TARGET REGION AMPLIFICATION POLYMORPHISM (TRAP) ANALYSIS OF PELARGONIUM
DISSERTATION
Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy
in the Graduate School of The Ohio State University
By
Rose E. Palumbo, M.S.
*****
The Ohio State University 2008
Dissertation Committee:
Dr. Guo-Liang Wang, Advisor Approved by
Dr. Charles Krause
Dr. Terry Graham Advisor Dr. Andrea Wolfe Plant Pathology Graduate Program
Copyright by
Rose Ellen Palumbo
2007
ABSTRACT
ABSTRACT
Characterization of plant collections with molecular markers is an ideal approach
for efficient conservation of plant genetic resources. In 2003, Pelargonium was a priority
genus at the Ornamental Plant Germplasm Center (OPGC), which accumulated
approximately 800 Pelargonium accessions. In order to make space for other important
ornamental species, the collection needed to be reduced to 25% of its peak size. The
goals of this project were to use molecular marker techniques to investigate phylogenetic relationships among the accessions in OPGC’s Pelargonium collection, and to assist
OPGC in purging redundant accessions. The molecular screening technique selected was
Target Region Amplification Polymorphism (TRAP), which combines the use of arbitrary primers and primers targeted to a gene of interest. Analyses were run to assess the utility of the TRAP technique for this genus, to evaluate the collection in terms of taxonomic sections, to compare potentially duplicated accessions, and to compare related accessions within section Ciconum using TRAP primers based on resistance genes. The first test on 46 Pelargonium accessions found the combined results from two sets of
TRAP primers were sufficient to distinguish all the accessions on a dendrogram. The entire collection was analyzed using 301 TRAP markers in a neighbor joining analysis, and the dendrogram produced in the analysis revealed a division of the collection into two groups: section Ciconum and non-Ciconum accessions. Dendrograms produced by
ii Bayesian analysis of the non-Ciconum accessions showed that section Pelargonium has the highest representation in the non-Ciconum group at OPGC, followed by section
Reniformia. Ultimately, 103 accessions were found to represent most of the diversity in the non-Ciconum group. A separate Bayesian analysis of potentially duplicated accessions from section Ciconum confirmed the identity of accessions in about half the sets of potential duplicates. For the most diverse 103 accessions in the non-Ciconum group to be retained, the accessions in section Ciconum have to be reduced from over 650 to just 97, keeping the total collection size below 200 accessions. In conclusion, TRAP analyses successfully identified the taxonomic section for the previously unidentified accessions in OPGC’s collection, and grouped the known accessions into the expected clusters.
iii ACKNOWLEDGMENTS
ACKNOWLEDGMENTS
Many thanks to my advisor and committee for their support and advice during the preparation of this thesis, and also to the many members of Dr. Wang’s lab for their patience and assistance over the last four years. I am very grateful for Dr. Jinguo Hu’s instruction and advice regarding the TRAP technique, and for Dr. MacDonald Wick’s generosity in sharing his lab and equipment with me. Also, thanks Kimberly Walters and
Jeff Pan from the statistical consulting service for their assistance with the statistics in chapter 2, and for their efforts to analyze the entire collection. This project would not have been possible without the generosity of Dr. Richard Craig and Charles Heidgen who donated the plants and provided background information about them. Finally, thanks to my family for being a source of unwavering support.
iv VITA VITA
2000-2001………………..Undergraduate Research - Virginia Tech
2001-2003………………..Graduate Research - Virginia Tech
2002………………………B.S. Crop and Soil Environmental Sciences – Biotechnology Virginia Tech
2002………………………B.S. Horticulture – Science Virginia Tech
2003………………………M.S. Horticulture – Virginia Tech
2004-2007………………..Graduate Research Associate – The Ohio State University
PUBLICATIONS
Palumbo, Hong, Hu, Craig, Locke, Krause, Tay and Wang. 2007. Target Region Amplification Polymorphism (TRAP) as a Tool for Detecting Genetic Variation in the Genus Pelargonium. HortScience 42(5): 118-1123
Palumbo, R. and Veilleux, R. 2007. GFP expression in potato. The American Journal of Potato Research. 84(5)
FIELDS OF STUDY
Major Field: Plant Pathology
Other Fields: Horticulture, and Crop and Soil Environmental Sciences
v TABLE OF CONTENTS
TABLE OF CONTENTS
ABSTRACT…...... ii
ACKNOWLEDGMENTS ...... iv
VITA…………...... v
LIST OF TABLES...... ix
LIST OF FIGURES ...... x
CHAPTER 1: INTRODUCTION...... 1
Pelargonium taxonomy ...... 1
Pelargonium systematics...... 6
Pelargonium hybrids...... 7
Molecular markers...... 11
Target gene: Arthropod resistance...... 15
Purpose...... 17
Goals...... 18
vi CHAPTER 2: Target Region Amplification Polymorphism (TRAP) as a tool for
detecting genetic variation in the genus Pelargonium...... 20
Abstract: ...... 20
Introduction...... 21
Materials and Methods...... 25
Plant material and TRAP primers:...... 25
Pelargonium DNA extraction:...... 26
TRAP amplification: ...... 27
Gel analysis: ...... 27
Results and Discussion...... 31
Optimization of DNA extraction method for TRAP reaction:...... 31
TRAP analysis of 46 Pelargonium accessions:...... 33
CHAPTER 3: Identifying the most diverse subset of OPGC’s Pelargonium
collection...... 45
Abstract: ...... 45
Introduction...... 46
Materials and Methods...... 48
Plant material and DNA extraction:...... 48
TRAP primers...... 49
TRAP amplification: ...... 50
Gel and data analysis: ...... 51
vii Results ...... 72
Phylogenetic analysis of non-Ciconum accessions ...... 72
Bayesian analysis...... 74
Pelargonium...... 75
Reniformia...... 78
Phylogentic analysis of section Ciconum ...... 90
Potential duplicates ...... 91
Accessions with known pedigree...... 92
Discussion...... 95
Conclusion ...... 98
CHAPTER 4: DISCUSSION AND GENERAL SUMMARY...... 106
LIST OF REFERENCES...... 111
viii LIST OF TABLES
LIST OF TABLES
Table 1.1: Sections in the genus Pelargonium grouped according to Van der Walt (1992)
and changes proposed since then...... 4
Table 2.1: Pelargonium Accession Details...... 29
Table 3.1: OPGC Pelargonium accessions with known pedigree...... 54
Table 3.2: Potential duplicates within OPGC’s Pelargonium collection...... 60
Table 3.3: OPGC accessions from section Ciconum with pedigree and/or known
resistance/susceptibility traits...... 65
ix LIST OF FIGURES
LIST OF FIGURES
Figure 2.1: A sample TRAP image generated by the primer combination w3a-700
