Click on author name to view affiliation information
Lantana camara, a member of the verbena family, is a popular ornamental yet highly invasive plant. It can escape from cultivation through seed dispersal and can contaminate native lantana species (Lantana depressa) through cross-pollination. Ploidy manipulation is being used as a genetic approach to produce sterile, noninvasive lantana cultivars. Polyploids have been observed in lantana (Lantana), but little information is available about the mechanisms for lantana polyploidization and the possible effects of natural polyploidization on the sterility (or fertility) of lantana triploids. In this study, we analyzed the ploidy level of more than 1500 lantana progeny from self, open, and/or controlled pollinations of 10 commercial cultivars and seven breeding lines. Our results confirmed the occurrence of unreduced gametes, specifically, unreduced female gametes (UFGs), in lantana. The frequency of UFG formation varied among commercial cultivars, and cultivars/breeding lines could be categorized into two groups: UFG producers and nonproducers. Tetraploid cultivars Gold, Pink Caprice, and Radiation fall into the UFG-producing group, while diploid cultivars Cream, Denholm White and Lola and tetraploid cultivars Carlos, Dallas Red and Irene belong to the nonproducer group. The frequency of UFG formation observed in nine UFG producers was 5.5% to 100%, varying with cultivar, growing condition, and/or pollination scheme. Progeny of the cross between ‘Carlos’ (seed parent) and ‘Gold’ (pollen parent) also showed the ability to produce UFGs, indicating that the trait (UFG formation) could be transmitted from ‘Gold’ to its progeny and is likely to be controlled by nuclear gene(s). Lantana triploids with or without the UFG-forming ability in its genetic background showed a significant difference in seed set: the former produced abundant seed when pollinated, while the latter produced little or no seed. The results stress the need to avoid using lantana with UFG-forming ability as parents in crosses designed to produce sterile triploids for invasiveness control. Additionally, the results from this study suggest multiple pathways for emergence and evolution of polyploids in cultivated and naturalized lantana populations.
Lantana is a member of the Verbenaceae and has been widely grown as a container plant, hanging basket plant, groundcover, hedge, or accent plant (Howard, 1969). It is popular in the nursery industry because of easy propagation and short production cycles, and is popular among gardeners because it attracts butterflies and is tolerant to droughts and poor soil conditions. Many nursery growers produce lantana, especially in the southern United States. For example, a survey of the Florida nursery industry, which consists of more than 5000 nurseries, indicates that 19% of the responding nurseries grew lantana and that the annual sales value in Florida alone was over $40 million (Wirth et al., 2004). The majority of the commercial lantana cultivars belong to Lantana camara. This species can escape from cultivation through seed dispersal and can invade agricultural and natural lands and hybridize with native lantana species (Lantana depressa). Because of these behaviors, L. camara has been listed as an invasive species in southern and central Florida (Florida Exotic Pest Plant Council, 2007). In several other countries including Australia, India, and South Africa, lantana is considered as a noxious weed or an invasive species (Sharma et al., 2005). Polyploid manipulation, particularly triploid production, has been proposed as a genetic approach to develop sterile, noninvasive lantana cultivars (Czarnecki et al., 2008). Similar genetic approaches (polyploid production and selection) are being used to sterilize other ornamental plants for invasiveness control (Ranney, 2004).
Polyploids are common in L. camara. Triploids, tetraploids, pentaploids, and hexaploids have been reported in cultivated and naturalized L. camara (Czarnecki et al., 2008; Natarajan and Ahuja, 1957; Spies and Stirton, 1982). Several lines of evidence suggest that polyploidization may be associated with the species’ invasive behavior (Sanders, 2001): Tetraploids are rare in the native populations of this species in tropical America, but are very common in the naturalized populations in Australia, India, and South Africa. Overall, tetraploids have courser leaves, grow more vigorously, and set more seeds (Sanders, 2001), whereas diploids tend to be dwarf and stunted. Tetraploids have a much wider range of distribution than diploids, and pentaploids are frequently found at high altitudes (Sinha and Sharma, 1984).
