Polymorphic DNA (Williams et al. 1990) based genetic marker in the beginning of 1990s probably revolutionized the application of till then sophisticated Polymerase Chain Reaction technique. The instant popularity of this technique was probably due to its simplicity since RAPD analysis does not involve hybridization/ autoradiography or high technical expertise. It uses one or sometimes two short arbitrary primers (usually 8–10 bases) to amplify anonymous stretches of DNA which are then separated and visualized usually by simple agarose gel electrophoresis. Many different fragments are normally amplified using each single primer; the technique has, therefore, proved to be a fast method for detecting polymorphisms. Furthermore, the requirement of only tiny quantities of target DNA and relatively easy procurement of the different series of random primers from the commercial manufacturer made the unit costs per assay quite low. Suddenly all the field of Biology jumped to it and to the Population Geneticists it was initially probably the perfect technique they were looking for so long. However, the theoretical and practical limitations of this technique soon became quite evident.

From the practical point of view it was observed that RAPD does suffer from sensitivity to changes in PCR conditions resulting in changes to some of the amplified fragments. Considering theoretically, it was understood that RAPD has problems of co-migration: and often it was found that result interpretation became problematic since it was difficult to predict whether same RAPD derived band is same DNA fragment and whether one band is solely one fragment. This is because the type of gel electrophoresis used, while able to separate DNA quantitatively (i.e., according to size), cannot separate equal-sized fragments qualitatively (i.e., according to base sequence). Finally, the dominant nature of the RAPD markers was found to be another bottleneck for distinguishing homozygotes from heterozygotes thus proving this system quite unsuitable in Marker Assisted Selection particularly at the stage of F1 hybrids.

In spite of the above shortcomings, the most popular markers for sex determination in plants surprisingly include RAPD (Milewicz and Sawicki 2013) and gymnosperms vis-à-vis cycads are no exception to that. The reason probably is simple since in the absence of any specific genome information in most of the non-model plants and particularly the gymnosperms, the workers have to adopt this fast yet random technique to look for polymorphism between plants of expressed sex and later try to correlate their findings in population of unknown or yet to be expressed sex (i.e., at pre sporangial stage).

The group of the present reviewers has successfully used RAPD technique for detecting sex related marker in Cycas circinalis and observed readily distinguishable sexual dimorphism of this Gymnosperm in members of contrasting sex within population (Gangopadhyay et al. 2007). Of the RAPD fingerprints generated from a number of random primers, the profiles of primers OPB 01 and OPB 05 were noteworthy, since they represented one male-specific (686 bp) and another female specific (2048 bp) band respectively (Fig. 2) (Gangopadhyay et al. 2007). Sequencing of these two cloned DNA fragments, followed by BLASTX searching, revealed maximum homology with reverse transcriptase of Ginkgo biloba (score 69.3 bits; NCBI accession no. AAY87195) in case of male-specific DNA fragment (NCBI accession DQ386640, dated 22.02.2006), while the female-specific DNA fragment did not result in any significant match.

Unique RAPD derived polymorphism was also detected by the present group between male and female Zamia fischeri plants (Roy et al. 2012). The RAPD profiles of both OPB03 and OPB04 primers showed one male and one female-specific DNA fragments in each of the cases (Fig. 2). The molecular mass of male and female specific fragments in case of OPB03 was 294 and 534 kb, while those were 1320 and 1015 kb respectively in case of OPB04. The specific DNA fragments of both male and female samples of OPB03 were eluted out from gel, cloned and subsequently sequenced. Sequencing of these two cloned DNA fragments, followed by BLASTN searching, revealed homology with Araucaria angustifolia (a conifer) clone AAng27 micro satellite sequence (maximum identity 83%; GenBank accession no. AY865591) in case of male specific DNA fragment (GQ141708), while the female specific DNA fragment (GQ141709) did not result in any relevant homology with the available database.

Similar RAPD based experimental approach was undertaken for sex identification in Encephalartos natalensis, another cycad (Prakash and Staden 2006). Initially, the workers used 140 primers to amplify the bulk DNA of five individuals each of known male and female sexuality. While a high degree of polymorphism was observed in the amplification profiles of male and female plants, only primer OPD-20 generated a specific band (~850 bp) in female DNA (Fig. 2). To validate this observation, this primer was re-tested with 69 individuals of E. natalensis. The ~850 bp DNA band was present in all 38 female individuals tested and it was consistently absent in all 31 male plants tested.

An even greater endeavor was undertaken by Ling et al. (2003) while searching for a sex-associated RAPD marker in the living fossil, Ginkgo biloba. The workers screened one thousand and two hundred random decamers and landed up to a single 682 bp RAPD marker, which appeared to be maleness associated after scoring a staggering figure of 8,372 amplicons.

Though RAPD is the first choice to look for sex related polymorphism in many plants including gymnosperms but due to some inherent shortcomings of this technique, adoption of next level of marker is often recommended (Milewicz and Sawicki 2013). Conversion of RAPD polymorphic band into SCAR or Sequence Characterized Amplified Region marker though technically demanding but greatly enhances the reliability of the marker. Longer (20–30 bp) and more specific primers for these markers facilitate amplification of the desirable sequence, and they would guarantee repeatability of measurement results. Furthermore, development of sex-linked SCAR markers support sex differentiation even of a single individual—the sex-specific band appears or does not. For other marker systems, which generate the whole band patterns, it would be difficult to identify a gender without the need of comparing the band patterns for both sexes.