(QHF6H21L + Sa12-700)...... 38
Figure 2.2: Dendrogram from the w3a-700 primer set...... 39
Figure 2.3: Dendrogram from w1-700 and w1-800 primers sets...... 41
Figure 2.4: Photographs of Pelargonium accessions...... 43
Figure 3.1: Diagrams of relationships between Ciconum selections based on records
from OPGC’s database...... 70
Figure 3.2: Non-Ciconum section of the Pelargonium dendrogram according to the
neighbor joining tree...... 79
Figure 3.3: Dendrogram from Bayesian analysis of all Pelargonium accessions from
outside section Ciconum and a few representatives of that section...... 81
Figure 3.4: P. fragrans accessions and three accessions with no pedigree that all
clustered together in Figure 3.3...... 83
Figure 3.5: Four accessions of Reniformia that were separated from the P. fragrans
accessions in Figure 3.3...... 84
Figure 3.6: Two groups of P. × domesticum accessions...... 85
Figure 3.7: The three accessions in section Pelargonium that are not part of a larger
cluster in Figure 3.3...... 86
x Figure 3.8: The differences in leaf shape and plant form found in P. crispum and P. ×
nervosum accessions...... 87
Figure 3.9: Variation in appearance among accessions in section Pelargonium...... 88
Figure 3.10: Dendrogram containing 26 sets of potentially duplicated accessions...... 100
Figure 3.11: Bayesian analysis of Ciconum accessions selected for their resistance and
breeding history...... 102
Figure 3.12: TRAP analysis (targeting resistance genes) of Ciconum accessions selected
for their resistance and breeding history...... 104
xi CHAPTER 1
INTRODUCTION
CHAPTER 1: INTRODUCTION
Species of Pelargonium are some of the most popular cultivated flowers in the
world and the Royal Horticultural Society (http://www.rhs.org.uk/) listed over 3,000 entries for distribution in 2004. In 2004 the combined wholesale value for all flats, hanging baskets and pots of Pelargonium plants in the United States was over $206
million (Jerardo, 2006). There are more than 230 species in the genus and most are
native to South Africa, and Van der Walt published descriptions of ~140 of those species
in three volumes (Van der Walt, 1977; Van der Walt and Vorster, 1988). Some species
have been domesticated to take advantage of their distinct flower color, leaf shape, and
scent (Van der Walt, 1977). The interest in breeding has lead to many improved or novel
cultivars. In fact, efforts have been made to improve cultivars through somaclonal
variation (Ravindra et al., 2004) and genetic transformation (Hassanein et al., 2005;
Winkelmann et al., 2005).
Pelargonium taxonomy
In a presentation to an international geranium conference, Van der Walt (1992) summarized the subdivisions of the genus Pelargonium. Van der Walt (1992) reported that previous multidisciplinary studies divided the genus into 16 sections of species
1 (number of species reported for each section in parentheses): Pelargonium (24),
Glaucophyllum (7), Campylia (no number given), Cortusinia (13), Reniformia (5),
Eumorpha (6), Ciconum (9), Dibrachya (1), Ligularia (6), Jenkinsonia (28), Myrrhidium
(8), Hoarea (52), Peristera (6), Isopetalum (1), Otidia (8) and Polyactium (7). Most of the taxonomic sections were identified by William Henery Harvey in 1860, and have been modified repeatedly since that time (Van der Walt, 1977) (Table 1.1). However,
Van der Walt (1992) also reported potential changes to those sections, including four groups of related sections: 1) Pelargonium, Glaucophyllum, and Campylia, 2) Cortusinia and Reniformia, 3) Ciconum, Eumorpha and Dibrachya, and 4) Ligularia, Jenkinsonia, and Myrrhidium.
The sections Glaucophyllum and Campylia group with section Pelargonium, whereas section Reniformia was split from section Cortusinia (Van der Walt, 1992). Van der Walt (1992) also reported a group containing section Ciconum with sections
Eumorpha and Dibrachya. Support for that group could be found in Oliver and Van der
Walt’s report in 1984 that all species in section Dibrachya should be condensed into P. peltatum instead. Additionally, karyological studies of P. peltatum confirmed that this one remaining section Dibrachya species belongs in section Ciconum (Gibby et al.,
1990), and agreed with earlier recommendations (Van der Walt and Vorster, 1988 and
Van der Walt et al. 1990a) to divide section Eumorpha into other sections according to chromosome number. That reduced this group to just section Ciconum.
The fourth group of related sections includes Ligularia, Jenkinsonia, and
Myrrhidium (Van der Walt, 1992). Although karyological studies suggested that section
Jenkinsonia was an artificial group resulting from convergence of flower structure
2 (Gibby et al., 1990), a comprehensive examination of morphology and chromosome number suggested that section Jenkinsonia is closely related to section Myrrhidium and possibly includes P. tetragonum, a hybrid of the two sections (Scheltema and Van der
Walt, 1990).
Likewise, section Ligularia was also highly varied, and had included natural hybrids of the species P. ionidiflorum, P. abronifolium, and P. exstipulatum with members of section Reniformia. These three species were identified as more appropriately belonging in section Reniformia, which led to a classification of member species with consistent chromosome number and appearance (Albers et al., 1992).
Albers et al. (1992) concluded that section Ligularia was probably an artificial section of differing species. In 1995, two studies divided two additional sections from section
Ligularia: section Chorisma (Albers et al., 1995) and section Subsucculentia (Van der
Walt et al., 1995).
3
Table 1.1: Sections in the genus Pelargonium grouped according to Van der Walt (1992) and changes proposed since then. Sections listed in grey are no longer recognized, and dark lines outline groups from Van der Walt (1992). Numbers of species in each section are indicated according to:
* Sections by Reinhard Kunth as summarized in Van der Walt (1977) included new species in William Henry Harvey’s sections from 1860
** Sections included in descriptions by Van der Walt (1988) in three volumes, according to the index of volume 3. The series includes some, but not all, of the species from regions other than South Africa.
*** Sections used in Van der Walt’s (1992) presentation to the Third International Geraniums Conference
**** Sections currently listed on GRIN (USDA-ARS Germplasm Resource Information Network: GRIN www.ars-grin.gov)
# Not all species are assigned to a taxonomic section, so the total number of recognized species is larger than the total of all sections. The total number of species has also changed through reclassified species.