In plants, natural polyploidization can occur through somatic chromosome doubling or gametic nonreduction (Bretagnolle and Thompson, 1995). The former results from mitotic abnormalities in somatic (zygotic and meristematic) cells, while the latter results from meiotic abnormalities during gamete or gametophyte genesis and the formation of unreduced gametes (pollen and/or eggs). When an unreduced gamete unites with another unreduced gamete (bilateral) or with a normal haploid gamete (unilateral), the union leads to sexual polyploidization. To a certain extent, polyploidization via somatic chromosome doubling bears similarity to inbreeding, while polyploidization via unreduced gametes can retain heterozygosity. Thus, the two polyploidization processes can have significant differences in terms of genetic and evolutionary consequences to polyploidized species (Bretagnolle and Thompson, 1995; Hermsen, 1984).
Little information is available regarding the origin of polyploids in L. camara, except for a report by Khoshoo and Mahal (1967). The authors observed several tetraploids and a pentaploid in the open-pollinated (OP) progeny of a triploid and two hexaploids in the OP progeny of a pentaploid. They inferred that these tetraploid and hexaploid progeny with chromosome numbers higher than their parents must have come from union of unreduced female gametes with normally reduced pollen. The occurrence of UFGs has been reported in several plant species (Ramsey and Schemske, 1998; Stelly and Peloquin, 1986). Unreduced pollen has been reported in many plants and seems to be a more common mode of sexual polyploidization in plants (Bretagnolle and Thompson, 1995). Several studies have examined the pollen size and morphology of Lantana, but none of them reported the occurrence of unreduced pollen in L. camara (Raghavan and Arora, 1960; Sanders, 1987). In a recent pollen viability study, we examined tens of thousands of pollen grains and did not notice highly variable viable pollen grains within cultivars (D. Czarnecki and Z. Deng, unpublished).
During the course of interploid pollination and triploid generation, we observed pentaploid progeny from a tetraploid by diploid cross, which seems to indicate the occurrence of UFGs in the lantana cultivars used. Therefore, a study was undertaken to confirm the occurrence of UFGs in lantana, to determine its frequency and distribution in major commercial lantana cultivars, to test its transmissibility from generation to generation, and to determine if this trait would affect seed set on lantana triploids. Toward these objectives, progeny of commercial cultivars from self-pollination (SP) and OP were first analyzed for ploidy levels, followed by controlled pollinations among cultivars and breeding lines and ploidy analysis of their progeny. This report presents the results from these pollinations and ploidy analyses. Based on the results, we further propose a model for the origin of the multiple levels of polyploids in cultivated (and naturalized) L. camara and discuss the possible mechanisms for the formation of unreduced gametes in lantana and potential implications of this trait for lantana ploidy manipulation, particularly triploid generation and selection for sterile lantana development.
Ten commercial cultivars and seven breeding lines were used in this study (Table 1). Cuttings or rooted cuttings of these cultivars were generously provided by B. Dehgan (University of Florida, Gainesville) in 2004, and Robrick Nursery (Hawthorne, FL) in 2005 and 2006. Plants were grown in plastic containers filled with a commercial soilless mix, VerGro container mix A (Verlite, Tampa, FL) amended with controlled-release fertilizer, Osmocote (15N–3.9P–10K, 8 to 9 months release at 21 °C; Scotts, Marysville, OH) at 7.12 kg·m−3.
Three types of pollination were completed to generate as many seeds (progeny) as possible for ploidy analysis: SP, OP, and controlled pollinations (CP). In SP, lantana plants in 11.4-L containers were grown in a greenhouse with double doors and a thrips-proof screen. Insecticides were applied at regular intervals in the greenhouse. As lantana is autogamous (Rambuda and Johnson, 2004), hand pollination is not needed for SP. Therefore, to avoid damages to flowers and maximize seed set, flowers were allowed to self-pollinate without manual emasculation or hand pollination. In OP, lantana plants in 26.5-L containers were arranged on a raised metal bench outside in open air. Different cultivars were interspersed to maximize cross-pollination mediated by insect pollinators. Insecticides were not applied. Common insect pollinators observed were butterflies, moths, bees, and wasps. CP were done on lantana plants grown in the greenhouse. Mature flowers were emasculated by pulling off the entire calyx to which the anthers were attached and pollinated immediately using a camel brush with fresh pollen from flowers that just opened. Between pollinations, the brushes were soaked in 100% ethanol to kill residual pollen. Controlled pollinations were done from Apr. to June 2006, Aug. to Dec. 2006, Feb. to July 2007, June to Nov. 2008, and Feb. 2009.
