Abstract
Morphological analysis historically has been used to determine parentage of unknown hybrids. This can be difficult when potential parents have similar appearance, as in the case of three azaleodendron cultivars, Rhododendron L. ‘Fragrans’, ‘Fragrans Affinity’, and ‘Fragrant Affinity’. These cultivars are similar in name and appearance, and all are purported hybrids of R. catawbiense Michx. or R. ponticum L. and R. viscosum (L.) Torr. Amplified fragment length polymorphism (AFLP) analysis was conducted to determine whether the cultivars are synonyms or distinct clones and to elucidate the parental species. The three cultivars, suspected to be hybrids between taxa in subgenera Hymenanthes (Blume) K.Koch (evergreen rhododendrons) and Pentanthera (G.Don) Pojarkova (deciduous azaleas), and related taxa from each subgenus were evaluated using 31 AFLP primer combinations. Genetic similarity, calculated using Jaccard's coefficient, among the hybrids ranged from 53% to 71%, indicating that they are distinct cultivars and not a single clone. Genetic similarity was highest between the hybrids and R. ponticum among the evergreen rhododendrons, and R. viscosum among the deciduous azaleas. A dendrogram generated using the genetic similarity matrix grouped taxa into their respective subgenera, with the three cultivars nested intermediately between subgenera but more closely with subgenus Hymenanthes and particularly R. ponticum, suggesting it is the evergreen rhododendron parent. Furthermore, principle components grouped R. ponticum more closely with the hybrids and there were 18 AFLP fragments unique to R. ponticum and the hybrids. However, no unique AFLP bands were shared exclusively among the hybrids and the purported deciduous azalea parent, R. viscosum, suggesting that the original azalea parents may have been hybrids.
Rhododendron L. is an extremely diverse genus with eight subgenera and more than 1000 species (Chamberlain et al., 1996). This diversity, combined with broad crossability, has led to the development of more than 28,000 cultivars registered with the Royal Horticulture Society (RHS), including 14,298 rhododendrons, 12,989 azaleas, and 108 azaleodendrons (intersubgeneric hybrids between azaleas and rhododendrons) (Leslie, 2002). Thus, this genus is among the most extensively developed and widely grown in the world. Many superior cultivars continue to be used in breeding programs, but are often of unknown parentage.
Three azaleodendron cultivars exist in the trade with similar names: ‘Fragrans’, ‘Fragrans Affinity’, and ‘Fragrant Affinity’. ‘Fragrans’ is a purported hybrid of R. catawbiense Michx. and R. viscosum L. (Torr.) that was introduced by Paxton, of Chandler & Sons Nursery, London, in 1843. It is described as, “A sweet-scented azaleodendron, fast-growing and compact. Trusses of small flowers, pale mauve with centers lighter to white” (Salley and Greer, 1986, p. 110). ‘Fragrans Affinity’ was found in a group of deciduous azalea seedlings at Greer Gardens, Eugene, Ore., in the 1950s. Harold Greer (pers. comm., May 2004) speculated that it may be a hybrid of R. ponticum L. and R. viscosum or R. catawbiense and R. viscosum. The plant was named ‘Fragrans Affinity’ because of its similarity to ‘Fragrans’. The history of ‘Fragrant Affinity’ is poorly documented. A plant was received in 2000 from the late Dr. A. Kehr, who indicated it was a hybrid between R. viscosum and R. catawbiense. The name ‘Fragrant Affinity’ is not registered with the RHS or the American Rhododendron Society, and the ultimate origin of this material is unknown. Although ‘Fragrant Affinity’ is sterile, an allotetraploid form of ‘Fragrant Affinity’, named ‘Fragrant Affinity Tetra’, was developed that exhibited restored fertility (Contreras et al., 2007). All three are semievergreen and closely resemble the description provided by Salley and Greer (1986) of ‘Fragrans’ (pers. obs.).