4
Section # of Species Section Changes 1912* 1988** 1992*** 2007**** Pelargonium 28 24 24 25 Glaucophyllum 6 6 7 7 Campylia 7 7 no # 14 Cortusina 19 9 13 7 Reniformia - - 5 10 Eumorpha 14 0 6 0 Distributed into Ciconum and Glaucophyllum Ciconum 4 10 9 22 Dibrachya 4 1 1 0 Moved into Ciconum Ligularia 23 30 6 28 Source of new sections Chorisma (4 sp.) and Subsucculentia (4 sp.) Jenkinsonia 4 4 28 6 Myrrhidium 9 6 8 8 Seymouria 5 0 0 0 Moved into Hoarea Hoarea 48 15 52 19 Peristera 21 10 6 21 Isopetalum - 1 1 1 Otidia 8 8 8 8 Polyactum 27 11 7 11 Total species # >230 >250 197 Not all in sections Table 1.1
5 Pelargonium systematics
Some molecular systematic studies of Pelargonium have focused on the use of
plastid and mitochondrial DNA sequences. Bakker et al. (2000) were able to align
mitochondrial (nad1 exons b and c) and chloroplast (trnL-F region) sequences with other angiosperms to observe many informative changes to both sequences including a missing intron from the Pelargonium mtDNA in between the two exons, which had not been
reported in angiosperms outside of Gernaniaceae. Combined analysis of the two data
sets grouped species from sections Jenkinsonia, Chorisma, Myrrhidium, Ciconum,
Subsucculentia, Ligularia, Polyactum, Otidia, Pelargonium and Perstera by chromosome
size (and number) and taxonomic section in a semi-strict consensus tree, with the
exception of P. caylae and P. mutans which were separated from the other accessions of
section Ciconum. Bakker et al. (2000) suggested that the placement of P. mutans as a
sister group to section Jenkinsonia is a clue that it should be part of section Jenkinsonia.
In 2004, Bakker et al. enhanced their analysis by including additional species, and
nuclear (rDNA ITS region) as well as the same mitochondrial and chloroplast DNA
sequences. The strict consensus tree resulting from the combination of data from all
three genomes revealed clusters of species in the recognized taxonomic sections with few
exceptions (Bakker et al., 2004). The taxonomic sections were divided into five clades:
1) Hoarea, Ligularia, Otidia, Polyactum and Cortusinia, 2) Pelargonium and Campylia,
3) Reniformia and Peristera, 4) Ciconum and Subsucculentia and 5) Myrrhydium,
Chorisma and Jenkinsonia. In clade 1, the only exception to clustering by taxonomic
section was that a subsection of section Polyactium was on a separate branch than the rest
of that section. In clade 2, two section Campylia species clustered within section
6 Pelargonium, and two section Pelargonium species clustered on a sister branch to the rest
of section Campylia. Also P. nanum was in clade 2 instead of clade 3 with the rest of section Peristera, and a single section Isopetalum species clustered with section
Peristera. In clade 4, P. caylae and P. transvealense clustered with section
Subsucculentia instead of section Ciconum, and two section Jenkinsonia species were in
clade 4 instead of clade 5. In clade 5, P. mutans was in a cluster with species from
section Jenkinsonia instead of in clade 4 with the rest of section Ciconum. In spite of
those exceptions, Bakker et al.’s (2004) phylogeny of about 140 Pelagonium species is
currently the best phylogenetic hypothesis for this genus.
Pelargonium hybrids
At the third geranium conference in 1992, Horn reported that natural hybrids can
occur both within and between sections, and Van der Walt (1992) reported that 23
examples of natural hybrids had been identified. For example, within section
Pelargonium, P. graveolens and P. radens are closely related and hybridize naturally,
blurring the line between the two species that is most obvious in the texture of their
leaves (Van der Walt and Demarne, 1988). Natural hybrids which occurred within
sections included ten in section Pelargonium, two in section Glaucophyllum, and one in
section Campylia, while the natural hybrids that occurred between sections included eight
between sections Pelargonium and Glaucophyllum, one between sections Glaucophyllum
and Campylia, and one between sections Eumorpha and Dibrachya (later Ciconum) (Van
der Walt 1992).
7 The larger numbers of hybrids found within sections compared to between them may be explained by the size and number of chromosomes in the species within each section. Horn (1992) reported that compatible chromosome number is very important to the production of hybrids, and that production of viable seed is more successful for hybrids within taxonomic sections as compared to intersectional hybrids. Van der walt
(1992) listed a wide range of basic chromosome numbers (x=4 to x=17) and ploidy (2n=8 to 2n=104) within the genus Pelargonium, but only five of the sections (Ligularia,
Jenkinsonia, Peristera, Hoarea, and Eumorpha) contained variation in chromosome number among member species. Coetzee et al. (1994) investigated the origin of a presumed natural hybrid of P. patalum (Glaucophyllum) and P. tomentosum
(Pelargonium), and their karyological study revealed that the hybrid had an intermediate chromosome number of 2n=33 compared to the two parents P. patalum (2n=22) and P. tomentosum (2n=44). However, all three had the same chromosome size (Coetzee et al.,
1994), which is consistent with Gibby et al’s (1996) report that chromosome size was even more important than chromosome number for successful hybridizations of species in Pelargonium section Hoarea.
Hybridization between the members of different sections has been used to confirm relationships between those sections. For example, in one case a fertile natural hybrid between Erodium incarnatum and P. patulum (Glaucophyllum), was partially responsible for the reclassification of E. incarnatum as Pelargonium incarnatum
(Campylia) (Van der Walt et al., 1990b). Van der Walt and Roux (1991) reported that the species in section Campylia are highly varied and naturally hybridize with species in sections Glaucophyllum and Pelargonium. A karyological study (Albers and Van der
8 Walt, 1992) of that hybrid and several accessions of each parent species from the same
location revealed that the hybrid was an allotetraploid with 2n=42 chromosomes. The
accessions of the parent species P. incarnatum were all tetraploids with 2n=40, whereas
the accessions of the other parent, P. patulum, included a rare tetraploid with 2n=44
among the typical diploids (2n=22) (Albers and Van der Walt, 1992). The hybridizations
confirmed in those studies provides evidence in favor of Van der Walt’s (1992) grouping
of all three sections, and are consistent with Bakker et al.’s (2004) clade with sections
Campylia and Pelargonium.
Most of the known natural hybrids Van der Walt (1992) reported are within section Pelargonium or between section Pelargonium and section Glaucophyllum. Van der Walt et al.’s (1990a) description of section Glaucophyllum is another good example
of hybrids helping to define the relationships among sections. Van der Walt et al.
(1990a) applied a multidisciplinary study to support the transfer of two species out of section Glaucophyllum and addition of several species from section Eumorpha, they also used the natural hybrids of section Glaucophyllum species with section Pelargonium species to define a relationship between the sections. Van der Walt et al. (1990a) observed that even though species in section Glaucophyllum show adaptations to a desert environment, distinctive traits such as pollen grain morphology and chromosome number are shared with members of section Pelargonium, which prefers more moisture.
Horn (1992) also reviewed the value of interspecific hybrids for producing commercial cultivars, pointing out that most Pelargonium cultivars are derived from hybrids of only ten species, and that embryo rescue has been used to produce hybrids from crosses that normally would not produce fertile hybrids. For example, Bentvelsen
9 (1992) reported that generating hybrid cultivars P. peltatum required embryo rescue to avoid aborted seeds. Many of the cultivars that were derived from hybridizations of numerous species have been assigned to new species, or “hybrid species,” rather than identified as hybrids of one of the contributing species. The most common “geranium” used in horticulture is the hybrid species P. × hortorum (zonal geranium); another common hybrid species is P. × domesticum (regal geranium). Additionally some P. peltatum (ivy geranium) cultivars are hybrids.
