Opening the Door on the Mysterious Sexual Antics of Pawpaws

Pawpaws (Carica papaya, Family Caricaceae, Order Brassicales, commonly called papaya in other countries) are highly productive fruit trees, native to meso-America and now grown in tropical and sub-tropical areas worldwide. They are enjoyed for their pleasant flavours and aromas, and also greatly valued as a rich source of vitamin A which is a major deficiency problem in much of the developing world. But for almost all the time they have been cultivated, growers have had to deal with their strange sexual behaviour which influences all the parameters of fruit production, without any real understanding of what's going on and how best to optimise outcomes.

About 90% of all the flowering plants (angiosperms) have flowers with both male and female parts in the same flower; these are called hermaphrodite or bisexual, and are normally self-fertile. Another 5% have individual male and female flowers on the same plant (monoecious), and most of the remainder have individual male and female flowers on separate plants (dioecious). A very small minority such as pawpaw are trioecious (subdioecious), with separate male, female and hermaphrodite plants. Seed germination is the principal way in which pawpaws are propagated and traditionally it has not been possible to determine whether a given plant will be male, female or hermaphrodite until it flowers. As only the latter two produce good quality marketable fruit, this results in much wasted time and resources raising and then eliminating plants. Only one in 10-20 plants need be male to ensure sufficient pollination of females, and depending on the climate, one or other of the bisexual or female plants may be preferred for fruit production. Bisexual plants are usually preferred in tropical climates and dioecious in sub-tropical, and this preference may mean other young plants are sacrificed. Traditional growers were also aware that varying environmental conditions each season, such as temperature, precipitation, nitrogen fertilisation, plant age, light intensity, photoperiod, mite infestation and mechanical injury (eg leaf shredding after storms or loss of storage tissue from pruning) could cause strange outcomes, with males sometimes producing fruits, females doing the equivalent of the virgin birth with fruit production in the absence of fertilisation (parthenocarpy), females temporarily becoming sterile or having long peduncles, trees producing normal-shaped fruit one year and deformed fruit the next, and individual trees changing to have both female and hermaphrodite flowers (gynomonoecious) or male and hermaphrodite flowers (andromonoecious).

Although a prominent feature of pawpaw behaviour, sex lability is not unique to this species and the phenomenon has been observed in more than 50 species across 25 families including such commonly known plants as asparagus, cannabis, castor oil, corn, cucumber, holly, orchids, spinach, summer squash, weeping willow, wheat, white mulberry and wild grapes. Usually hermaphrodites are more labile than the other sex forms. Dioecious plants under stressful conditions may convert to hermaphrodites or the opposite sex, and monoecious plants may change the ratio of male and female flowers. These changes are usually consistent with evolutionary adaptations that maximise reproductive success under difficult conditions, as the stress on females carrying fruits through to maturity is much greater than males producing pollen for only a limited time. Generally, stress of one form or another results in conversion from female forms to male, and optimal conditions the reverse.

Substantial advances in understanding the pawpaw story only began with systematic study in the early 1900s. Unravelling the mechanisms controlling sexual expression illustrates how scientific knowledge progresses and also how fundamental interpretations of data are always dependent on techniques available in different eras. While the story is still incomplete, many key features are now firmly established. Some of the major steps along the way were:

• In the 1930-40s breeding experiments of the different sex types showed that offspring occurred in the following ratios:
  male X male – 1 female, 2 males,
  female X male – 1 female, 1 male,
  bisexual X bisexual – 1 female, 2 bisexual,
  female X bisexual – 1 female, 1 bisexual, and
  bisexual X male – 1 female, 1 male and 1 bisexual.
  
  These non-classical ratios provided support for the existence of a sex-determining gene that had three forms (alleles, definitions of common genetic terms are given in the 'More Information' section on the site) – M from male trees, Mh from hermaphrodite and m from female. These studies also suggested that male and hermaphrodite trees were heterozygous, Mm and Mhm resp, and females homozygous, mm. M and Mh were assumed dominant and m recessive, and dominant combinations MM, MhMh and MMh were lethal. This interpretation formed the basis of most theories over the next several decades, but it had to be successively embellished as new data came to hand.
• Flowers determining sex were found to be highly variable (more than 30 types), particularly amongst hermaphrodites, and in the 1950s it was suggested they could broadly be classified into 8 basic types, with (i) and (ii) being male-type, (iii)-(vii) hermaphrodite and (viii) female. The different types had the following characteristics:

1. staminate, - unisexual male flower on long peduncles, 10 stamens present in 2 whorls of 5 each
2. teratological staminate – found on sex-reversing males, with some degree of carpel initiation and vestigial carpels (hair-like processes) at the base, can fruit in cooler weather
3. reduced elongata - modified elongata flower, aborted pistil and reduced carpel size, more frequent in warmer conditions.
4. elongata - refers to the shape of the pistil, develops into pyriform or cylindrical fruit, 5 lateral-fused carpels, and petals fused two-thirds length
5. carpelloid elongata – transformation of the inner 5 stamens into carpel-like structures. Many sub-types depending on how many stamens are carpelloid, misshapen fruit
6. pentandria – normal hermaphrodite, stepwise transformation of stamens to carpels, short corolla tube, 5 stamens in the outer whorl on long filaments, globose and furrowed pistil with 5-10 carpels
7. carpelloid pentandria – all 5 stamens in the outer whorl become carpelloid, producing carpellodic fruit, especially under cool conditions, original carpels abort and flowers resemble pistillate type
8. pistillate – unisexual flowers, larger than hermaphrodites, 5-carpellate, lack of stamens and the most environmentally stable of all types.

