Parent material.

Parents were collected from nurseries, gardens, and arboreta from 2009 to 2014 (Table 2) that provided cultivar and trademark names. Full scientific names, cultivars, and trademarks are reported (Table 2), but for simplicity only market names (cultivar or trademark) are used hereafter. Taxonomic designations reflect current phylogenies and revisions, including the use of subspecies designations in Pubescentes (Chen et al., 2009). Representative species and cultivars were obtained from series Syringa, Pubescentes, and Villosae focusing on elite cultivars improved for ornamental traits including flower colors and forms, leaf pigments, and novel growth habits including dwarf habits.

Table 2.

Source material for lilac breeding population.

View Table

Flower colors included white, pink, blue, and purple, with one taxon, S. vulgaris ‘Sensation’, having picotee flowers in which the petal edges lack pigment. Flower forms included single and double flowers, with some exhibiting hose-in-hose flowers. Double flowers in lilac often represent a case of neoheterotrophy where additional floral whorls lead to supernumerary petals (Dadpour et al., 2011). Double flowers can also arise from mutations leading to petaloid sepals (Fiala, 1988). Both cases leave reproductive whorls intact and allow for double-flowered cultivars to be used in reciprocal crosses. Foliar pigments were rare across the breeding population, occurring in a spring flush of yellow or purple leaves (Fig. 1). Syringa emodii in series Villosae was the only taxon that produced a yellow flush of leaves (Fig. 1A). Purple color of spring foliage was limited to purple-flowered taxa in series Syringa, but was most pronounced in S. ×hyacinthiflora ‘Old Glory’ (Fig. 1B). Novel variations in form were limited to two dwarf taxa in series Syringa, S. vulgaris ‘Tiny Dancer’ and S. vulgaris ‘Prairie Petite’.

Spring leaf color phenotypes in lilac breeding population: (A) yellow-green leaves of Syringa emodii, (B) purple leaves of Syringa ×hyacinthiflora ‘Old Glory’, (C) hybrid seedling (H2013-150-001) from the cross Syringa vulgaris ‘President Grévy’ x Syringa vulgaris ‘Sensation’, and (D) hybrid seedlings from the cross Syringa vulgaris ‘Ludwig Spaeth’ x Syringa vulgaris ‘Angel White’ segregating for abaxial leaf pigment.

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 142, 4; 10.21273/JASHS04155-17

• Download Figure

View Full Size

Spring leaf color phenotypes in lilac breeding population: (A) yellow-green leaves of Syringa emodii, (B) purple leaves of Syringa ×hyacinthiflora ‘Old Glory’, (C) hybrid seedling (H2013-150-001) from the cross Syringa vulgaris ‘President Grévy’ x Syringa vulgaris ‘Sensation’, and (D) hybrid seedlings from the cross Syringa vulgaris ‘Ludwig Spaeth’ x Syringa vulgaris ‘Angel White’ segregating for abaxial leaf pigment.

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 142, 4; 10.21273/JASHS04155-17

• Download Figure

Spring leaf color phenotypes in lilac breeding population: (A) yellow-green leaves of Syringa emodii, (B) purple leaves of Syringa ×hyacinthiflora ‘Old Glory’, (C) hybrid seedling (H2013-150-001) from the cross Syringa vulgaris ‘President Grévy’ x Syringa vulgaris ‘Sensation’, and (D) hybrid seedlings from the cross Syringa vulgaris ‘Ludwig Spaeth’ x Syringa vulgaris ‘Angel White’ segregating for abaxial leaf pigment.

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 142, 4; 10.21273/JASHS04155-17

• Download Figure

Crosses.

During the spring and summer of 2013, 2014, and 2015, a total of 27,645 cross-pollinations were made. Of these, 114 crosses were performed with an average of 243 ± 27 flowers pollinated per cross. For each series, three types of crosses were attempted: intraspecific, interspecific, and interseries (Tables 3–5, respectively). Within each series, crosses made with at least one interspecific parent (e.g., S. ×hyacinthiflora x S. oblata) were classified as interspecific. Most cultivars in Villosae were interspecific hybrids; each replicate of an unimproved species was collected from a single source. For example, we received S. emodii from a single Index Seminum source. Consequently, crosses with Villosae focused on interseries and interspecific combinations because intraspecific crosses would likely involve significant inbreeding.

Table 3.

Intraspecific cross-compatibility within series Pubescentes and Syringa in lilac.

View Table

Table 4.

Interspecific cross-compatibility within series Pubescentes, Syringa, and Villosae in lilac.

View Table

Table 5.

Interseries cross-compatibility among series Pubescentes, Syringa, and Villosae in lilac.