When lantana fruit (berries) turned dark purple to black and became ripe, they were collected from SP, OP, or CP flower heads. Seeds were extracted, cleaned, and sown within 1 to 2 weeks after harvest, with the exception of the SP seeds of ‘Pink Caprice’, which were cleaned and stored for over 1.5 years before sowing.
Seeds were sown on the surface of a peat/vermiculite mix (VerGro container mix A) and germinated under intermittent mist in a greenhouse. Temperatures in the greenhouse ranged from 16 °C (night) to 30 °C (day), and no artificial lighting was used. The majority of seeds germinated within 4 months after sowing, but some took as long as 1 year to germinate. After young seedlings had developed true leaves, they were transferred to 12.7-cm-diameter plastic containers filled with VerGro container mix A. Plants were fertilized by incorporating Osmocote (15N–3.9P–10K) at 7.12 kg·m−3 in the soilless mix.
Analysis was performed on all plants using fully expanded young leaves and the Partec PA I ploidy analyzer and the CyStain ultraviolet Ploidy Precise P dye (Partec, Münster, Germany). The manufacturer-recommended ploidy analysis procedure was followed with minor modifications. The ploidy level of cultivars and progeny was determined by comparing to one or more commercial cultivars (reference cultivars) with known ploidy levels that were included in the same analysis.
The ploidy levels of the reference cultivars had been confirmed by counting chromosomes in root tip cells. Growing root tips of ‘Cream’, ‘Gold’, ‘Pink Caprice’, and ‘Radiation’ were collected from rooted cuttings, chilled at 8 °C overnight, and pretreated with 0.05% colchicine at ambient temperature for 4 h, fixed in Carnoy's fluid for 2 d, and stored in 70% ethanol at 4 °C. Fixed root tips were hydrolyzed in 1 N HCl at 60 °C for 5 to 10 min, squashed in acetic carmine on glass slides, and observed under a 100× objective on a microscope (BH-2; Olympus, Tokyo). Chromosome counting results showed that ‘Cream’ is a diploid (2n = 2x = 22), and ‘Gold’, ‘Pink Caprice’, and ‘Radiation’ are tetraploids (2n = 2x = 44).
One diploid (‘Lola’) and two tetraploid cultivars (‘Pink Caprice’ and ‘Gold’) were allowed to self-pollinate naturally in the greenhouse (without emasculation and hand-pollination). More than 900 seeds were collected and sown, and more than 500 progeny were analyzed for ploidy levels. All 103 progeny of ‘Lola’ were diploids, indicating no UFG formation in this cultivar. The majority (88.3% and 92.3%) of the progeny of tetraploid cultivars Pink Caprice and Gold were tetraploids, as expected for normal n (2x) gamete formation and fertilization (Table 2), but 5.5% of ‘Gold’ SP progeny and 11.7% of ‘Pink Caprice’ SP progeny were hexaploids (Table 3). The occurrence of these hexaploids indicates that a 2n gamete (4x) had been formed, combined with an n gamete (2x), and developed into a hexaploid embryo. Additionally, two diploids were recovered from ‘Gold’ SP seeds, but no diploids were identified in the SP progeny of ‘Pink Caprice’, even though much more progeny (311 of ‘Pink Caprice’ vs. 91 of ‘Gold’) were examined. This is the first time that such ploidy level reduction from a tetraploid to a diploid has been observed in lantana. Understanding the origin of these diploids will require the use of cytological and molecular genetic analysis tools, which are not yet available in lantana.