Endeavor in this direction has resulted in conversion of RAPD marker into female sex specific co dominant SCAR marker, which has provided a possibility of identifying the sex of Cycas tanqingii before sexual maturation, which is very important for in situ or ex situ conservation (Jing et al. 2007).  There are further issues in SCAR marker development. Ideally, a researcher should use one or two different SCAR markers which create products of different length in males and females in the same amplification. Sex-linked markers for Ginkgo biloba were determined in line with the above method. SCAR markers generated products with the length of 571 bp for males and 688 bp for females (Fig. 2) (Liao et al. 2009). Annealing temperature differed for both primer pairs. Situations such as those encountered in the study of Ginkgo biloba happen rarely. Even if the makers of both sexes are found in the same species, they are rarely discovered by the same research team, and their identification is a laborious process, leaving aside the luck or chance factor (Milewicz and Sawicki 2013).

Described as the most full-proof of all molecular marker techniques, AFLP or Amplified Fragment Length Polymorphism (Vos et al. 1995) involves the following steps: Digestion of DNA with two specific restriction enzymes, one frequent cutter and the other rare cutter; ligation of oligonucleotide ‘adapters’ to the ends of each fragment ensuring amplification of only the fragments, which have been cut by both frequent and rare cutters; designing of primers from the known sequences of the adapters plus 1–3 selective nucleotides, which extend into the fragment sequences; PCR followed by visualization of fragments in gel after autoradiography. The high number of bands eases analysis by providing better chance of polymorphism. Though theoretically sound but this time, skill and cost intensive technique is hardly suitable for evaluating large number of individuals of a population effectively. Hence, report of use of AFLP in determining sex of gymnosperm is relatively scarce. Only one report of identify cation of three female and one male specific AFLP polymorphism is available in Ginkgo biloba (Wang et al. 2001).

Micro satellites (SSR)
Micro satellite loci or simple sequence repeats (SSR) exhibit high levels of variability because of differences in the number of repeated units. The high allelic diversity and abundance of micro satellites in the eukaryotic genome make these co dominant molecular markers popular for detailed genetic studies as genetic diversity and genetic structure (Chase et al. 1996). Unfortunately, plausibly due to complexity of most of the gymnosperm genome as well as paucity of research materials in case of endangered ones, research is yet to be flourished in this direction unlike angiosperm crop plants. Only one report of development of twenty-nine species-specific and highly polymorphic micro satellite loci is available in Araucaria angustifolia from a genomic library enriched for AG/TC repeats (Schmidt et al. 2007). Future endeavor to link those SSRs with sex loci may throw some light in early detection of sex in gymnosperms.

Functional Genomics in Sex Determination
A promising functional genomics approach was undertaken by Zhang et al. (2002) to understand the molecular mechanisms controlling development of sexual characters in Cycas edentate. Their attempt towards cloning genes expressing differentially in male or female reproductive organs culminated in a novel gene, named Fortune-1 (Ft-1), with enhanced expression in male reproductive organs. The 593-base-pair Ft-1 cDNA was predicted to encode a 77-amino-acid protein, and exists as a single copy gene in the C. edentata genome. Ft-1 expression was enhanced in male cones, including the cone axis, microsporophylls and microsporangia, but was reduced in ovules and undetectable in megasporophylls. The secondary structure prediction and homology search of Ft-1 protein showed that it has a helix–loop–helix motif, and predictably it was without any homologue in the database indicating paucity of basic research work in non-model, rare plant materials.

Since sex is the queen of problems in evolutionary biology (Bell 1982), understanding the molecular factor(s) behind sex expression has immense importance both in basic and applied research. Despite the growing body of research, the mechanism of sex determination in many plant species remains unexplained, and this is even truer in case of gymnosperms and cycads to be precise. Though the search for molecular sex-linked markers paves the way for future scientific discoveries, on the other hand, sex-linked markers alone do not explain the molecular mechanism of sex determination in dioecious plants, but the number of markers, their sequence structure and homology between sequences characteristic of males and females provide a certain venture point for studies into sex determination mechanisms. Still, it appears that there remains a large gap between the theoretical understanding and the technique/marker of choice for early sex determination (in pre sporangial condition) that will be helpful for breeding program prior undertaking a proper conservation strategy. If an epigenetic control behind the conflict regarding the existence of sex chromosome having unequal length is finally envisaged on the backdrop of occasional sex change in cycads, then a proper technique, which considers the epigenetic mechanism, especially DNA methylation has to be adopted for future research towards sex determination in cycads. In this regard adoption of technique like Methylation-Sensitive Amplification Polymorphism (MSAP) seems to be promising, which has been first tried for sex determination in some cycads (Kanchanaketu et al 2007). The attempt of their work using modified AFLP technique using isochizomer enzyme (MspI and HpaII) or MSAP was carried out to assess the pattern of cytosine methylation in both sexes of Cycas and Zamia. Using seven pairs of primers, a total of 364 bands, some of which showing sex-specific were produced and classified into three groups. The first group was non-polymorphic markers, whereas the second group was chosen from the results of differentiation ability of MspI and HpaII to cut the methylated sequences, but sex-different markers were still not obtained. Markers in the third group were methylation-sensitive and they also showed some polymorphic patterns between the two sexes. Their end suggestion that sex in cycads may be associated with DNA methylation is probably justified, which however, warrants further studies in this direction with state-of-art techniques to reach a final conclusion.