View Table

Each year, fresh pollen was collected and stored in small petri dishes over desiccant (Drierite; W.A. Hammond Drierite, Xenia, OH) in a refrigerator at 4 °C. Before pollination, open flowers were removed on all inflorescences and saved in glassine bags for pollen collection and reciprocal crosses. Individual flowers were emasculated before anthesis and pollinations were made in a glasshouse free of pollinators with day/night temperatures of 25/20 °C and a 16-h photoperiod. Each flower was pollinated using a brush two to three times postemasculation over consecutive days. Brushes were sterilized before and between pollinations using 70% ethanol. Incidences of self-pollination were tested by covering two inflorescences per plant with organza bags and shaking occasionally. Self-pollination can also occur if pollen is released during emasculation. Self-pollination during emasculation was tested by emasculating multiple inflorescences (200+ flowers) on two fertile parents (S. vulgaris ‘Angel White’ and S. vulgaris ‘Ludwig Spaeth’) and covering the inflorescences with organza bags. Developing fruit were counted throughout the summer and dry fruit were collected before dehiscence during fall. Data were collected on number of pollinated inflorescences, pollinated flowers, and developing fruit. During fall, seed were cleaned and counted before cold stratification during winter.

Seed germination.

Seed were placed in plastic bags filled with moist stratification media consisting of half perlite (Supreme Perlite Co., Portland, OR) and half growing medium (Metro-Mix; Sun Gro Horticulture, Agawam, MA). Seed were cold-stratified for 10 weeks at 4 °C. After stratification, seed were sown in 1.3-L containers filled with the growing medium and treated once with copper hydroxide (Kocide 2000; DuPont, Wilmington, DE) at 0.3 mg·L⁻¹. For each cross, no more than 30 seed were sown per pot. Seedlings were germinated in a glasshouse under the conditions described previously. Data collected over winter included number of germinated seed, albino seedlings, and viable green seedlings.

In vitro germination.

Capsules progressed from pollination to dehiscence over a 20- to 30-week period in the glasshouse. In 2013, observations of early fruit abortion at 6 weeks postpollination in the interseries cross S. oblata x S. pubescens Bloomerang^® Purple prompted an in vitro germination trial on a subset of developing fruit (Table 6). In subsequent years, fruit were allowed to dehisce early, and seed were sown according to the methods previously mentioned. For in vitro germination, interseries hybrid fruit from five crosses (Table 6) were collected 7 weeks after pollination. Green fruit were collected and immediately surface-sterilized by rinsing in a 70% ethanol solution for 30 s followed by a soak in a 6.15% (v/v) sodium hyporchlorite solution with several drops of surfactant (Tween 20; Acros Organics, Fair Lawn, NJ). Fruit were triple rinsed and temporarily stored in filter-sterilized, autoclaved water. Fruit were dissected in a sterile, laminar flow hood using a dissecting microscope. Green seed were removed from capsules into sterile petri dishes containing an aqueous solution of l-ascorbic acid at 25 mg·L⁻¹ to reduce oxidation. Embryo extraction at this early stage proved too damaging to young tissues. Therefore, intact dissected seed were cultured on an embryo rescue medium.

Table 6.

Attempted pollinations, recovered seed, and in vitro germination of interseries lilac hybrids in 2013. All seed collected from green capsules and cultured on cultured on Monnier’s medium as described by Zhou et al. (2003).

View Table

Dissected seed were cultured on Monnier’s medium according to the lilac embryo rescue protocol of Zhou et al. (2003) and incubated under standard culture conditions (24 ± 2 °C and a 16-h photoperiod of 60 μmol·m⁻²·s⁻¹ provided by cool-white fluorescent lamps). Seed were incubated on 10 mL of the embryo rescue medium in 150-mm culture tubes. Tubes were capped and sealed with paraffin film (parafilm; American National Can Co., Menasha, WI). A total of 161 seed were placed on the germination medium, representing five interseries crosses. Culture tubes were completely randomized and maintained in racks of 40 tubes. Germination was observed and recorded over 3 months. On germination, seedlings and callus were transferred to the shoot regeneration medium composed of Murashige and Skoog basal salts and vitamins, 5 μM BAP, 0.5 μM indole-3-butyric acid (IBA), 100 mg·L⁻¹ myo-inositol, 100 mg·L⁻¹ 2-(N-morpholino) ethanesulfonic acid (MES monohydrate), and 30 g·L⁻¹ sucrose. The solution containing basal salts, vitamins, and hormones (Phytotechnology Laboratories, Shawnee Mission, KS) was adjusted to pH 5.8 and solidified with 7.5 g·L⁻¹ agar (Sigma-Aldrich, St. Louis, MO).