The above results clearly indicate the formation of 2n gametes in ‘Gold’ and ‘Pink Caprice’, but do not elucidate which side (maternal or paternal) produced the 2n gametes.
To confirm the above findings and to determine if 2n gametes occur in other lantana cultivars, OP seeds (more than 2800 in total) were collected from ‘Lola’, ‘Pink Caprice’, and ‘Gold’, as well as another two diploids (‘Cream’ and ‘Denholm White’) and three more tetraploids (‘Carlos’, ‘Dallas Red’, and ‘Irene’) grown in a large screen house. If 2n gametes were not present, it would be expected that most OP progeny of a diploid cultivar would be diploid (from SP or CP by other diploid cultivars) and some progeny would be triploid (from CP by tetraploid cultivars) (Table 2). Similarly, if 2n gametes did not occur, most OP progeny of a tetraploid cultivar would be tetraploids (from SP or CP by other tetraploid cultivars) and some progeny would be triploids (from CP by diploid cultivars) (Table 2). On the other hand, should 2n gametes have been formed on the seed parent side, some OP progeny of a diploid cultivar would be expected to be tetraploids (from 2n gametes of the diploid seed parent and the normal gametes of tetraploid pollen parents), and some OP progeny of a tetraploid cultivar would be pentaploids and hexaploids (from 2n gametes of the tetraploid seed parent and the normal n gametes of diploid and tetraploid pollen parents) (Table 2).
A limited number of seeds (and progeny) were obtained from ‘Cream’ and ‘Denholm White’ due to their very low female fertility, especially in ‘Denholm White’, where 2472 flowers were pollinated from five crosses and 0.04% seed set (one seed) was observed. All their progeny were diploids, indicating a possible lack of 2n gamete formation in these cultivars (more supporting data below).
The majority (96.4%) of ‘Lola’ OP progeny were diploids and 2 of 55 progeny were triploids (Table 3). In an OP environment and with diploids and tetraploids present, these two triploids could have resulted from normal gamete formation and fertilization, or 2n gamete formation and fertilization (Table 2). To determine the origin of the two triploids, we examined the leaf and flower morphology of these triploids, as well as all the potential diploid and tetraploid parents present at the time when OP was conducted. The most obvious difference was in flower color: all diploids had yellow, creamy yellow, or white flowers, while all tetraploids had pink, magenta, or red flowers. Our lantana flower color inheritance studies indicated that crosses among yellow, creamy yellow, or white flowers would result in progeny with shades of yellow to white, and crosses between yellow, creamy yellow, or white flowers and pink, magenta, or red flowers would result in progeny with shades of pink or red (D. Czarnecki and Z. Deng, unpublished). Flowers of the two OP triploids of ‘Lola’ were in shades of magenta, indicating a possible origin from union of normal n (x) female gametes of ‘Lola’ (yellow) with n (2x) male gametes of tetraploids (pink or magenta). Thus, 2n gametes were probably not involved in the formation of the triploid progeny of ‘Lola’.
Four ploidy levels (3x, 4x, 5x, and 6x) were observed in 14 ‘Pink Caprice’ OP progeny (Table 3). The presence of 5x and 6x progeny indicates the occurrence of 2n gametes in ‘Pink Caprice’ (Table 2). Further, the presence of 5x suggests that, more specifically, UFGs occurred in ‘Pink Caprice’ (Table 2). The percentage of 5x and 6x progeny was 35.7%, a much higher frequency of 2n gamete formation than observed when ‘Pink Caprice’ was grown in the greenhouse and self-pollinated. Formation of 2n gametes was observed again in ‘Gold’, as indicated by the presence of 11 pentaploids and 4 hexaploids in the 81 ‘Gold’ OP progeny (Table 3). Similarly, the frequency of 2n gamete formation in ‘Gold’ was much higher in the OP progeny than in the SP progeny (14.3% vs. 5.5% in the SP progeny). One diploid appeared in ‘Gold’ OP progeny, and its causes remain to be understood.