• In the 1940-60s a gene balance hypothesis proposed that the M and Mh alleles were on normal chromosomes (autosomes) and the m allele on 'sex chromosomes', although at that time there was no supporting cytological evidence of these. The M and Mh regions on the autosomes were degenerate, missing genes that were necessary for plant development. This model explained why homozygous dominant MM, MMh and MhMh were lethal, as heterozygous plants with at least one copy of m were viable.
• Also in the 1960s it was suggested that rather than a single gene controlling sexual expression, a number of these genes resided in a small region of the 'sex chromosome'. This theory suggested there were 5 genes involved and various combinations of them produced the highly variable observed behaviour. It was also suggested the existence of hermaphrodite pawpaw could have been due to human selection ie it was a relatively recent development in the species. A real problem was that, like other theories, there was no definitive means of testing them.
• Another proposal in the 1960s was that pawpaw sex determination was of the XX-XY type that we're familiar with in humans and most other mammals; homozygous XX being female and heterozygous XY being male. It was further posited that the Y chromosome has a region containing a lethal factor, and that Mh was a modified form of M while still including the lethal factor.
• In the 1990s a dominant male allele SEX1-M was proposed that promotes stamen development but suppresses that of the carpels ie it is a masculinizing factor. The dominant hermaphrodite SEX1-H allele was suggested to be an intermediate with the ability to induce stamens but only partially suppress carpels. The recessive female allele sex1-f promotes carpel development but is a null allele in terms of inducing stamens. Functional stamens and carpels would develop in heterozygous SEX1-H/SEX1-f plants, and the lethal factor linked to the SEX1-M and SEX1-H alleles would explain why homozygous dominant alleles do not survive and heterozygotes are viable.
• Finally in the last 15-20 years, numerous molecular biological studies have provided unequivocal answers to a number of these perplexing observations, interpretations and management issues. Up to the present, none of the vegetative pawpaw features (eg, stem and flower colour) suggested for their ability to discriminate sex type before flowering is reliable, and growers still have to put several seedlings in each planting hole and wait for flowering to know which can be culled. Modern DNA techniques can discriminate sex type well before this, but currently are only feasible in a laboratory setting. This could change in the not too distant future as simple, cheap diagnostic genetic tests become available.

Pawpaws have 9 chromosome pairs and a small genome of 442 million base pairs with approximately 24,000 genes. The gene count may seem large but in fact is made up of many repeats, non-functional pseudogenes and non-expressed DNA. It has been estimated that the minimum gene count for viability in Angiosperms is about 13,000. Approximately 96% of the genome has now been sequenced and mapped. The sex determining region is on an incipient Y chromosome which could not be differentiated from others using classical techniques. Separation began about 7 million years ago and the male specific region (MSY) is only about 10% of Y compared to 95% in humans with more mature chromosome divergence (240-320 million years ago). For such evolution to occur, it had to become non-recombinant to allow the MSY to develop and build up differences from the X. Regions just outside the MSY are changing about 7 times faster than the genome average, ie they are 'hot spots' responsible for the rapid changes that have occurred in this relatively short time. The corresponding MSY region on the X chromosome is changing much slower. Non-recombinant changes have led to degeneration and expansion in the MSY, mainly through insertions that occurred in 2 major evolutionary events, with loss of some key genes necessary for survival while still being present on the matching X chromosome; dominant homozygotes are thus lethal and heterozygotes viable. In breeding experiments the MSY behaves like a single (linked) genetic unit since it's all on one chromosome and there's no recombination with X. The hermaphrodite sex chromosome has an equivalent region (HSY) to MSY that diverged from it only 73,000 years ago, an incredibly short time in evolutionary history but still probably too early to be the result of human selection. It consists of 8.1 million bases compared to 3.5Mb on the matching X region, and it codes for 72 genes compared to 84 on the corresponding X chromosome region; 50 of these genes are common to both and there are 24 and 14 pseudogenes in the HSY and X regions resp. To date, pawpaw is the only plant species with fully sequenced and annotated sex chromosomes. The small pawpaw genome is probably the reason for genetic diversity between different varieties being relatively minor (correlation coefficient about 0.9) compared to other flowering plants. Self-pollinated hermaphrodites (cleistogamy) are just as variable as open-pollinated dioecious lines, seemingly from rare but still finite cross-pollination.