Jan 12, 2017

Pollen-ovule Interactions in Gymnosperms

After pollen is delivered to ovules via wind or insects, there are complex interactions between pollen and ovules prior to fertilization in seed plants. These interactions are integral in the functional reproductive biology of seed plants. Among extant gymnosperms there are diverse pollen delivery mechanisms that bring pollen into the ovule, where it germinates. Pollination mechanisms can be classified into types according to presence and absence of pollination drops, and by pollen grain and/or ovular modifications. Although most pollination mechanisms of gymnosperms can be grouped into four pollen capture mechanisms (PCMs) that describe pollen-ovule interactions, there are a few gymnosperms in which ovules neither capture the pollen directly, nor are they the site of germination. We classify these mechanisms as extra-ovular capture and germination (ECGs). In most gymnosperm genera, pollen grains germinate in pollination drops—liquid secretions containing minerals, carbohydrates and proteins, including active enzymes. Once germinated, pollen interacts primarily with the nucellus. The fossil record of seed plants suggests now extinct modes of pollen-ovule interactions. We provide a phylogenetic framework to show that pollination drops were probably a fundamental part of pollination in the earliest seed plants. To this we have added recent research to shed light on the evolutionary, developmental and biochemical aspects of pollen-ovule interactions. Pollination drop biochemistry offers a particularly rich area of exploration. Ongoing and future research in pollen ovule interactions are expected to provide new insights into this aspect of gymnosperm reproductive biology.


Pollen released from male cones of gymnosperms is delivered to ovules either by the wind (anemophily) or by insects (entomophily). From this moment until fertilization, pollen will have many interactions with the ovule. The capture of pollen that will eventually produce sperm for fertilization, is described as a Pollination Mechanism. Pollination mechanisms are diverse among the extant lineages of both anemophilous and entomophilous gymnosperms. The variation in the structure of pollen and ovules are considered to be reproductive adaptations (Tomlinson et al. 1997; Doyle 2013). There is also a significant diversity in biochemical constituents of liquids associated with pollination capture, e.g., pollination drops (Fig. 1; Gelbart and von Aderkas 2002). As drops withdraw, pollen is transported into the interior of the ovule. This retreat of the drop is either caused by evaporation, active retraction of the drop, or by a combination of both active and passive processes (Mugnaini et al. 2007). Among modern gymnosperms, pollen grains inside the ovule germinate and produce outgrowths that penetrate the nucellus.

In the case of Cycadales and Ginkgo biloba L., this outgrowth is haustorial in nature, but in all other modern gymnosperms (Gnetales and conifers sensu lato) the outgrowth is a more or less linear pollen tube, lacking or only having minimal haustorial side branches, which delivers non-motile sperm to the eggs. Pollen tubes, whether haustorial or linear, grow into the ovule nucellus. The time from pollination until fertilization shows substantial variability among different groups of modern gymnosperms that have been studied (Willson and Burley 1983; Williams 2012). The shortest time from pollination to fertilization among gymnosperms is found in Ephedra L., a gnetophyte; its pollen germinates, the pollen tube grows through the nucellus and releases male gametes, all within a day (El-Ghazaly 1998; Williams 2009). However, in most gymnosperms, this period is much longer. In conifers, pollination to fertilization is two weeks in Picea A. Dietr. (Runions and Owens 1999 a, b), two months in Pseudotsuga Carrière (von Aderkas and Leary 1999a), and over a year in Pinus L. (Owens et al. 2005). These long periods of pollen growth within the nucellus require coordination of pollen and ovule development. Delivery of male gametes to briefly receptive eggs requires more complex physiological and molecular mechanisms than, for example, pollen capture.

Of the three stages of pollen-ovule interaction that we discuss in this chapter, some have received much more attention than others. For example, studies of first contact of pollen with ovules (pollen capture) are the most common type of study, as this stage is easier to study than the prolonged and complex interactions of microgametophytes growing within ovular tissues. The advantage to these numerous descriptive pollen capture studies is that they allow more comparisons and inferences to be made. In contrast, there are few experimental studies, e.g., wind tunnel experiments on ovule capture (Niklas 1982; Niklas and Kyaw 1982), or the physiological study of pollination drop withdrawal (Xing et al. 2000; Mugnaini et al. 2007; Leslie 2010; Jin et al. 2012). Interactions between pollen (including the microgametophyte), and nucellus are the least studied. It is not well-known whether there is an influence of the nucellus on pollen growth, or of the pollen on ovule development.

In this chapter, these various interactions will be outlined in three parts: i. pollination capture mechanisms, ii. Germination, and iii. Nucellus pollen interactions. Where we have sufficient comparative evidence, e.g., pollination mechanisms, we will provide a guide to the evolutionary history of pollen-ovule interactions through deep-time, including both extant and extinct gymnosperms. We outline these evolutionary, developmental and biochemical aspects of pollen-ovule interactions, with a view towards identifying hypotheses for future lines of research and providing new insights into this key group of seed plants.

Pollination Mechanisms
Pollen can enter the ovule in one of two ways: i. pollen is captured by the ovule, or ii. Pollen germinates outside the ovule and produces a pollen tube that grows towards, and then into, the nucellus. The first we will call a pollen capture mechanism (PCM), the second an extra-ovular capture and germination (ECG). These syndromes can be broadly considered as pollination mechanism sub-types.