In ‘Carlos’ and ‘Dallas Red’, the majority of the 77 or 50 OP progeny were tetraploids and a few were triploids. The absence of pentaploids or hexaploids indicates a lack of 2n gamete formation in these tetraploid cultivars.
OP seeds were also collected from a commercial cultivar grown at a hotel site in Pittsburgh, PA. This unknown Pittsburgh lantana (UPL) and its progeny were tetraploid or hexaploid, suggesting a high frequency (57.7%) of 2n gamete formation in this cultivar (Table 3).
These results from ploidy analysis of OP progeny confirm the results from SP progeny analysis. They also suggest considerable differences among lantana cultivars in 2n gamete formation and the nonreduction is likely to occur on the maternal side.
A total of over 2000 seeds were sown from CP. ‘Carlos’ (4x) produced tetraploid progeny when pollinated with ‘Dallas Red’ (4x) and produced triploids (214) when pollinated with ‘Denholm White’ (2x), ‘Lola’ (2x), or a diploid breeding line LAOP-9 (progeny of ‘Lola’) (Table 4). These results support the above-mentioned observation in SP and OP progeny: ‘Carlos’ did not produce 2n gametes. It seems reasonable to make a similar conclusion for ‘Dallas Red’, as it produced only tetraploid progeny when pollinated with ‘Carlos’ and produced only triploids when pollinated with ‘Denholm White’, ‘Lola’, or LAOP-9 (Table 4). When pollinated with three tetraploids (‘Carlos’, ‘Dallas Red’, and ‘Irene’), ‘Lola’ produced triploids, and when pollinated with diploid LAOP-9 or selfed, ‘Lola’ produced diploids (Table 4). ‘Cream’ produced triploids when pollinated with ‘Carlos’, ‘Dallas Red’, or ‘Irene’, which indicates that ‘Cream’ did not form 2n gametes and supports the observations from SP and OP progeny (Table 4).
Although 1137 flowers of ‘Gold’ were pollinated with ‘Carlos’ or ‘Dallas Red’ in 2007 and 2008, only 20 seeds were obtained and only 15 germinated for ploidy analysis. The high percentages of hexaploids (60% and 100%) appearing in these progeny indicate high percentages of 2n gamete formation, more specifically, UFG formation, in ‘Gold’ (Table 4), because, as mentioned previously, ‘Carlos’ and ‘Dallas Red’ were shown not to produce 2n gametes from controlled pollinations. When ‘Carlos’ was pollinated with ‘Gold’, all the progeny were tetraploids (Table 4). This contrasts with the ploidy level distribution in progeny of ‘Gold’ × ‘Carlos’ and indicates a lack of 2n pollen formation in ‘Gold’ and a lack of UFG formation in ‘Carlos’. ‘Pink Caprice’ produced pentaploids and hexaploids when pollinated with diploid ‘Denholm White’ or tetraploid ‘Radiation’, respectively, which supports the formation of UFGs in ‘Pink Caprice’ (25%–100%) (Table 4).
To gain understanding of the transmission of this 2n gamete formation trait, progeny of ‘Gold’ and ‘Pink Caprice’ were analyzed for ability to form UFGs. Two ‘Gold’ progeny were selected because of the availability of flowers for pollination. Progeny 604–1 (P604–1, 4x) was from a CP between ‘Carlos’ (non-2n gamete producer) and ‘Gold’ (UFG producer). P604–1 produced hexaploids at high frequencies in its progeny when self- or open-pollinated (Table 5). This indicates that P604–1 was able to produce 2n gametes or it had inherited the trait from ‘Gold’. Transmission of the trait from a paternal parent to progeny also suggests that the trait may be under nuclear gene control. Progeny GDGHOP-36 was a diploid progeny of self-pollinated ‘Gold’. When pollinated with ‘Lola’ (2x, non-2n female gamete producer), GDGHOP-36 produced triploids only (100%; Table 5), indicating 2n gamete formation in GDGHOP-36, another line of evidence showing the transmission of 2n gamete formation from ‘Gold’ to its progeny.