Intraspecific hybridization.

Within series Pubescentes, a total of 2180 pollinations resulted in an average of 17.18 ± 11.75 viable seedlings per cross. However, only four crosses yielded viable seedlings of the 11 attempted. The most prolific cross was between two remontant taxa, S. pubescens Josee × S. pubescens Bloomerang^® Purple, which produced 131 viable seedlings at 0.53 seedlings per pollinated flower (Table 3). The reciprocal cross produced 20 viable seedlings at 0.14 seedlings per pollinated flower (Table 3). Syringa pubescens Josee was also an effective seed parent in crosses with S. pubescens Tinkerbelle^®, yielding 28 viable seedlings at 0.19 seedlings per pollinated flower (Table 3). The fewest viable seedlings were produced from S. pubescens Tinkerbelle^® x S. pubescens Bloomerang^® with only 10 viable seedlings produced at 0.03 seedlings per pollinated flower (Table 3).

Within series Syringa, a total of 3209 pollinations resulted in an average of 24.24 ± 11.28 viable seedlings per cross. Of the 17 crosses attempted, eight yielded viable seedlings. The most prolific was S. vulgaris ‘Ludwig Spaeth’ x S. vulgaris ‘Angel White’ producing 186 seedlings at 0.68 seedlings per pollinated flower (Table 3). Of the dwarfs, S. vulgaris ‘Prairie Petite’ was the smallest and slowest growing, producing few inflorescences each year. One cross performed with S. vulgaris ‘Prairie Petite’ used 74 pollinations with S. vulgaris ‘Sensation’ and yielded 33 capsules but no seed (Table 3). The dwarf S. vulgaris Tiny Dancer was used successfully as both a seed and pollen parent in intraspecific crosses. As a seed parent, S. vulgaris Tiny Dancer was compatible with S. vulgaris ‘Sensation’ at 0.14 seedlings per pollinated flower, whereas the reciprocal cross yielded 0.21 seedlings per pollinated flower. As a pollen parent, S. vulgaris Tiny Dancer was compatible with S. vulgaris Blue Skies^® at 0.61 seedlings per pollinated flower and S. vulgaris ‘Lavender Lady’ at 0.27 seedlings per pollinated flower (Table 3). Syringa vulgaris ‘President Grévy’ had few successful crosses. As a seed parent, 807 intraspecific pollinations were performed with four taxa, and only one seedling was produced from a cross with S. vulgaris ‘Sensation’ (Table 3). As a pollen parent, 740 pollinations were performed with S. vulgaris ‘President Grévy’ on three taxa. Only S. vulgaris Blue Skies^® proved an effective seed parent, with 12 seedlings produced at 0.05 seedlings per pollination (Table 3).

Interspecific hybridization.

Within series Pubescentes, interspecific crosses were performed between cultivars of S. pubescens and S. meyeri ‘Palabin’. A total of 2649 pollinations resulted in an average of 96.29 ± 37.12 viable seedlings per cross (Table 4). Of the seven crosses, two failed to produce viable seedlings. As a seed parent, S. pubescens ‘Miss Kim’ yielded 149 nonviable, albino seedlings, whereas reciprocal crosses yielded no seedlings after 522 pollinations (Table 4). The most prolific cross, S. meyeri ‘Palabin’ x S. pubescens Bloomerang^® Purple, yielded 278 seedlings at 1.16 seedlings per pollinated flower (Table 4).

Within series Syringa, interspecific crosses were performed among taxa of S. vulgaris, S. ×hyacinthiflora, and S. oblata. A total of 3518 pollinations resulted in an average of 31.71 ± 10.93 viable seedlings per cross (Table 4). Of the 24 crosses, 15 resulted in viable seedlings. The most prolific cross was S. vulgaris Blue Skies^® x S. ×hyacinthiflora ‘Betsy Ross’ producing 1.41 seedlings per pollinated flower (Table 4). Although S. oblata was successful in a number of interspecific crosses, the white-flowered S. oblata var. alba produced large seedling populations only in crosses with S. vulgaris Tiny Dancer. This cross produced 68 viable seedlings at 0.40 seedlings per pollination, whereas the reciprocal cross produced 47 seedlings at 0.23 seedlings per pollination (Table 4).