Molecular marker and sequence data have been used to examine the parentage of
Pelargonium hybrids. James et al. (2004) used RFLP data from chloroplast and nuclear
DNA to confirm that P. inquinans and P. zonale were most likely the largest contributors to P. × hotorum. The plants included by James et al. (2004) were from section Ciconum, including some former section Eumorpha species, and they also illustrated that the former Eumorpha species might be a sub-section of Ciconum. In multiple analyses two section Ciconum species, P. mutans and P. caylae, were separated from the main cluster of section Ciconum. P. mutans was treated as an out group, because visual screening of the data identified it as the most unique (James et al., 2004). The separation of those two groups is in agreement with Bakker et al. (2000, 2004), and exists in spite of the inclusion of additional species to the study (James et al., 2004). James et al. (2004) also suggested that at least some of the ivy-leaved cultivars are P. peltatum accessions that have been hybridized, and that P. zonale may have contributed to those cultivars.
Loehrlein and Craig (2001) reviewed the history of the hybrid species P. × domesticum, which they reported originated in Europe from interspecific crosses beginning in 1690 with P. cucullatum (section Pelargonium) and continuing with the
10 addition of up to eight other contributing species from South Africa. Although they
reported that P. cucullatum hybridizes naturally in the wild, they also pointed out that the
complex pedigree of the hybrid species, P. × domesticum, does not occur in the wild.
Loehrlein and Craig (2001) summarized several reports of other potential contributing species, including several scented species from section Pelargonium and large flowering species from section Glaucophyllum that P. × domesticum hybridizes with naturally.
The earliest P. × domesticum cultivars had small pansy-like flowers, and were developed into the first large flowered P. × domesticum cultivars by Carl Faiss in 1920 (Loehrlein and Craig, 2001).
Molecular markers
Molecular markers have been used in analyses of the genus Pelargonium, using either arbitrary markers or chloroplast sequences to clarify relationships among species.
Renou et al. (1997) tested the potential of RAPD markers to differentiate some cultivars of zonal (P. × hortorum) and ivy leaved (P. peltatum) geraniums from each other.
Although RAPD (Random Amplified Polymorphic DNA) markers showed promise by successfully differentiating species, the relatively small number of markers produced was not sufficient to distinguish all 34 cultivars from each other (Renou et al., 1997).
Starman and Abbitt (1997) used highly polymorphic arbitrary molecular markers to differentiate eight cultivars; the results clustered cultivars by species and ploidy level.
Another method worth considering for the categorization of cultivars within a species or section, is molecular characterization by Target Region Amplification
Polymorphism (TRAP) markers (Hu and Vick, 2003). The TRAP marker technique
11 combines the AT- and GC-rich primers of SRAP (Sequence-Related Amplification
Polymorphism; Li and Quiros, 2001) with a third “fixed” or “targeted” primer that matches a gene of interest. The advantages of using TRAP markers for cultivar identification are that 1) there is no need for extensive pre-PCR treatment of the DNA samples as there would be for SSRs (Simple Sequence Repeats), 2) many more fragments can be amplified in a single TRAP PCR reaction than would be with single-primer RAPD reactions, and 3) potential use of previously reported sequence information for a targeted primer. In addition to amplifying many fragments with one arbitrary primer and one targeted primer, TRAP reactions can be multiplexed with the additional arbitrary primers.
Each of the two arbitrary primers has a different fluorescent label, so the PCR amplified
DNA fragments can be detected by two different detectors, each producing a separate profile for the two fluorescently labeled primers used in the same reaction. These reactions are separated via gel electrophoresis using a Li-Cor Global Genotyper (Li-Cor
Bioscience, Lincoln, NE).
Although targeting a gene of interest for TRAP markers may enable fishing for markers to map that gene, it is important to remember the markers are not all directly linked to a particular gene. Many of the fragments amplified using a fixed primer are the product of slight mismatches, because the first five cycles of the PCR reaction have a lower stringency than subsequent cycles. Also, some fragments are amplified between the arbitrary primers, similar to RAPD markers, and would therefore not be linked to the gene of interest all. Hu and Vick (2003) developed the TRAP method in order to have an application for the increasingly large number of DNA sequences available. Hu et al.
(2003) employed the new TRAP technique to evaluate 42 accessions from 18 perennial
12 Helianthus species (or hybrids) using two targeted primers based on sunflower ESTs
(Expressed Sequence Tags) (A21B09b and B18I19b) for LRR (Leucine Rich Repeat) and
NBS (Nucleotide Binding Site) regions of disease resistance genes. Each of the four
multiplexed reactions combining one of the two targeted primers with two of four
arbitrary primers yielded over 100 polymorphic fragments (Hu and Vick, 2003). Hu et
al. (2003) presented a phylogenetic tree based on 257 of those fragments, which
demonstrated that the TRAP markers resulted in clusters of species consistent with
classical taxonomy of sunflowers. In order to evaluate the reproducibility of the TRAP
markers Hu and Vick (2003) used the same primer pairs to test segregating populations of
wheat from a recombinant inbred line and from an F2 generation. The polymorphic markers segregated in a 1:1 ratio for the recombinant inbred line and in a 3:1 ratio for the
F2 accessions. The results of both brief studies allowed Hu and Vick (2003) to conclude
that the TRAP technique is a useful way to apply the available sequence data to plant
genomics.
The TRAP technique has been used with various plants, including lettuce
(Lactuca sativa L.), common bean (Phaseolus vulgaris L.), sugarcane (Saccharum species), wheat (Triticum aestivum L. and Triticum turgidum L.), sunflower (Helianthus annuus L.), caladium (Caladium ×hortulanum Birdsey) and a cotton hybrid (Gossyplum hirsutum L. × G. barbadense L.). Hu et al. (2005) amplified 107 polymorphic markers from ten TRAP reactions, which discriminated 53 lettuce cultivars, and was fairly stringent in grouping the cultivars by horticultural types. Miklas et al. (2006) used a disease resistance gene in the common bean to demonstrate that TRAP identified new markers linked to disease resistance. Alwala et al. (2006) used TRAP on sugarcane,
13 which confirmed known systematics and also sorted cultivars into a phylogeny according to the gene of interest. Yang et al. (2005) reported using TRAP for quantitative trait loci
(QTL) mapping in wheat. They observed that TRAP produced more markers, but not in the targeted QTL regions. Liu et al. (2005) compared TRAP and SSR markers in wheat and determined that both could be used to generate genetic maps for QTL identification, but that TRAP produced more data from a single reaction. Li et al. (2006) used substitution lines of durum wheat to identify chromosome specific TRAP markers, and observed that targeted primers based on a mapped EST always produced at least one marker specific to where it was mapped. Also, approximately 15 percent of markers were on the same chromosome as the targeted EST (Li et al., 2006). Chen et al. (2006) identified six TRAP markers linked to a male-sterility gene in sunflower, and by using
SSR markers for comparison with the public linkage map, they mapped the gene to linkage group 10. Deng et al. (2007) used TRAP markers to evaluate the diversity of
Caladium cultivars. While their results confirmed fears that there was very limited diversity available in caladium cultivars, Deng et al. (2007) also found much more diversity among the species. Since it was more common for fragments found in the species to be missing in the cultivars than vise versa, they concluded that the loss of diversity in cultivars could be repaired by using the available species (Deng et al. 2007).