The transmission of 2n gamete formation from ‘Pink Caprice’ to its progeny was examined using PCOP-6 (4x), a progeny of ‘Pink Caprice’ and PKGHOP-1 (2x), a second-generation progeny of ‘Pink Caprice’. PCOP-6 produced 2n female gametes, as shown by the occurrence of 5x progeny when pollinated with LAOP-9 (2x) (Table 5). PKGHOP-1 produced 4x progeny when pollinated with DROP-25, a 4x OP progeny of ‘Dallas Red’ (Table 5), indicating that PKGHOP-1 also carried the UFG formation trait from ‘Pink Caprice’.
Three 3x breeding lines with or without 2n gamete formation background were pollinated with one common cultivar ‘Lola’ (2x) (Table 6). Breeding lines 605–35 and 624–1 were from ‘Carlos’ and ‘Dallas Red’, respectively, and both did not carry the UFG trait. Their seed set ranged from 0 to 1%. Breeding line GDOP-4 carried the UFG formation trait from ‘Gold’, and its seed set was 25%.
In the 2n-gamete-producing cultivars, the frequency of 2n gamete formation varied widely, from 5.5% to 100.0% in ‘Gold’ and from 11.3% to 100.0% in ‘Pink Caprice’ (Table 1). Similar variation seems common in other plants and has been noted among individuals, flowers of an individual, and even within different anthers of a single flower (McCoy, 1982; Veilleux et al., 1982). It has been documented that the frequency of 2n gamete formation in a plant may be influenced by genotype, environment, and their interactions. Temperatures, in particular, seem to have a strong impact on the production of 2n gametes (McHale, 1983). Additional important factors noted in previous studies are variable degrees of penetrance and expressivity in the genes responsible for 2n gamete formation (Mok and Peloquin, 1975; Watanabe and Peloquin, 1989). In this study, except for the variable frequencies, a cultivar's 2n gamete formation activity remained quite consistent under different pollination schemes (OP, SP, or CP) and under different growing conditions (shade house and greenhouse): 2n-gamete producers always produced some 2n gametes, and non-2n-gamete producers did not.
Two basic processes have been described as the mechanisms for 2n gamete (pollen and egg) formation in plants: first division restitution (FDR) and second division restitution (SDR) (see review by Bretagnolle and Thompson, 1995). In FDR, the first meiotic division occurs abnormally, with homologous chromosomes not pairing during prophase I and/or not separating to opposite poles during anaphase I. In SDR, the second meiotic division occurs abnormally, with sister chromatids not separating to opposite poles during anaphase II. It has been shown that univalents and multivalents (during prophase I) and laggards (during anaphase I) are very common in lantana polyploids (Natarajan and Ahuja, 1957; Spies and Stirton, 1982). It is expected, therefore, that 2n gametes formed through SDR would be aneuploids and contain various numbers of chromosomes. The progeny of ‘Gold’ and ‘Pink Caprice’ resulting from UFGs were pentaploids or hexaploids (Tables 3 and 4). Similarly, the progeny of ‘Red Cap’ (3x) were tetraploids or pentaploids, and the progeny of ‘Purple Prince’ (5x) were hexaploids (Khoshoo and Mahal, 1967). These results suggest that the UFGs leading to the production of these higher level polyploids should contain all the chromosome complements in the respective seed parents. This would be possible only through FDR, and the meiotic abnormalities occurred before or during the first division and all the unpaired (univalents) as well as paired homologous chromosomes (bivalents, trivalents, and quadrivalents) did not separate during anaphase I. To confirm or refute the above inference, we are developing simple sequence repeat markers that can be used for segregation analysis and for assessing the levels of heterozygosity in the progeny of 2n-gamete-producing cultivars. This will help determine the stage of nonreduction.