Not surprisingly, cultivars of S. ×hyacinthiflora, crossed successfully with cultivars of S. vulgaris and wild-type S. oblata. Crosses with S. ×hyacinthiflora ‘Old Glory’ were of interest for future breeding because of the lack of flower and form diversity in purple-leaved cultivars. Syringa ×hyacinthiflora ‘Old Glory’ proved a successful seed parent in crosses with white-flowered S. vulgaris ‘Angel White’, yielding 37 seedlings at 0.19 seedlings per pollinated flower (Table 4). Despite 176 pollinations and 76 recovered seed, S. ×hyacinthiflora ‘Old Glory’ produced no viable seedlings with the double-flowered S. vulgaris ‘President Grévy’. When crossed with the dwarf S. vulgaris Tiny Dancer, S. ×hyacinthiflora ‘Old Glory’ proved an efficient seed parent producing 37 seedlings at 0.30 seedlings per pollination. The reciprocal cross was even more efficient, yielding 160 seedlings at 0.96 seedlings per pollination (Table 4).

Within series Villosae, only three of 14 crosses produced seedlings. A total of 2997 pollinations yielded an average of 9.29 ± 7.19 seedlings per cross combination (Table 4). The most prolific cross was S. julianae x S. ×prestoniae ‘Miss Canada’ which produced 100 viable seedlings at 0.48 seedlings per pollinated flower (Table 4). In addition, S. julianae had some of the largest flowers in series Villosae with a fragrance reminiscent of S. vulgaris. The only other seed parent to produce viable seedlings in interspecific crosses in series Villosae was S. wolfii. The cross between S. wolfii and the yellow-leaved S. emodii (Fig. 1A) produced 24 seedlings at 0.11 seedlings per pollinated flower (Table 4). Only six seedlings resulted from crosses between S. wolfii and S. villosa yielding 0.03 seedlings per pollinated flower (Table 4).

Interseries hybridization.

Interseries crosses proved the most challenging because of differences in bloom time. It took 3 years to complete a range of interseries crosses. A typical lilac will bloom for 6 weeks in an average season with reliable patterns of bloom across series and species (Fiala, 1988). Our observations agreed with Fiala (1988) with members of series Syringa blooming first in spring, followed by series Pubescentes in late spring to early summer and series Villosae in early to midsummer. During the first year, pollen was collected from early-blooming plants and the sequence of blooming was noted across all taxa. Bloom data were used to design reciprocal interseries crosses over the following 2 years where temperature was altered to hasten or slow bloom. Earlier bloom times were induced in late-blooming taxa using a heated glasshouse. Delayed bloom times were induced in early-blooming taxa using a walk-in cooler.

For series Pubescentes, Syringa, and Villosae, there were six possible combinations of interseries crosses (Table 5). A total of 41 crosses were performed across these six combinations representing 13,092 pollinations (Table 5). No viable seedlings were recovered although some crosses produced capsules and seed (Table 5). Across all the interseries combinations, 975 capsules were recovered which produced 480 seed. Seed-producing crosses could provide a foundation for future studies on embryo abortion and embryo rescue.

Interseries crosses between Pubescentes and Villosae included seven crosses that produced capsules and seed. These crosses included S. emodii and cultivars of S. pubescens and S. ×prestoniae (Table 5). Capsules and seed were produced when taxa in Pubescentes and Villosae were used as seed parents. The most prolific cross was S. ×prestoniae ‘Redwine’ × S. pubescens Josee which produced 129 seed from 174 capsules after 602 pollinations (Table 5).

Interseries crosses between Pubescentes and Syringa included five crosses that produced capsules and seed. Only one cross produced seed with series Syringa as a seed parent, S. vulgaris ‘Ludwig Spaeth’ x S. pubescens Bloomerang^®. This cross resulted in 27 capsules and 18 seed after 2206 pollinations (Table 5). Four crosses produced seed using series Pubescentes as a seed parent. The cross S. oblata x S. pubescens Bloomerang^® Purple produced the most capsules of any interseries cross at 238 capsules from 547 pollinations. However, these capsules aborted 6 weeks after pollination, unlike most interseries capsules which persisted for the 20- to 30-week development period. The most prolific cross between series Pubescentes and Syringa was S. pubescens Josee x S. oblata, yielding 77 seed from 138 pollinations at 0.56 seed per pollination (Table 5).

No interseries cross between series Syringa and Villosae produced seed-filled capsules. Of the 2016 pollinations, capsules were produced from three crosses: S. ×hyacinthiflora ‘Old Glory’ x S. villosa, S. ×hyacinthiflora ‘Old Glory’ x S. wolfii, and S. vulgaris ‘Sensation’ x S. ×prestoniae ‘Miss Canada’ (Table 5). The most prolific cross was S. ×hyacinthiflora ‘Old Glory’ x S. villosa which yielded 46 capsules from 176 pollinations (Table 5).

In vitro germination.