Yu et al. (2007) incorporated TRAP into their mapping project for cotton using silver staining rather than fluorescent labels. The number of polymorphic markers from each primer set was drastically reduced in Yu et al.’s (2007) results possibly because of their targeted primers or the clarity silver stain. Regardless, they still found markers that contributed to their mapping project (Yu et al. 2007).
14 In mapping applications, TRAP also provides an opportunity to extend existing maps closer to the ends of chromosomes, through the use of primers based on sequences from the telomere region. In 2006, Wang et al. attempted to use TRAP to map the physical end of a linkage map by using a fixed primer for the telomere region of the chromosome a Hessian fly resistance gene was mapped to, but the linked marker was not in the telomere region. Using the fixed primers derived from the Arabidopsis-type telomere sequences, Hu (2006) succeeded in defining 21 of the 34 linkage group ends of the sunflower linkage map, by using primers targeted to the telomere region with the addition of one extra base to avoid repetition.
As for the fragments amplified by TRAP, Hu (2006) speculated from two experiments in cloning TRAP fragments, that the three-primer TRAP reaction (initial ratio of fixed primer to arbitrary primers 30:1:1) could amplify six types of fragments with relative frequencies of 900:30:30:1:1:2. The most abundant fragment type has the fixed primer on both ends, and is not observed because it is not fluorescently labeled.
The next two types combine the fixed primer on one end with one of the two arbitrary primers on the other end, and represent most of the observed fragments. The two least frequent fragment types have the same arbitrary primer on both ends, and the final type has a different arbitrary primer at each end.
Target gene: Arthropod resistance
TRAP markers have been used for both phylogenetic analyses (Alwala et al.,
2006; Deng et al., 2007; Hu et al., 2005; Hu et al., 2003; Hu and Vick, 2003) and gene mapping (Chen et al., 2006; Hu, 2006; Li et al., 2006; Liu et al., 2005; Miklas et al.,
15 2006; Wang et al., 2006; Yang et al., 2005; Yu et al., 2007) studies. For TRAP analysis of Pelargonium a trait of interest to the horticultural industry is insect resistance. In addition to the damage caused by feeding, some insects also spread disease. For example, Chen and Williams (2006) reported that thrips are a damaging greenhouse pest that can transmit viruses.
In a review of the discovery of arthropod resistance in Pelargonium Mumma et al.
(1992) explained that the initial infestation of the two-spotted spider mite (Tetranychus urticae Koch) was surprising because most P. × hortorum cultivars are resistant to mites.
Unfortunately, cultivars from Europe tend to be susceptible to mites, and Mumma et al.
(1992) reported that susceptibility to mites also means susceptibility to the foxglove aphid (Acyrthosipon solani Kaltenbach). Microscopic observations of the mites and aphids on resistant leaves revealed that a sticky orange liquid excreted from tall trichomes could trap the bugs (Mumma et al., 1992). That sticky exudate was an unsaturated anacardic acid; in its saturated form it is dry and flaky and has very little affect on pests.
Schultz et al. (1996) reported that the resistance is due to a single dominant locus involved in anacardic acid production and that the ∆9 14:0-acyl carrier protein fatty acid
desaturase gene at that locus that might encode the desaturase gene. The fatty acid
desaturase causes a change in the anacardic acid excreted by the Pelargonium trichomes,
which causes it to be sticky and toxic to insects. Grazzini (1992) reported that the
resistant form of the anacardic acid (unsaturated) has only been found in three species:
16 P. acetosum, P. inquinans, and P. × hortorum. Since arthropod resistance is not found in
P. zonale (a parent species of P. × hortorum), the source of the resistance in zonal
hybrids (P. × hortorum) was most likely P. inquinans (another parent species of P. ×
hortorum ) (Grazzini, 1992).
Grazzini et al. (1999) examined the fatty acid content of various plant tissues, in
order to identify which ones had the fatty acid that could produce the resistance causing
unsaturated anacardic acid. Ultimately, the necessary types of fatty acid were only found
in glandular trichomes similar to the ones that were observed to have sticky exudates
(Grazzini et al, 1999). In addition to the sticky toxic trap for tiny insects, Schultz et al.,
(2006) found that larger insects (Colorado Potato Beetle) avoid food sources containing
this anacardic acid. Their goal was to find an appropriate application of this compound
as a natural, nonspecific, plant pest repellant (Schultz et al., 2006). The sequence of the
single dominant ∆9 14:0-acyl carrier protein fatty acid desaturase gene (PHU40344:
Schultz et al., 1996) responsible for insect resistance was used to develop targeted primers for TRAP analysis of Pelargonium accessions.
Purpose
The US floricultural industry ranked Pelargonium as one of the three most important floral species for germplasm conservation (Tay, 2003), and in 2003 the OPGC
(Ornamental Plant Germplasm Center, Columbus, Ohio) Pelargonium collection included approximately 800 accessions representing approximately 60 species. Plants were donated by Dr. Richard Craig of Penn State University and Mr. Charles Heidgen of
Shady Hill Gardens. The three most represented sections among the species at OPGC
17 (Table 3.1) are Ciconum, Pelargonium and Reniformia. Sections Chorisma,
Subsucculentia, Peristera, Ligularia and Myrrhidium are not included in OPGC’s
collection, whereas all of OPGC’s accessions from section Campylia and Polyactium are
hybrids; five hybrid accessions (hybrids of species from other sections in some cases)
were assigned to section Campylia, and a species from Polyactium contributes to just one
hybrid accession. Accessions from sections have been hybridized as well, but those
sections are represented by true species as well. Sections Glaucophyllum, Isopetalum,
Otidia, Cortusinia, Jenkinsonia, and Hoarea are represented at OPGC by only one or two
accessions. This small sample size is problematic with regards to section Hoarea (turnip shaped tubers), because it is the largest section within Pelargonium (Marais, 1989).
Since most of the Pelargonium cultivars at OPGC require vegetative propagation to remain true to type, the germplasm can be more efficiently maintained at OPGC if only the most genetically dissimilar accessions are retained. A large scale molecular screening of the existing collection was conducted. A purging of redundant accessions was proposed to provide space for a more balanced representation of the sections of
Pelargonium species in OPGC’s collection, leaving space for the acquisition of additional accessions not yet represented in the collection.