In the present study, diploids produced triploids and tetraploids, and tetraploids produced pentaploids and hexaploids through UFG formation and fertilization with n male gametes (Tables 3 and 4). In a previous study, a triploid produced tetraploid and pentaploid progeny, and a pentaploid produced hexaploid progeny through the same process (Khoshoo and Mahal, 1967). As summarized in Fig. 1, these results suggests that all the observed polyploid levels in L. camara (triploids to hexaploids) can be evolved through UFG formation and fertilization with normal n male gametes, and that there exist two or more pathways for polyploidization at each of the ploidy levels in L. camara. It is conceivable that the existence of multiple polyploidization pathways could lead to an increased diversity (or complexity) in L. camara.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 134, 5; 10.21273/JASHS.134.5.560
One interesting phenomenon in L. camara has been the lack of correlation between meiotic irregularities during microsporegenesis and seed set in naturalized polyploids. In spite of high frequencies of meiotic irregularities in pollen mother cells, lantana polyploids produced abundant viable seeds (Natarajan and Ahuja, 1957; Raghavan and Arora, 1960). Khoshoo and Mahal (1967) suggested that apoximis was responsible for the seed set in these highly male-sterile polyploids, but Spies and Stirton (1982) could not find definitive evidence for apomixis in embryo sac examination. The finding that lantana can produce UFGs may provide an explanation for such a lack of correlation in polyploids. Without undergoing the normal meiotic first division during megasporegenesis, polyploids, even though showing high levels of meiotic abnormalities during microsporegenesis in their pollen mother cells, could still produce megaspores and egg cells that contain all the chromosome complements, which would be viable and able to be fertilized and produce viable seeds.
Sterile cultivars are needed to control the invasive potential L. camara has shown in southern United States and other countries. Compared with diploids and tetraploids, triploid lantana plants show the lowest pollen stainability and seed set (Czarnecki et al., 2008). Production of triploids has been proposed as a main genetic approach to develop new sterile lantana cultivars. Results from the present study suggest the possibility of producing triploids through crosses between diploids and tetraploids, and through crosses between diploids as well (Fig. 1). More importantly, the results indicate a strong need to screen and select breeding parents carefully and to avoid using UFG-producing plants in crosses intended for high sterility. If triploids have inherited the ability to form UFGs from seed or pollen parents, these triploids will be likely to produce viable UFGs (3x), which can be fertilized with pollen from surrounding diploids or tetraploids and produce viable seeds, which will defy the purpose of producing such triploids. None of the diploids (‘Cream’, ‘Denholm White’, and ‘Lola’) in this study showed tendency to produce UFGs, but four of the seven tetraploids (‘Gold’, ‘Pink Caprice’, ‘Radiation’, and UPL) showed such tendency. To find more tetraploids that do not produce UFGs and can be used as breeding parents, screening of more commercial cultivars or germplasm and progeny through ploidy analysis will be required.
The above results showed the occurrence of 2n female gametes and revealed characteristics of this phenomenon in lantana. Frequencies of UFG formation varied among cultivars and breeding lines. It was common in ‘Gold’, ‘Pink Caprice’, ‘Radiation’, and UPL, but it did not occur in ‘Carlos’, ‘Dallas Red’, and ‘Irene’. The UFG-forming trait can be transmitted from one generation to the next, and it seems to be under nuclear gene control. This characteristic should facilitate hybridization between different ploidy levels, and thus transference of morphological or physiological traits from a diploid background to a higher ploidy background for cultivar development. Conversely, the trait could increase triploids’ seed set, thus becoming a barrier to producing new sterile, triploid Lantana cultivars.
Contributor Notes
This project was funded in part by the Florida Nursery Growers and Landscape Association, the Southwest Florida Water Management District, and the USDA/CSREES/TSTAR (United States Department of Agriculture/Cooperative States Research, Extension and Education Service/Tropical and Subtropical Agriculture) program.
We thank Joyce Jones, Gail Bowman, and Sarah Smith for their technical assistance, and Dr. Jude W. Grosser, J.L. Chandler, and Patricia Brickman for ploidy analysis of lantana commercial cultivars in 2004.
Corresponding author. E-mail: zdeng@ufl.edu.