Pollen Capture Mechanisms (PCMs)—Extant Taxa
Of these two general classes of pollination mechanisms, PCMs are the most common type. They are found in the only species of Ginkgo, and all species of Gnetales and Cycadales. PCMs also are found in most of the Pinaceae and Podocarpaceae, though there are some notable exceptions that will be discussed later. There are four general types of PCMs: i. pollination drop capture and delivery with non-saccate pollen, ii. Non-drop capture, i.e., physical entrapment of saccate pollen, followed by drop-mediated transport into the ovule, iii. Non-drop capture of saccate pollen by inverted ovules with scavenging pollination fluid, and iv. Non-drop capture of non-saccate pollen followed by physically-mediated pollen delivery into the ovule followed some weeks later by a drop.

The first PCM, pollination drop capture and delivery with non-saccate pollen, is the most widespread (Tomlinson 2012). A secreted drop appears at the entrance to the ovule. The drop persists, depending on species, for days or weeks. Wind- or pollinator-borne pollen is captured by the drop and sinks into the interior of these typically upright ovules (Xing et al. 2000). Ginkgo, as well as all extant Cycadales and Gnetales have this mechanism (Gelbart and von Aderkas 2002). Additionally, all members of some conifer families capture pollen in this way (Cupressaceae, Taxaceae, Sciadopityaceae, and Cephalotaxaceae).

The second PCM initially traps pollen on paired micropylar extensions. The pollen is modified with two hemispheric projections called sacci that promote buoyancy in pollination drops. Windborne pollen grains adhere to lipid microdrops on the micropylar extension surfaces. Soon afterwards, a pollination drop is secreted from the ovule that flows into the space between the flaps, filling it and liberating the pollen. The saccate pollen float upwards into the downward-oriented ovule via the micropyle. This mechanism is found in some Pinaceae, e.g., most species of Picea (Doyle and Kane 1943) and Pinus (Owens et al. 1981), Cedrus Mill. (Takaso and Owens 1995), and Tsuga mertensiana (Bong.) Carrière (Owens and Blake 1983). Additionally, Abies Mill could be classified as having PCM ii, but no pollination drop has yet been observed for this genus (Singh and Owens 1982).

The third PCM is restricted to Podocarpaceae, in which there are genera that have both inverted ovules and saccate pollen (Tomlinson 1994; Tomlinson et al. 1991, 1997). Unlike most other pollination drops, these are involved in pollen scavenging. These ovules produce pollination drops that do not form a hemispheric drop because the entrance to the ovule as well as the nearby surrounding bract surfaces have little wax and are hydrophilic. As a consequence the drop spreads along the surfaces adjacent to the ovule. Saccate pollen captured by the liquid eventually float upwards into the ovule, or are drawn into the ovule as the drop retracts. Over a number of days, the pollination drop is repeatedly secreted and retracted, scavenging pollen grains from the surfaces adjacent to the ovule.

In the fourth PCM, pollen is trapped on hairs found on two flap-like integumentary extensions at the ovule’s entrance. The hairs synchronously collapse inwards, pushing the pollen grains further into the micropyle (Barner and Christianson 1962). The pollen grain swells, and stays in this state for approximately two months (Takaso and Owens 1994). Pollen grains only form tubes after a drop is secreted (Said et al. 1991). Because of the substantial delay in secretion, this drop has been called the post-pollination pre-fertilization drop to differentiate it from a pollination drop (i.e., PCM i), although they both share functions such as transport of pollen and induction of pollen germination (von Aderkas and Leary 1999a, b). Like the second mechanism, this one is also restricted to the Pinaceae, in particular to Pseudotsuga and Larix Mill. (Doyle and O’Leary 1935a).

Extra-ovular Capture and Germination (ECG)—Extant Taxa
Unlike PCMs, extra-ovular capture and germination (ECG) is rare among extant gymnosperms. This type of pollination mechanism, found in several families, has no drop at any point in the process. In all cases, pollen tubes germinate outside the ovule, growing towards, and then into, the nucellus.

This is in contrast to some conifer pollen, e.g., Pinus, which can germinate readily in high humidity conditions outside of ovules, but fails to fertilize the ovule because the pollen tubes are unable to find the ovule entrance. In the case of conifers with ECGs, pollen must find either the nucellus and or the micropyle opening. The distance that tubes are able to grow and still reproduce successfully, i.e., reach the nucellus, varies among species. In ECGs, how pollen tubes are attracted to the nucellus is not known; the expectation is that there is form of chemotaxis present.

The ovules of Saxegothea Benth. & Hook. f. (Podocarpaceae) have a wide micropyle with a protruding nucellar beak (Doyle and O’Leary 1935b; Tison 1908). Air-borne pollen that lands on the bract, the ovule exterior, or on the nucellus will germinate. This is similar to what is seen in Araucariaceae.

In Agathis australis (D. Don) Lindl. and in the genus Araucaria Juss. (Araucariaceae), pollen lands on the bract surface at a distance from the micropyle or on the ovule. It germinates and the tube grows along towards the micropyle until it reaches the nucellus, which, typical of species in the family, extends out of the micropyle (Owens et al. 1995).

Pollen of Tsuga pattoniana Engelm. (Pinaceae), captured on micropylar extensions of the integument will germinate and tubes grow into the micropyle, entering the ovule interior (Doyle and Kane 1943). Tsuga heterophylla Sarg. Pollen behaves like that of Araucaria. Its spiny pollen grains are trapped in cuticle hairs on the bract surface, sometimes at substantial distances from an ovule entrance. Pollen germinates and the pollen tubes seek the micropylar entrance to the ovule (Doyle and O’Leary 1935a).