Goals
The goal of the first stage of this study (Chapter 2) was to test the potential of
TRAP markers for differentiating Pelargonium accessions using a small group of accessions. The goal of the second stage of this study (Chapter 3) was to apply TRAP analysis to OPGC’s Pelargonium collection in order to observe the overall diversity with
18 in the population, and based on those results advise OPGC on how to keep most of the diversity while shrinking their collection to 25 percent of its initial size. A subsection of that study used primers targeting resistance genes for TRAP analysis of accessions from section Ciconum. The use of TRAP primers targeting resistance genes had the dual goals of observing redundant accessions and identifying TRAP markers that could be used to screen for resistance in Pelargoniums without background data regarding resistance.
19 CHAPTER 2
TARGET REGION AMPLIFICATION POLYMORPHISM (TRAP) AS A TOOL
FOR DETECTING GENETIC VARIATION IN THE GENUS PELARGONIUM. *
*Published as: Palumbo, Hong, Hu, Craig, Locke, Krause, Tay and Wang. 2007. HortScience 42(5): 1118-1123
CHAPTER 2: Target Region Amplification Polymorphism (TRAP) as a tool for detecting genetic variation in the genus Pelargonium
Abstract:
Pelargonium was a priority genera collected by the Ornamental Plant Germplasm
Center (OPGC) until a recent reorganization. To preserve genetic diversity for future breeders, OPGC collects heirloom cultivars, breeding lines, and wild species. The current Pelargonium collection at OPGC consists primarily of cultivars originating from
P. × hortorum and P. × domesticum. Target region amplification polymorphism (TRAP) has the advantage of producing a large number of markers through use of sequence information that is already available. Our first goal was to determine the feasibility of
TRAP for the analysis of this large collection, so that in the future the most diverse genotypes may be retained. To achieve this goal, we first modified existing DNA extraction techniques to account for the high levels of phenolic compounds present in some Pelargonium species by combining several washes to remove the phenolics with the addition of high levels of antiphenolic compounds. Second, we evaluated the TRAP procedure using the DNA isolated from 46 accessions. For 44 accessions, one or two
20 primer combinations generated enough fragments to discriminate each of the accessions, and similar clades were produced by cluster analysis of the polymorphic fragments amplified by different primer combinations. All the scorable fragments were polymorphic, for one primer combination there were 148 markers from one gel image and the other produced 160 markers on two gel images. These results demonstrate that
TRAP is an effective method for molecular characterization of ornamental collections.
Introduction
(See Chapter 1 for more detail)
Pelargonium species are some of the most popular flowers in the world, and the
Royal Horticultural Society (http://www.rhs.org.uk/) listed more than 3000 entries for distribution in 2004. In 2004, the combined wholesale value for all flats, hanging baskets, and pots of Pelargonium plants in the United States was over $206 million
(Jerardo, 2006). There are over 280 species in the genus, and most are native to South
Africa. Some species have been domesticated to take advantage of their distinct flower color, leaf shape, and scent (Van der Walt, 1977; Van der Walt and Vorster, 1988). The interest in breeding has led to many improved or novel cultivars. In fact, efforts have been made to improve cultivars through somaclonal variation (Ravindra et al., 2004) and genetic transformation (Hassanein et al., 2005; Winkelmann et al., 2005).
The United States’ floricultural industry ranked Pelargonium as one of the three most important floral species for germplasm conservation (Tay, 2003), and in 2003, the
OPGC (Ornamental Plant Germplasm Center, Columbus, OH) Pelargonium collection included approximately 900 accessions representing approximately 60 species. Plants
21 were donated by Richard Craig of Penn State University and Charles Heidgen of Shady
Hill Gardens. Because the Pelargonium cultivars require vegetative propagation, the germplasm can be more efficiently maintained at OPGC if only the most genetically dissimilar accessions are retained. This necessitates a large-scale molecular screening of the current collection to identify redundant cultivars and provide space for additional accessions representing the more than 200 remaining species.
Ornamental plants are often selected for their aesthetic qualities rather than their ability to survive in any particular environment. As a result, the genetic base of most modern flower cultivars risks loss of other important traits. Gene banks (like OPGC) serve an important function by maintaining populations with traits that could otherwise be lost before anyone knew their importance (Tanksley and McCouch, 1997). By preserving the diversity of ornamentals, OPGC is protecting consumers and breeders against loss of genetic diversity required for future breeding. Many ornamental plants also produce compounds that have potential use in agriculture and medicine. For example, the essential oils that scented geraniums (Pelargonium sp.) produce have been used as perfumes and food flavoring (Becker and Brawner, 1996; Ravindra et al., 2004), and their inhibitory effects on bacteria and nematodes have also been studied (Lis-
Balchin et al., 1995). In addition, some Pelargonium accessions are resistant to arthropods, which a team at The Pennsylvania State University attributed to the anacardic acid composition of glandular exudates (Shultz et al., 1996).
22 Previous molecular systematics of Pelargonium species have focused on the use of chloroplast DNA (Bakker et al., 2000; James et al., 2004). Although maternal inheritance of chloroplast DNA can provide an advantage in studies of many plant species, this is not necessarily the case with Pelargonium species. Pollen can carry chloroplast DNA, providing for potential biparental inheritance (James et al., 2001).
We intend to determine the genetic similarity of the accessions through molecular characterization by target region amplified polymorphism (TRAP) markers (Hu and
Vick, 2003). TRAP is a technique that combines the AT-and GC-rich primers of SRAP
(sequence-related amplification polymorphism; Li and Quiros, 2001) with a third
‘‘fixed’’ primer that matches a gene of interest. The advantages of using this technique are that there is no need for extensive pre-PCR treatment of the DNA samples, that many fragments can be amplified in a single PCR reaction, and that previously reported genetic information has the potential to be used as the targeted primer. Each of the two arbitrary primers has a different fluorescent label, so the PCR-amplified DNA fragments can be detected by two different channels, each producing a separate image for one of the two fluorescently labeled primers on the same gel, using the LI-COR Global Genotyper (LI-
COR Bioscience, Lincoln, NE). Although targeting a gene of interest can provide an advantage for finding markers to map that gene, it is important to remember the markers are not all directly linked to the gene of interest. Many of the fragments amplified with the fixed primer will be products of slight mismatches, because the first five cycles have a lower stringency. Also, some fragments are amplified between the arbitrary primers and would therefore not be linked to the gene of interest at all. Hu and Vick (2003) demonstrated that TRAP provides reproducible results by analyzing wheat from a
23 recombinant inbred line and from an F2 generation. They also reported that marker data
generated from 16 perennial Helianthus species with two TRAP reactions by six primers
(two arbitrary and one fixed in each TRAP reaction) produced a phylogenetic tree that
had similar clustering as those produced using morphological characteristics.