The suspected parents of these azaleodendrons are distinct taxonomically and are classified in separate subgenera. Rhododendron catawbiense and R. ponticum are in subgenus Hymenanthes (Blume) K.Koch, section Ponticum G.Don, subsection Pontica Sleumer. This subsection contains evergreen species from North America, Europe, and Asia, including R. hyperythrum L. and R. maximum L. Rhododendron viscosum is in subgenus Pentanthera (G.Don) Pojarkova, section Pentanthera G.Don. This section contains other fragrant, deciduous species from North America including R. arborescens (Pursh) Torrey, R. atlanticum (Ashe) Rehd., R. canescens (Michx.) Sweet, and R. periclymenoides (Michx.) Shinners (Chamberlain et al., 1996).
Parentage of these azaleodendrons is difficult to determine because these hybrids were not the result of controlled pollinations. Furthermore, use of traditional morphological comparison to clarify parentage is complicated because of the number of species that could potentially be involved and the similarity of traits among species within each subgenus. Molecular techniques have been used widely to assess genetic relationships among plants. The amplified fragment length polymorphism (AFLP) technique (Vos et al., 1995; Zabeau and Vos, 1993), in particular, has been used by many scientists to distinguish between species as well as cultivars of the same species (DeHaan et al., 2003; Mellish et al., 2002; Paul et al., 1997; Perera et al., 1998; Zhang et al., 2000). Because of the inclusion of a restriction endonuclease digestion step, AFLPs have the advantage of being highly reproducible in comparison with random amplified polymorphic DNA markers. Milla et al. (2005) reported that reproducibility of markers used in a study of Arachis L. germplasm ranged from 96% to 100%. Also, AFLP analysis has a very high resolution, requires no prior knowledge of the genomes being studied (Vos et al., 1995), and has the capacity to screen simultaneously for many DNA regions distributed throughout the genome, thus producing hundreds of genetic markers (Mueller and Wolfenbarger, 1999).
Amplified fragment length polymorphism markers have been used to confirm hybridity in a number of plants. Teo et al. (2002) confirmed the hybrid status of Mangifera odorata Griff. Kiew et al. (2003) assessed the hybrid status of four genera of Malesian plants. Beismann et al. (1997) differentiated between Salix alba L., S. fragilis L., and their hybrid S. ×rubens Schrank when morphological analysis was determined to be inconclusive.
The objectives of this study were 1) to elucidate the progenitor species of ‘Fragrans’, ‘Fragrans Affinity’, and ‘Fragrant Affinity’; and 2) to determine whether these cultivars are all distinct clones or possibly synonyms.
Materials and Methods
Plant material.
Genotypes evaluated in this study included Rhododendron cultivars ‘Fragrans’, ‘Fragrans Affinity’, and ‘Fragrant Affinity’; putative parents; and related taxa. Putative parental and related taxa included R. catawbiense, R. ponticum, R. maximum, and R. hyperythrum from subgenus Hymenanthes; and R. viscosum, R. arborescens, R. canescens ‘Varnadoes Phlox Pink’, and ‘Marydel’ (R. atlanticum × R. periclymenoides) from subgenus Pentanthera. Kalmia latifolia L. ‘Sharon Rose’ was used as an outgroup. Plants were grown in 11.4-L containers with a pine bark medium amended with 0.59 kg·m−3 dolomitic lime and 1.0 kg·m−3 micronutrient blend (Micromax, Scotts, Marysville, Ohio) under 50% shade and fertilized using 17N–7.4P–14.1K controlled-release fertilizer (Multicote, Vicksburg Chemical Co., Vicksburg, Mo.). Plant material was maintained at the Mountain Horticultural Crops Research Station (Fletcher, N.C.) and the J.C. Raulston Arboretum (JCRA; Raleigh, N.C.; Table 1). Plants maintained at JCRA were grown in display beds.
Taxa used in amplified fragment length polymorphism analysis to elucidate parentage and hybridity of Rhododendron ‘Fragrans’, ‘Fragrans Affinity’, and ‘Fragrant Affinity’.
Amplified fragment length polymorphism analysis.