Fossils and Interpretation of PCM Evolution
The major lineages of extant seed plants, angiosperms, Ginkgo, Gnetales, Pinaceae, Cycadales, and non-pinaceous conifers represent only a small portion of the diversity of seed plant lineages that first arose in the late Devonian (Rothwell and Serbet 1994; Hilton and Bateman 2006; Doyle 2013). Thus, meaningful interpretations of origins and evolutionary changes in pollen-ovule interactions require the inclusion of extinct seed plants. A seminal paper by Doyle (1945), established various evolutionary models of how pollination mechanisms may have evolved over time, mainly using transformational series with key fossils as ancestors to illustrate a model for the origin of modern diversity. Recently, Tomlinson (2012) further developed these ideas, incorporating temporal aspects of seed plant reproductive syndromes established by Robert Brown (1844), including fossil and modern taxa. Pollination drops are generally accepted to be a fundamental aspect of the evolution of PCMs in gymnosperms, and pollination drops play key roles in pollen-ovule interactions. However, preservation of fossil pollination drops is rare (Rothwell 1977), and as a result, much of the evolutionary inferences regarding drops rely on a synthesis of phylogeny, modern biology, and over a century of paleobotanical research.

We used the phytochrome gene duplication rooting of seed plants (Mathews et al. 2010) as well as the sister-group relations of the 6 major extant seed plant lineages based on various DNA sequences (Graham and Iles 2009) as a backbone constraint for the phylogenetic analysis of a morphological matrix of seed plants that includes fossils (Doyle 2008). We produced 10 trees of 348 steps, using the program TNT (Goloboff et al. 2000), arriving at the same 10 trees using both the ratchet as well as typical TBR heuristic searches. Mapping the presence of a pollination drop on the phylogeny using parsimony reconstruction (Maddison and Maddison 2010) suggests that the occurrence of pollination drops was probably fundamental to PCMs of the earliest seed plants, not just in crown-group seed plants/ modern lineages (Fig. 2).

Inference of ancestral pollination drops is further supported by using the presence of saccate pollen in fossil taxa as a proxy for the presence of pollination drops. Although virtually all modern PCMs with saccate pollen have a pollination drop, they all have an intervening non-drop based pollen capture mechanism (i.e., PCMs ii and iii). The evolution of saccate pollen is considered part of an evolutionary syndrome that would have included inverted ovules at the time of pollination (Doyle 1945; Runions et al. 1999; Leslie 2008). However, there are no known modern gymnosperms with a simple case of inverted ovules with exuded pollination drops that serve as both the pollen capture and transport mechanism (i.e., PCM i); this might be expected to have occurred in the evolution of PCMs that include saccate pollen. Making the assumption that there is a pollination drop present if a species has saccate pollen, has been part of evolutionary interpretations of PCMs since Doyle (1945). Thus, for the scoring of presence of pollination drops in fossils, we score each taxon as having a drop if it has saccate pollen grains.

Medullosans had large pollen grains that could not have been wind transported (Niklas 1983; Schwendemann 2007). The medullosan pollen organ Parasporotheca Dennis and Eggert (1978) was found with Parasporitestype pollen in situ. Parasporotheca pollen have saccus-like structures, suggesting that species of the genus may have had pollination drops and that Parasporotheca-bearing plants had evolved to having inverted ovules at, or around, the time of pollen capture. Thus we score Medullosans as having pollination drops, with the assumption that saccate pollen evolves in species with existing drops.

Glossopterids are scored as having pollination drops and having zoodiogamy. Species in this genus had saccate pollen (Ryberg et al. 2012a), supporting the scoring of droplet presence in the lineage. Nishida et al. (2003, 2004) observed fl agellated sperm within a pollen chamber. Most of the Paleozoic gymnosperms associated with coniferophytes and conifer evolution have saccate pollen. Some species or all species may be saccate, these details are succinctly summarized by Doyle (2010).

Inferring ancestral PCMs without information from the fossil record is probably misleading since logically there would be several evolutionary steps between PCM i and any other pollen capture mechanism. For example, several PCMs are only found in Pinaceae and nowhere else, such as those with saccate pollen and non-drop primary pollen capture, a minimum of two trait changes from PCM i. The fossil record of Pinaceae begins in the Lower Cretaceous (131–129 million years ago; Ryberg et al. 2012b), relatively recent in comparison to the oldest known fossil gymnosperms ca. 300 million years (Clarke et al. 2011). Determination of PCMs from the paleobotanical record would aid in inference of how extinct seed plants shifted pollination syndromes, but is rare. An additional complication of inferring PCM evolution in deep-time is that several pollination mechanisms that have been inferred from the fossil record have no modern analogue such as that of the Mesozoic cheirolepidiaceous conifer Alvinia (Kvaček 2000; Labandeira et al. 2007; Labandeira 2010), in Medullosales (Stewart 1951; Niklas1983), although many PCMs reconstructed from fossils appear to be more or less modern (e.g., the Paleozoic conifer Otovicia; Kerp et al. 1990).