The TRAP technique has been applied with various plants, including lettuce
(Lactuca sativa L.), common bean (Phaseolus vulgaris L.), sugarcane (Saccharum
species), wheat (Triticum aestivum L.), and sunflower (Helianthus annuus L.). Hu et al.
(2005) amplified 107 polymorphic markers from 10 TRAP reactions, which discriminated the 53 lettuce cultivars analyzed and was fairly stringent in grouping the cultivars by horticultural types. Miklas et al. (2006) used a disease-resistance gene in the common bean and demonstrated that TRAP identified new markers linked to disease resistance. Alwala et al. (2006) used TRAP on sugarcane, which confirmed known systematics and also sorted the phylogeny according to the gene of interest. Yang et al.
(2005) reported using TRAP to detect markers for quantitative trait loci (QTL) mapping
in wheat. They observed that TRAP produced more markers but not in the targeted QTL
regions. Liu et al. (2005) compared TRAP and SSR markers in wheat and determined
that both could be used to generate genetic maps for QTL identification, but that TRAP
produced more data from a single reaction. Chen et al. (2006) identified six TRAP
markers linked to a male-sterility gene in sunflower, and by using SSR markers for
comparison with the public linkage map, they mapped the gene to linkage group 10.
Wang et al. (2006) attempted to use TRAP to map the physical end of a linkage
map by using a fixed primer for the telomere region of the chromosome to which a
Hessian fly resistance gene was mapped, but the linked marker was not in the telomere
24 region. Using the fixed primers derived from the Arabidopsis-type telomere sequences,
Hu (2006) succeeded in defining 21 of the 34 linkage group ends of the sunflower
linkage maP. As for the fragments amplified by TRAP, Hu (2006) speculated from two experiments in cloning TRAP fragments, that the three-primer TRAP reaction (initial ratio of fixed primer to arbitrary primers, 30:1:1) could amplify six types of fragments with relative frequencies of 900:30:30:1:1:2. The most abundant fragment type has the fixed primer on both ends and is not observed because it is not fluorescently labeled. The next two types combine the fixed primer on one end with one of the two arbitrary primers on the other end, and they represent most of the observed fragments. The two least- frequent fragment types have the same arbitrary primer on both ends, and the final type has a different arbitrary primer at each end.
TRAP was chosen as an ideal method for analyzing the collection of a gene bank because it allows for evaluation of genetic variation that emphasizes specific traits of interest. As it had not yet been tested on Pelargonium species, our objective was to
demonstrate that TRAP could provide markers that distinguish between Pelargonium
species. Successful results will segregate according to known species designation of
selections, indicating that our long-term goal of applying TRAP to evaluate the similarity
of all the accessions in the OPGC collection should do the same.
Materials and Methods
Plant material and TRAP primers:
Fortysix Pelargonium accessions were selected from the 800 accessions available
at OPGC in 2003, including some with known pedigrees to demonstrate clustering within
a species and some representing the diversity of species in the OPGC collection to test for
25 future applicability to the entire collection. Selected accessions included nine species and
22 cultivars (Table 2.1). The primers used here were selected as a result of Jinguo Hu’s
experience that these primers, which have been developed for other projects in his
laboratory, produce acceptable TRAP results for most plant families. Based on the goal
to evaluate the procedure rather than a specific gene of interest, these primers were ideal.
The two arbitrary primers were Sa12–700 (TTCTAGGTAATCCAACAACA; Hu et al.,
2005) and Ga5–800 (GGAACCAAACACATGAAGA; Hu et al., 2005), and the fixed
primers were QHA21B09a (TGTCATTCAATTCGGTGC, homolog to an Arabidopsis
thaliana gene coding for an unknown protein At5g65840.1) and QHF6H21L
(ACAGGAAAAGCCTGTCAC, homolog to an A. thaliana gene coding for the BEL1-
like homeobox 1 protein). Information regarding both fixed primers was obtained from
The Compositae Genome Project website (http://compgenomics.ucdavis.edu).
Pelargonium DNA extraction:
This high-salt CTAB procedure was modified from techniques for cactus (Tel-zur et al., 1999) and for barley (Saghai-Maroof et al., 1984). Pelargonium leaf tissue (1 g)
ground in liquid nitrogen was added to 8 mL extraction buffer (as in Tel-zur et al., 1999).
The samples were centrifuged at 3300 rpm for 15 min, and the pellet was resuspended in
8 mL fresh extraction buffer. As in Tel-zur et al. (1999), this was repeated for a total of 3
washes. After the third wash, the pellet was suspended in a mixture of 1 mL extraction
buffer and 5 mL high salt-CTAB buffer [from Tel-zur et al. (1999), with modifications:
doubled Tris-HCl concentration, added sorbitol (25 g/L), 10% SDS, PVPP (Sigma
P6755; 20 g/L), and proteinase K (0.002 g/L)], and incubated in a 55–65 °C water bath for 60 min. An equal volume of chloroform was mixed in by shaking vigorously for 35 s
26 and then separated by centrifuging at 3300 rpm for 15 min. The supernatant was
transferred to a new tube, precipitated with isopropyl alcohol/sodium acetate, centrifuged
(3300 rpm for 15 min), and washed in ethanol as in Tel-zur et al. (1999). After
centrifuging again at 3300 rpm for 10 min, the pellet was air-dried for 10–15 min. The pellet was dissolved in 300 µL TE buffer with 10 µL RNase (10 mg/mL) and then incubated, extracted (twice), and precipitated according to Tel-zur et al. (1999), centrifuging when called for at 13,200 rpm for 10 min. The pellet was washed with 500
µL ice-cold ethanol (75%) and centrifuged at 13,200 rpm for 5 min. The pellet was air-
dried for 15 min and was then dissolved in 50 µL TE before being stored at -20 °C.
TRAP amplification:
The PCR reaction was conducted using 96-well plates holding 15 µL per well,
~50 ng DNA, 1.5 µL Qiagen 10x buffer, 1.5 µL 25mM MgCl2, 1 µL 5mM dNTPs, 0.3 pmol of each fluorescently labeled arbitrary primer, 1 pmol fixed primer, and 1.5 units of
Taq polymerase. The reaction ran at 94 °C for 2 min; 5 cycles of 94 °C for 45 s, 35 °C for 45 s, 72 °C for 1 min; 35 cycles of 94 °C for 45 s, 50 °C for 45 s, 72 °C for 1 min; and finally 72 °C for 7 min. After loading dye was added to the TRAP products, the samples were loaded onto sequencing gels in LI-COR sequencers, which recorded digital images of the fluorescent banding patterns.