Total genomic DNA was extracted using a cetyltrimethyl–ammonium bromide extraction method described by Stein et al. (2001), using ≈100 to 200 mg tissue from newly opening leaves. Amplified fragment length polymorphism analysis was conducted using the protocol described by Milla et al. (2005) with 31 primer combinations (Table 2).
Thirty-one primer combinations used in amplified fragment length polymorphism analysis of Rhododendron catawbiense, R. ponticum, R. maximum, R. hyperythrum, R. viscosum, R. arborescens, R. canescens ‘Varnadoes Phlox Pink’, R. ‘Marydel’ (R. atlanticum × R. periclymenoides), and Kalmia latifolia ‘Sharon Rose’, and the number of polymorphic bands scored for each combination.
Data analysis.
The AFLP-Quantar 1.0 (Keygene Products B.V., Wageningen, the Netherlands) software package was used to score distinct, major, unambiguous bands. Presence or absence of each AFLP fragment was scored as a binary unit character (present, 1; absent, 0). Jaccard's coefficient of similarity was calculated using the SIMQUAL function of NTSYSpc 2.1 software (Exeter Software, Setauket, N.Y.) and subsequently used to construct a dendrogram using the unweighted pair group method with arithmetic averages. Principle component analysis was based on the variance–covariance matrix of the data using the PRINCOMP function of SAS 9.1 software (SAS Institute, Cary, N.C.). SigmaPlot 9.0 (Systat Software, Richmond, Calif.) was used to create a scatter plot of the first three principle components.
Results
Amplified fragment length polymorphism analysis.
The 31 primer combinations used generated a total of 152 useful AFLP polymorphisms ranging in size from 107 to 614 bp. Only bands that were unambiguous were scored and used for analysis. The number of scored polymorphic bands generated by each primer combination ranged from 1 to 10 (mean, 5). A total of 18 polymorphic bands specific only to R. ponticum and the hybrids were observed. No polymorphic bands specific only to R. viscosum and the hybrids were observed.
Genetic similarity matrix and cluster analysis.
The pairwise genetic similarities ranged from 0.04 to 0.72 (mean, 0.25; Table 3). Genetic similarities among the hybrids ranged from 53% to 71%, indicating they are distinct cultivars and not a single clone. The degree of similarity for the hybrids was highest with R. ponticum (53% to 61%) among the evergreen rhododendrons, and with R. viscosum (26% to 48%) among the deciduous azaleas. The dendrogram generated from the similarity matrix (Fig. 1) had subgenera nested in accordance with recognized taxonomic groups (Chamberlain et al., 1996). The upper branch contained the evergreen rhododendrons from subgenus Hymenanthes with the exception of R. ponticum. Rhododendron ponticum was nested within the same clade as the three azaleodendron cultivars, suggesting that it was a parent. The clade below the hybrids contained the deciduous azaleas from subgenus Pentanthera, with the exceptions of ‘Marydel’ and R. arborescens. Kalmia latifolia ‘Sharon Rose’ was most distantly related and formed the rooting branch.
Genetic similarity matrix based on Jaccard's coefficient of similarity of the 14 taxa evaluated to elucidate parentage and hybridity of Rhododendron ‘Fragrans’, ‘Fragrans Affinity’, and ‘Fragrant Affinity’.
Dendrogram created using the unweighted pair group method with arithmetic averages based on Jaccard's coefficient of similarity of the 14 taxa evaluated to elucidate parentage and hybridity of Rhododendron ‘Fragrans’, ‘Fragrans Affinity’, and ‘Fragrant Affinity’. R. ‘Marydel’ is a hybrid of R. atlanticum and R. periclymenoides.
Citation: HortScience horts 42, 3; 10.21273/HORTSCI.42.3.740
Principle component analysis.
Principle components one, two, and three accounted for 22%, 16%, and 15% of the variance observed between all samples respectively, for a total of 53% of the observed variation (Fig. 2). A plot of the first principle component separated the taxa into two groups. In one group were the deciduous azaleas and K. latifolia ‘Sharon Rose’. The second group contained the hybrids and the evergreen rhododendrons. Addition of the second principle component resulted in a separation of the hybrids and the evergreen rhododendrons. The third component separated K. latifolia ‘Sharon Rose’ from the deciduous azaleas. Also, it more closely allied the R. arborescens and ‘Marydel’ group with the other deciduous azaleas, in contrast to what was observed in the dendrogram (Fig. 1). The taxonomic groups were well defined with the exception of the two accessions of R. ponticum (Fig. 2).