Since the discovery of swimming sperm in Cycas L. and Ginkgo L. (Hirase 1896a, b; Ikeno 1896), the prevailing thinking is that free swimming sperm is a pleisiomorphic trait of seed plants. This is also logical since all free-sporing/seed-free land plant lineages have free swimming sperm. Thus, given that the earliest seed plant lineages had nonsaccate pollen, this suggests that some version of PCM i and zoodiogamy, may have been present in the earliest gymnosperms. If this is the case, then pollination drops could have performed the dual function of pollen capture and delivery into the ovule interior (i.e., PCM i), but may have also provided the medium for swimming sperm to reach and fertilize the egg (Fig. 2). This idea is not new to paleobotanical research (e.g., Brongniart 1881; Renault 1887; Stewart 1951), but is interesting in that such interpretations remain reasonable in light of modern phylogenetic hypotheses. Further support from the fossil record comes from rare observations of microgametophytes, such as two probable sperm in pollen enclosed in a medullosan pollen chamber (Stewart 1951), pre-germination microgametophytes in pollen of Idanothekion, Callistophytales (Millay and Eggert 1974), and direct observation of cycad-like top-shaped sperm in the pollen chamber of Homevaleia gouldii H., a Glossopterid (Nishida et al. 2003, 2004).

Hydrasperman reproduction, found in Elkinsia Rothwell related plants (Elkinsiales) and in Lyginopteridales probably had a variant of PCM i, but with possible zoodiogamy directly within pollination drops (Fig. 2). These groups, along with medullosans, and several extinct gymnosperms had pre-pollen, in which microspores have proximal germination. Proximal is in relation to the meiotic mark, the face that is toward the meiotic divisions (Rudall and Bateman 2007). For the purpose of character mapping, we score all fossil taxa with prepollen as zoodiogamous. All modern pollen has distal or non-proximal germination (Poort et al. 1996). Zoodiogamy from haustorial microgametophytes is found in modern Cycadales and Ginkgo, and their position in the seed plant phylogeny suggests that at least late stem seed-plants were also zoodiogamous (Fig. 2). Thus the ancestral seed plant PCM probably had no modern analogue. These early gymnosperms would have had non-saccate pollen, delivered by wind that was captured by pollination drops on erect ovules, and had pre-pollen which germinated in pollination drops that may have directly released free-swimming sperm.

Pollination Drops as a Medium for Germination
Pollination drops serve as the medium for pollen germination in most extant gymnosperms. Pollen germination can occur soon after pollination. In Cycadales, pollen germinates in the residual droplet contained in the pollination chamber (Choi and Friedman 1991). Anhaeusser (1953) was able to germinate pollen of Taxus in isolated pollination drops. In Ephedra, pollen germinates in the droplet within hours of pollen capture (Williams 2009, 2012; El-Ghazaly 1998). In other cases, pollination and germination are separated by a number of weeks, yet the drop still acts as the trigger for germination. In both Pseudotsuga and Larix, pollen is captured and brought into the ovule by stigmatic hairs. Only weeks later, when the post-pollination pre-fertilization drop is released does pollen germination occur (Said et al. 1991; Takaso and Owens 1996). In vitro studies of pollen germination support the idea that specific biochemical components are required for pollen germination (Brewbaker and Kwack 1963; Nygaard 1977; Dehoux and Pham Thi 1980). Many taxa have slower pollen tube growth rates in vivo versus in vitro (Williams 2012) which suggests that there are components in pollination drops that mediate/control germination. This leads one to the question, what is contained in pollination drops that promotes and/or moderates germination rates? A range of organic and mineral compounds has been identified in pollination drops of gymnosperms. Together these constituents create the biological environment in which pollen germinates.

Sugars are present in the drops of conifers (McWilliam 1958; Ziegler 1959), Cycadales (Tang 1987), Gnetum L. (Kato et al. 1995), and Welwitschia Hook. f. (Carafa et al. 1992), and Ephedra (Ziegler 1959; Bino et al. 1984a, b). It is likely that sugars are present in Ginkgo drops as well. Glucose, fructose and sucrose were identified in a number of conifers (McWilliam 1958; Ziegler 1959; Seridi-Benkaddour and Chesnoy 1988) and Cycadales (Tang 1993). Other sugars, such as mannitol (Mugnaini et al. 2007), galactose (Carafa et al. 1992), xylose and melezitose (von Aderkas et al. 2012) have also been identified. Sugars are also present in some conifers as polymers containing arabinose, galactose, glucose, mannose, and rhamnose (Seridi- Benkaddour and Chesnoy 1988). Total sugar concentrations vary between groups. For conifers, total sugar concentrations between 1–2% have been found (McWilliam 1958). Other gymnosperms have higher concentrations: 10% for Ephedra (Ziegler 1959); 4–14% for cycads (Tang 1993); 3–15% for Gnetum (Kato et al. 1995; Nepi et al. 2009).

Drop sugars have potential roles in pollen germination. They could provide a source of energy for pollen. Monosaccharides are taken up and used by pollen during germination in vitro to support the growth of the pollen tube and the accumulation of polysaccharide storage molecules (Nygaard 1977). Additionally, sugars could be involved in mechanisms of osmotic regulation, such as at higher concentrations to provide an osmotic environment that inhibits microbial growth. Total sugar concentration varies greatly between groups, and specific pollen types may have optimal osmotic conditions for germination, thus providing germination specificity in different osmotic environments, a probable adaptation. Sugar concentration and composition have been observed to be controlled by enzymes present in the drop, in some taxa. A functional invertase is present in Pseudotsuga, breaking down sucrose into glucose and fructose (von Aderkas et al. 2012) thus affecting proportions among these sugars.

Proteomic studies have revealed that pollination drops of conifers and Welwitschia contain a number of proteins. These include xylosidases, invertases, aspartyl proteases, peroxidases, serine-carboxypeptidases, and galactosidases (Poulis et al. 2005), thaumatin-like proteins (Wagner et al. 2007; O’Leary et al. 2007), and chitinases (Poulis 2004; Wagner et al. 2007). Additional pollination drop proteins found in cuppresaceous conifers include a glucanase-like protein, a glycosyl hydrolase, glucan 1, 3-ß-glucosidases, a ß-D-glucan exohydrolase, subtilisin-like proteinases (Wagner et al. 2007). Several arabinogalactan proteins occur in Taxus x media Rehder, discovered using immunohistology (O’Leary et al. 2004).