Two genes are likely involved as the initiators of sex determination, one a carpel suppressor or masculinizing gene for carpel abortion in male flowers, and the other a stamen suppressor or feminizing gene. These two genes operate in different time frames - abortion of stamens occurs before initiation of stamen primordia whereas the male sex determination gene aborts carpels at later developmental stages. Remnants of the aborted gynoecium are a feature of the male flower structure, and with good growing conditions, a few male flowers may not undergo complete carpel abortion and can form fruit. Arabidopsis thaliana is the closest relative to pawpaw in the Brassicales that also has sex chromosomes, and this species has been extensively studied by plant geneticists. Many genome parallels exist between the two species that indicate the pawpaw female sex determining gene is likely to be an upstream regulator of 2 genes called APETALA3 (AP3) and PISTILLATA (PI) that cause early abortion of stamens. A. thaliana also produces a gene called ATA1 that is expressed only in male flowers during development and is homologous to similar genes found in corn and white campion. ATA1 is associated with normal pollen formation and although it's probably also involved in the masculinizing pathway in pawpaws it might not be the initial gene that sets the whole flower development cascade underway. Approximately 180 genes are thought to be involved in producing fully functional flowers.

At least 2 genes differentiate the M and Mh chromosomes; one controls the long peduncle on male trees and the other is a masculinizing gene that controls carpel abortion in male flowers. As embryo abortion occurs 25-50 days after pollination, there is a regulatory gene that's essential to early embryo development that resides in this region and has degenerated on both the M and Mh chromosomes but is still functional in the X region; this is the cause of homozygous dominant allele lethality. This year (2015) it was reported that one of the key genes in pawpaw sex differentiation between males and hermaphrodites and not found on the X is similar to one found in A thaliana called SHORT VEGETATIVE PHASE (SVP) . SVP has been shown to initiate flower development, and in pawpaws it is functional in the MSY but not in the HSY. Thus it would seem that in males, SVP sets in train carpel abortion whereas in hermaphrodites a varying number of carpels remain functional and can therefore be fertilised and bear fruit. This whole field is advancing rapidly and the hope is that it won't be too long before all the key sex determining genes are identified and characterised. The evidence so far is that these few genes are the initial triggers for setting in train male, hermaphrodite or female development through subsequent genetic, epigenetic and plant hormone control.

Some examples of other practical developments and possibilities that have followed this genetic work include:

• Cloning of a single gene that codes for an enzyme called lycopene beta-cyclase which determines red fruit flesh colour. In many countries red flesh pawpaws are preferred over the more usual yellow/orange types and accordingly fetch higher prices. The allele for red colour, which produces a non-functional enzyme, is recessive, and this is why typically with open pollination this feature can be quickly lost and fruit revert to the less strongly coloured forms. Healthy pro-vitamin A carotenoids in red flesh fruit consist mainly of lycopene whereas in the yellow/orange forms with the dominant allele, almost all lycopene is enzymatically converted to other carotenoid forms.
• The accelerated impetus for genetic work on pawpaw stemmed in large part from the disease Papaya Ring Spot Virus (PRSV). This is a serious problem in all papaya growing regions worldwide and can devastate whole orchards. In Hawaii, pawpaw had been one of the top 3 fruit crops and they were threatened with total collapse of the industry because of the spread of the virus across all the islands. C. papaya is the only species in the genus and with a small genome there were limited opportunities to use classical breeding strategies to develop more resistance – none was sufficiently successful. Following sequencing of the PRSV and identification of the viral resistance gene that produced a coat protein, a major collaborative research effort was undertaken to develop a transgenic solution by inserting this gene into pawpaw. Sunup (red flesh) and Rainbow (yellow flesh) became the first such tree crop fruit to be approved for general release in the US and local growers quickly moved to incorporate them in production; subsequent approval in foreign markets such as Japan cemented their place as a valuable commercial crop and saved the industry. Other major pawpaw producers in Latin America and South East Asia have found PRSV infections in their regions are due to different strains, and with assistance and a technology transfer program are developing their own transgenic solutions.
• As genes are characterised, particularly those in the MSY and HSY, and their role in plant and fruit development clarified, there is the prospect of having pure breeding lines that avoid the multi-planting problem associated with conventional seedling propagation. As a number of other species in the Brassicales have already been cloned and have larger genomes with many more genes whose purpose has been identified, parallel genes can be seen in the C. papaya genome that regulate features such as precocity, stem height, photoperiod response, climacteric behaviour, fruit levels of sugar, starch, phytonutrients and aromatic volatiles, cellulose synthesis for tree strength, fruit firmness and size, storage properties, disease resistance etc. Exploitation of this expanding collective knowledge will support improved economic production of superior fruit with greater control and/or predictability under varying environmental conditions.

Barry Madsen