A preliminary trial extracting open-pollinated lilac seed and embryos from green capsules revealed the difficulty of embryo extraction at this young stage. Oxidation of the young seed progressed rapidly during excision and phenolics proved damaging when immature seed were cultured in vitro. Excising seed while submerged in an antioxidant solution reduced oxidation and allowed extraction of undamaged green seed. Immature seed obtained from interseries crosses in 2013 had low germination in vitro on Monnier’s medium (Table 6). Most seed failed to germinate and eventually became necrotic. Of the 161 seed cultured in vitro, only three germinated, all from the cross S. pubescens ‘Penda’ Bloomerang^® Purple x S. vulgaris ‘Ludwig Spaeth’ (Table 6). Seedlings failed to grow post germination and tissues, including cotyledons, subsequently converted to callus. This result may be due to lack of proper transfer media postgermination or lack of proper combination of genotype and in vitro germination medium. Surprisingly, Zhou et al. (2003) did not report the genotype used in their protocol. Further research will be necessary to design protocols for in vitro seed germination and embryo rescue in lilac. Callus obtained from cotyledons could provide source material for somatic embryogenesis in future studies, as demonstrated by Liu et al. (2013).

Though interseries crosses and in vitro germination failed to achieve hybrids, the quantity of seed produced combined with several seed that germinated in vitro provides evidence that future work on wide hybridization in lilacs may prove fruitful. Anatomical studies have shown that low-germination lilac seed contain embryos that progress to walking stick stage before abortion (Chen et al., 2012). This study lists individual cross combinations between series Pubescentes and Villosae, as well as Pubescentes and Syringa, which produced large numbers of seed (Table 5). In addition, current research at North Dakota State University also observed fruit development in interseries crosses between series Villosae and the tree lilacs in Ligustrina that persisted well into the summer before fruit abortion (N.G. Maren, personal communication).

This study represents a comprehensive investigation of lilac cross-compatibility. Intraspecific and interspecific crosses produced hybrid progeny from a diverse set of crosses. The resulting seedlings will be used to study phenotypic segregation of flower traits such as color, double flowers, picotee petals, and remontancy. Seedlings from crosses with S. vulgaris Tiny Dancer will be used to study segregation of the dwarf phenotype. Seedlings from crosses Syringa ×hyacinthiflora ‘Old Glory’ and S. emodii will be used to study segregation of the purple- and yellow-leaved phenotypes, respectively. Selections among hybrid seedlings may identify novel combinations of traits such as purple-leaved dwarfs, double-flowered dwarfs, or double-picotee flowers.

Cross-incompatibility may be due to differences in genome size and ploidy level. Future evaluation of genome size and ploidy among taxa may provide insight into cross-compatibility in the current study. Cytological studies describe lilacs to be primarily diploids (Darlington and Wylie, 1956) with holoploid genome sizes near 2.5 pg (Olszewska and Osiecka, 1984; Siljak-Yakovlev et al., 2010). Cross-incompatibility can also be a function of pollination biology. Self-incompatibility systems have been discovered in related genera (Phillyrea, Fraxinus, and Olea) and sporophytic cross-incompatibility systems with S-allele dominance relationships have been discovered in cultivars of Olea (Breton et al., 2014; Collani et al., 2012; Koubouris et al., 2014; Saumitou-Laprade et al., 2017; Vernet et al., 2016). Future studies investigating rates of gametophytic/sporophytic incompatibility and unreduced gametes may provide insights into cross-compatibility. Environmental conditions have also been shown to affect seed development in lilacs (Junttila, 1973), and the current study may provide crosses useful for determining optimum greenhouse temperatures for pollination.

While many intraspecific and interspecific crosses were successful, no interseries crosses yielded viable seedlings. Some interseries crosses did, however, produce capsules and seed. A preliminary in vitro germination trial yielded low germination percentage and callus production. These results provide evidence that interseries hybrids among series Syringa, Pubescentes, and Villosae may be possible. Interseries seed development in hybrids with Ligustrina have also shown promise. In future research, seed-producing interseries crosses could be repeated and anatomical studies on embryo development, similar to Chen et al. (2012), may yield important information for targeting future embryo rescue efforts. Open-pollinated seed could be used to identify suitable culture media. Tissue culture and embryo rescue protocols are highly genotype-specific; therefore, cross-specific embryo rescue and embryogenic callus media could be fine-tuned for interseries hybrids that produce viable embryos early in seed development. Recovery of interseries hybrids in lilac will likely prove difficult. However, the current study in combination with other breeding and tissue culture studies provides a foundation for development of novel hybrid lilacs. For a group of ornamental shrubs and trees that have been bred for nearly 500 years, there are still new horizons for breeders to pursue in modern lilac breeding.