Drop proteins likely play an active role in pollen germination. Like sugars, their presence may alter the osmotic environment of the drop (Wagner et al. 2007). If broken down to free amino acids, they may also provide a source of nutrients for germinating pollen by supplying key components for protein synthesis within pollen tubes as they grow (Zhang et al. 1982). In vitro, externally supplied free amino acids have been observed to increase pollen tube growth and development (Dehoux and Pham Thi 1980). Proteases present in the drop are the expected driver of free amino acid concentrations (Poulis et al. 2005). Other active enzymes may also affect germination. Xylosidases and galactosidases could loosen the pollen cell wall by cleaving xyloglucans, a group of hemicelluloses that support the cellulose microfibrils of the cell wall (Poulis et al. 2005). This would help prime the pollen wall for tube emergence. Additional proteins, such as chitinases (Coulter et al. 2012) and thaumatin-like proteins may function to inhibit microbial growth (O’Leary et al. 2007).

Mineral components are also present in pollination drops and are known to affect pollen germination and growth. Calcium has been found in Taxus baccata L. (Fujii 1903), Larix and Pseudotsuga (von Aderkas et al. 2012). Application of calcium in pollen germination media was a key discovery for development of culture methods (Brewbaker and Kwack 1963). In vitro, calcium is required for pollen tube elongation in Norway spruce (Lazarro et al. 2005). Calcium sustains pollen viability and increases the percentage of pollen grains producing pollen tubes in Pseudotsuga (Fernando et al. 1997). Calcium-regulating proteins have been identified in pollen grains of Pinus yunnanensis Franch. (Gong et al. 1993) and Cryptomeria japonica D. Don (Yokota et al. 2004), suggesting an active role for calcium during pollen germination in conifers. Pollination drops are complex mixtures of organic and mineral compounds in which germination takes place in many gymnosperms. There is evidence that pollination drops provide genus-specific conditions, unique enough to provide a form of pre-zygotic selection (von Aderkas et al. 2012), which includes pollen-nucellus interactions.

Pollen-Nucellus Interactions
Among modern gymnosperms, after germination of pollen within the drop, or without a drop as in ECGs, pollen tubes produced from the distal sulcus grow into the nucellus. In the siphonogamous gymnosperms (Gnetales and conifers), the tube acts as a conduit for delivering the sperm to the egg. In the zoodiogamous gymnosperms (Cycadales and Gingko), the tube functions mainly in the transfer of nutrients from the nucellus to the developing male gametophyte, and motile sperm are released from the swollen, proximal end of the pollen (Rudall and Bateman 2007). Among fossil seed-plants, pollen tubes are only known in a single case in Callistophytales (Rothwell 1972, 1981) and in Williamsonia Carruthers, Bennettitales (Stockey and Rothwell 2003). It is not known if prepollen (with proximal germination) produced pollen tubes.

Tube morphology is variable among lineages, perhaps as a consequence of variable function (Friedman 1993). In Gingko and Cycadales, tubes form branches, which facilitate the uptake of nutrients by increasing the surface area through which exchange occurs. In Gingko, tubes are highly branched and slender, and grow between the cells of the nucellus (intercellular growth; Friedman 1987, 1993). In contrast, the tubes of Cycadales are wide, and may be unbranched or form small side branches. Growth can be intercellular or tubes may penetrate the nucellus cells (intracellular growth). In either case, cells adjacent to the tube are degraded by tube growth and possibly by enzymes emitted by the tube; hydrolases and acid phosphatases have been found in the intine of several cycads (Pettit 1977, 1982). Degradation may increase the availability of nutrients for absorption (Pettit 1977, 1982; Choi and Friedman 1991), although mechanisms of absorption are unknown.

The proximal end of the pollen does not branch. Instead, it swells inside the pollination chamber just above the archegonia. The spermatogenous cell, which remains inside the proximal end of the pollen, divides to form the sperm. At maturity, the proximal end bursts, and releases the mobile gametes that swim through the fluid-filled pollination chamber to the archegonia (Friedman 1987). Ultimately, one sperm nuclei will fuse with the egg nuclei after entry of the sperm into the archegonia via the neck cells.

In the extant, siphonogamous conifers and the Gnetales, the tube is usually unbranched, and grows inter-cellular, more or less unbranched, and in a direct path through the nucellus towards the archegonia. In vitro studies suggest that a chemical signal from the female gametophyte may attract the pollen tube (Dumont-Beboux et al. 1998). Each tube penetrates an archegonium, bursts, and releases two sperm, one of which will fuse with the egg (von Aderkas et al. 2012). In Gnetum and Welwitschia, sperm are released into surface cells of the mega gametophyte which seem to be able to function as eggs, even though the cells do not have obvious anatomical modifications seen in eggs and archegonia in other gymnosperm groups (Singh 1978; Fernando et al. 2010). As the tube grows, adjacent cells are often damaged, although less so than what is observed in Cycadales. Cell damage may result from the physical impact of the passing tube, or may be induced by enzymes released by the tube, in a manner similar to that proposed in Cycadales (Friedman 1993; Pettit 1985). It may also be a result of programmed cell death induced by the pollen tube (Hiratsuka et al. 2002). The tube grows more easily through space created by damaged cells, and may absorb nutrients released by cell break-down (Fernando et al. 2005a).