Plot of first three principle components based on the variance–covariance matrix of the data using the 14 taxa evaluated to elucidate parentage and hybridity of Rhododendron ‘Fragrans’, ‘Fragrans Affinity’, and ‘Fragrant Affinity’. 1, R. viscosum; 2, R. canescens ‘Varnadoes Phlox Pink’; 3, R. arborescens; 4, R. hyperythrum; 5, R. viscosum ‘Roseum’; 6, R. catawbiense; 7, R. ‘Fragrans’; 8, R. maximum; 9, R. ‘Fragrant Affinity’; 10, R. ponticum 2005–217; 11, R. ponticum 2004–076; 12, R. ‘Fragrans Affinity’; 13, Kalmia latifolia ‘Sharon Rose’; 14, R. ‘Marydel’ (R. atlanticum × R. periclymenoides).
Citation: HortScience horts 42, 3; 10.21273/HORTSCI.42.3.740
Discussion
The current study demonstrates that the AFLP technique can be used to elucidate information about the parentage and hybridity of cultivars in the genus Rhododendron. Results of the molecular analysis provided strong evidence that R. ponticum, not R. catawbiense, was most likely the evergreen rhododendron parent of ‘Fragrans’, ‘Fragrans Affinity’, and ‘Fragrant Affinity’. Genetic similarities among the three azaleodendron cultivars ranged from 53% to 71%, indicating that the three hybrids are each distinct cultivars, not a single clone.
Amplified fragment length polymorphism data supported the hypothesis that R. ponticum was a parent of the three cultivars. Genetic similarity with the hybrids was highest with R. ponticum among the evergreen rhododendrons, ranging from 53% to 61% genetically similar. The other species ranged from 21% to 37% similarity with the hybrids. Cluster analysis nested the three azaleodendrons with subgenus Hymenanthes. Of species in this group they were most closely allied with R. ponticum. Also, principle component analysis grouped the hybrids more closely with R. ponticum than any other species included. Furthermore, there were 18 species-specific markers unique to R. ponticum and the hybrids, providing very strong evidence that it is a parent.
Data on the deciduous azalea parent was less conclusive. Rhododendron viscosum had the highest coefficient of similarity with the hybrids (26% to 44%) among deciduous azaleas, but there were no markers specific to R. viscosum that were also common to the hybrids. Deciduous azaleas in Pentanthera hybridize freely in cultivation and in the wild, thus it often is difficult to differentiate between species and hybrids (Towe, 2004). It is probable that a hybrid azalea was a parent in these azaleodendron crosses. Considering that the hybrids have fragrant flowers of pale lavender, the potential candidates from the deciduous azaleas are somewhat limited. Rhododendron luteum (L.) Sweet and austrinum (Small) Rehd., for instance, are unlikely because of their brilliant yellow flower color. The limited number of species that could have contributed the morphological traits seen in the hybrids, their overlapping distribution, and their propensity for natural hybridization supports the hypothesis that the deciduous azalea parent is a hybrid of species included in the study.
The current study provided evidence that these three azaleodendron cultivars are intersubgeneric hybrids between an evergreen rhododendron and a deciduous azalea. All three azaleodendrons were nested between subgenus Hymenanthes and subgenus Pentanthera in the dendrogram. Confirmation that wide hybridization is possible may encourage more rhododendron breeders to attempt intersubgeneric hybridization to develop novel cultivars with traits from diverse taxa. The AFLP technique has proved to be a useful tool in determining parentage of rhododendron cultivars of uncertain origin. Elucidating parentage of the numerous superior cultivars of unknown parentage will allow breeders to use available germplasm more effectively. In contrast to previous literature, it is clear that R. ponticum, not R. catawbiense, was the evergreen rhododendron parent of the three azaleodendrons, but additional analysis would be necessary to determine better the deciduous azalea parents.