Pollen tube branching may occur in some conifers. In Pinus contorta Douglas many short branches form during the first year of cone development (Owens et al. 1981). In Podocarpus, the tube grows in a straight line, then branches near the female gametophyte, enlarging to form a disk-like structure over the neck cells; small side branches grow through the megaspore membrane and nucellus (Wilson and Owens 1999). In Araucaria and Agathis, tube branches form after penetration of the megagametophyte (Owens et al. 1995). Although nutrient absorption and tube branches may occur in some siphonogamous taxa, sperm transfer is viewed as the primary tube function. Sperm travel through the tube to reach the archegonia. In contrast, sperm do not enter the haustorial tube in Ginkgo or Cycadales (Rudall and Bateman 2007).

The ovule may influence the growth and development of the pollen, and pollen selection may occur before fertilization (Gelbart and von Aderkas 2002; Takaso and Owens 1994). In angiosperms, it is well-known that pollen selection occurs as the pollen grows through carpellary tissues to reach ovules. Self-incompatibility (SI) reactions are well understood (for review, Takayama and Isogai 2005), but selection mechanisms that act on pollen from foreign species and genera also exist (e.g., de Nettancourt 2001). In gymnosperms, selection was widely thought to occur mainly during embryo development; the over-expression of lethal alleles causes abortion of some embryos (Williams 2008).

In self-pollinated ovules, several studies suggest that low seed set may sometimes be linked to the failure of pollen tube growth in the nucellus, for example: Larix (Kosiński et al. 1986), Abies (Kormutak et al. 1999), Picea (Runions et al. 1998), and Thuja (Owens et al. 1990). However, statistical analyses did not support this finding in any of the studies, and several of the studies did not include data to support this claim. Other studies have looked for and not found evidence of self-incompatibility (Orr-Ewing 1956; Plym-Forshell 1974). Thus, further studies should explore self-incompatibility across gymnosperms before general conclusions can be drawn.

Pollen growth and development in some Pinaceae is dependent on whether the pollen is conspecific or heterospecific with respect to the ovule: Pinus (Stockwell 1939; Buchholz 1944; McWilliam 1959; Hagman 1975; Fernando et al. 2005b), Picea (Mikkola 1969), and Pseudotsuga and Larix (von Aderkas et al. 2012). Several of the studies had little or no data to support this observation, but others did have ample data that were supported by statistical analyses. McWilliam (1959) performed interspecific crosses using four species of Pinus, and found that growth of foreign pollen was halted in the nucellus or at the time of germination. In reciprocal crosses between Larix x marschlinsii Coax and Pseudotsuga menziesii (Mirb.) Franco, von Aderkas et al. (2012) found that foreign pollen germinated less frequently than conspecific pollen, and that growth of foreign pollen was interrupted at various subsequent stages of development. Fernando at al. (2005b) performed reciprocal crosses between Pinus lambertiana Douglas and P. monticola Douglas, and found that pollen tube growth aborted in the nucellus of one cross, whereas in the other, male and female gametes did not fuse. In contrast, one study examined the in vitro interaction between pollen tubes and female gametophytes of four different Pinaceae genera, and found that no barriers to the entry of foreign pollen tubes into the female gametophyte exist (Dumont-Beboux et al. 1998). Further studies that examine these phenomena in a wider range of taxa and at various stages of pollen development are needed.

Not only can the ovule affect the growth and development of pollen, but pollen can also affect ovule growth and development. In many extant taxa, ovule development is dependent on pollination and subsequent pollen tube growth. Requiring pollination for further ovule development has been observed in Ginkgo biloba (Nakao et al. 1998), Pinus (Owens et al. 2005), Picea (Dogra 1967; Owens and Blake 1984), Thuja (Owens et al. 1990), and Juniperus (Ortiz et al. 1998). Thus it is likely that there is a signal from developing pollen that is required for continued ovule development; if pollination and germination does not occur, the ovule will abort. In other taxa, ovule development continues whether or not pollination occurs, but instead abortion of ovules occurs if fertilization does not occur (e.g., Pseudotsuga, Owens et al. 1991).

The interdependency of pollen growth and development with that of the nucellus indicates that a strong feedback exists between the male and female reproductive tissues. Interactions between the tissues are probably mediated by the exchange of chemical signals, but no information regarding such mechanisms is known. In angiosperms a diversity of molecules, from proteins to ions, are known to be involved in pollen-ovule interactions. It is unclear what types of molecules may serve as signaling molecules in gymnosperms. The pollination drop contains a suite of proteins, some of which may influence pollen growth and development. These proteins originate from the nucellus, thus it is reasonable to assume that such proteins may also be active within the nucellus as well as the drop. Arabinogalactan proteins (AGPs) are known to function in pollen tube guidance in angiosperms (Cheung et al. 1996). AGPs are present in pollen tube walls of all gymnosperms, and also in the cell walls of the nucellus of some taxa (Fernando et al. 2010; O’Leary et al. 2004), and may be involved in signaling. Multiple proteins are released from pollen, which may also be involved in communication (Pettit 1982, 1985).

Current studies are exploring the molecular mechanisms that underlie pollen and ovule growth and development as well as the factors that affect these processes, but more work is needed. These studies, coupled with research that identifies molecules involved in pollen-ovule signaling will lead to a more complete understanding of pollen-ovule interactions in gymnosperms.