Literature Cited
Beismann, H. , Barker, J.H.A. , Karp, A. & Speck, T. 1997 AFLP analysis sheds light on distribution of two Salix species and their hybrid along a natural gradient Mol. Ecol. 6 989 993
Chamberlain, D.F. , Hyam, R. , Argent, G. , Fairweather, F. & Walter, K.S. 1996 Genus Rhododendron: Its classification and synonymy Royal Botanic Garden Edinburgh Edinburgh, U.K
Contreras, R.N. , Ranney, T.G. & Tallury, S.P. 2007 Reproductive behavior of diploid and allotetraploid Rhododendron ‘Fragrant Affinity’ HortScience 42 31 34
DeHaan, L.R. , Ehlke, N.J. , Sheaffer, C.C. , Muehlbauer, G.J. & Wyse, D.L. 2003 Illinois bundleflower genetic diversity determined by AFLP analysis Crop Sci. 43 402 408
Kiew, R. , Teo, L.L. & Gan, Y.Y. 2003 Assessment of the hybrid status of some Malesian plants using amplified fragment length polymorphism Telopea (Syd.) 10 225 233
Leslie, A.C. 2002 The international Rhododendron register and checklist 2nd ed. Royal Botanic Garden Edinburgh Edinburgh, U.K
Mellish, A. , Coulman, B. & Ferdinandez, Y. 2002 Genetic relationships among selected crested wheatgrass cultivars and species determined on the basis of AFLP markers Cell Biol. Mol. Genetics 42 1662 1668
Milla, S.R. , Isleib, T.G. & Stalker, H.T. 2005 Taxonomic relationships among Arachis sect. Arachis species as revealed by AFLP markers Genome 48 1 11
Mueller, U.G. & Wolfenbarger, L.L. 1999 AFLP genotyping and fingerprinting Trends Evol. Ecol. 14 389 394
Paul, S. , Wachira, F.N. , Powell, W. & Waugh, R. 1997 Diversity and genetic differentiation among populations of Indian and Kenyan tea (Camellia sinensis (L.) O. Kuntze) revealed by AFLP markers Theor. Appl. Genet. 94 255 263
Perera, L. , Russell, J.R. , Provan, J. , McNicol, J.W. & Powell, W. 1998 Evaluating genetic relationships between indigenous coconut (Cocos nucifera L.) accessions from Sri Lanka by means of AFLP profiling Theor. Appl. Genet. 96 545 550
Salley, H.E. & Greer, H.E. 1986 Rhododendron hybrids: A guide to their origins Timber Press Portland, Ore
Stein, N. , Herren, G. & Keller, B. 2001 A new DNA extraction method for high-throughput marker analysis in a large genome species such as Triticum aestivum Plant Breed. 120 354 356
Teo, L.L. , Kiew, R. , Set, O. , Lee, S.K. & Gan, Y.Y. 2002 Hybrid status of kuwini, Mangifera odorata Griff. (Anacardiaceae) verified by amplified fragment length polymorphism Mol. Ecol. 11 1465 1469
Towe, C.L. 2004 American azaleas Timber Press Portland, Ore
Vos, P. , Hogers, R. , Bleeker, M. , Reijans, M. , van de Lee, T. , Hornes, M. , Frijters, A. , Pot, J. , Peleman, J. , Kuiper, M. & Zabeau, M. 1995 AFLP: A new technique for DNA fingerprinting Nucl. Acids Res. 23 4407 4414
Zabeau, M. & Vos, P. 1993 Selective restriction fragment amplification: A general method for DNA fingerprinting European patent application no. 92402629.7, publication no. EP 0534858
Zhang, D. , Cervantes, J. , Huamán, Z. , Carey, E. & Ghislain, M. 2000 Assessing genetic diversity of sweet potato (Ipomoea batatas (L.) Lam.) cultivars from tropical America using AFLP Gen. Res. Crop Eval. 47 659 665