Gel analysis:
The gel images were analyzed using Crosschecker (Buntjer, 1999). Any data points that were not clearly present or absent were manually designated as missing data to avoid biasing the results. The binary interpretation was transferred to NTSYSpc 2.11S
(Rohlf, 2000), in which a matrix of ‘‘simple matching coefficients’’ was generated
27 assigning a numerical value to the similarities between each pair of individuals. Then a dendrogram was generated using the UPGMA method of the SAHN function. All the dendrograms that could be produced from different combinations of tied similarity values were combined by majority rule into a consensus dendrogram with branch probabilities indicating the percentage of dendrograms that contain that subset. Using the cophenetic values and matrix comparison modules of NTsys, the cophenetic correlation was generated and the r value was reported as an estimate of the cluster’s ‘‘goodness of fit.’’
28
Table 2.1: Pelargonium Accession Details. Accessions are numbered with OPGC identification numbers, source refers to where OPGC acquired the accession, and the cultivar column after the species identification includes cultivars in single quotes, breeding numbers as provided by the source, and parentage as provided by the source.
29 # Source Species Cultivar/Breeding # [Parentage] Ploidy/scent 511 Shady Hill P. ×fragrans 'Apple' (scented) 529 Shady Hill P. ×fragrans 'Golden Nutmeg' (scented) 542 Shady Hill P. ×fragrans 'Nutmeg' (scented) 543 Shady Hill P. ×fragrans 'Old Spice' (scented) 526 Shady Hill P. ×fragrans 'Snowy Nutmeg' (scented) 346 R. Craig P. ×fragrans (scented) 495 Shady Hill P. ×hortorum 'Frank Headley' 2x 238 R. Craig P. ×hortorum 'Juliette' syn 'Risque' [80-191-3 x 'Honseler's Glorie Rot'] 4x 548 Shady Hill P. ×hortorum 'Madame Salleron' 2x 503 Shady Hill P. ×hortorum 'Petals' 2x 505 Shady Hill P. ×hortorum 'Wilhelm Langguth' 2x 240 R. Craig P. ×hortorum 203 syn. 81-344-4 ['Honseler's Glorie Rot' x 'Karminball'] 4x 218 R. Craig P. ×hortorum 60-58-5 [G 9-18] 4x 265 R. Craig P. ×hortorum 71-17-7 2x 221 R. Craig P. ×hortorum 78-125-4 4x 219 R. Craig P. ×hortorum 78-51-5 2x 222 R. Craig P. ×hortorum 79-32-5 ['Stadtbern' x 'Wilhelm Langguth'] 2x 223 R. Craig P. ×hortorum 79-33-6 [('Stadtbern' x 'Wilhelm Langguth') x 'Berlin'] 2x 226 R. Craig P. ×hortorum 80-167-26 4x 234 R. Craig P. ×hortorum 81-18-2 ['Stadtbern'] 2x 285 R. Craig P. ×hortorum 82-127-6 [80-196-1 x 80-210-45] 4x 283 R. Craig P. ×hortorum 83-185-32 2x 287 R. Craig P. ×hortorum 83-186-11 2x 282 R. Craig P. ×hortorum 83-186-4 2x 338 R. Craig P. ×hortorum 837 4x 256 R. Craig P. ×hortorum 86-106-15 ['Jean Billes'] 4x 273 R. Craig P. ×hortorum 86-33-17 [82-107-1] 4x 244 R. Craig P. ×hortorum 86-54-26 4x 337 R. Craig P. ×hortorum 87-16-1 2x 284 R. Craig P. ×hortorum 'Jubilee' syn. 175 83-16-3 ['Jean Billes' x 'Honseler's Glorie Lila'] 4x 264 R. Craig P. ×inquinans 86-22-1 syn. G 630 2x 258 R. Craig P. caylae 'Caylae' 4x 417 R. Craig P. caylae Steu 2198 4x 550 Shady Hill P. cotyledonis 531 Shady Hill P. crispum 'French Lace' (scented) 523 Shady Hill P. crispum 'Lemon' (scented) 551 Shady Hill P. dasycaule 514 Shady Hill P. denticulatum 'Balsam' (scented) 516 Shady Hill P. graveolens 'Chocolate Mint' (scented) 519 Shady Hill P. graveolens 'Peppermint Rose' (scented) 524 Shady Hill P. graveolens 'Snowflurry' (scented) 549 Shady Hill P. sidoides 512 Shady Hill P. sp. 'Apple Cider' (scented) 518 Shady Hill P. sp. 'Cloves' (scented) 527 Shady Hill P. sp. 'Lavender Lad' (scented) 535 Shady Hill P. sp. 'Pine' (scented)
Table 2.1
30 Results and Discussion
In evaluating TRAP’s potential to discriminate a large population of Pelargonium accessions, two issues were considered. The first was isolation of high-quality DNA for
TRAP reactions, as many Pelargonium species produce phenolic compounds. After several rounds of testing, high-quality Pelargonium DNA was produced by combining the multiple washes of an existing procedure by Tel-zur et al. (1999) with increased antiphenolic chemicals from a procedure by Saghai-Maroof et al. (1984). The second issue was whether TRAP markers could distinguish Pelargonium accessions and group them according to phenotypes. Appropriate gel images were obtained from most of the
46 accessions using two fixed TRAP primers that have already been used in other plant species (Hu et al., 2005). Two dendrograms from separate data sets placed P. ×
hortorum with related accessions and formed a separate group including scented species
and cultivars. Our results demonstrate that most of the Pelargonium accessions analyzed
could be differentiated from each other using the TRAP markers and that groups formed
based on similar results from TRAP were consistent with what was previously known
about the accessions.
Optimization of DNA extraction method for TRAP reaction:
A basic CTAB method was adequate for many Pelargonium accessions; however,
it failed with almost every cultivar that was not a zonal type Pelargonium, especially the
scented species. In tests of several methods, a procedure by Saghai-Maroof et al. (1984),
which was in use in our laboratory for rice DNA extractions, produced DNA from some
plants for which the basic CTAB method failed. Because there were still many
31 accessions from which DNA could not be extracted, increased the concentrations of antiphenolic chemicals were used along with the multiple-wash technique used by Tel- zur et al. (1999). In comparison with the procedure by Tel-zur et al. (1999), our procedure’s most significant differences were the addition of sorbitol, SDS, PVPP, and proteinase K directly to the high-salt CTAB solution, in which the concentration of Tris-
HCl was doubled. In extractions that excluded the multiple-wash technique, some samples still failed despite the extra antiphenolic chemicals. We determined that both high levels of antiphenolic compounds and multiple washes were essential for DNA extraction from scented-leaved Pelargonium accessions.
All of the DNA used for these TRAP analyses was produced using the above technique. However, the DNA concentration was sometimes too low, so additional extractions were completed using three times the leaf tissue and extraction buffer for the first three washes. In those cases, the amount of both buffers was doubled for the CTAB step. To reach a higher throughput, we took advantage of the recommendation of Tel-zur et al. (1999) that the DNA could be stored at –20 °C before continuing the procedure.
This provided an opportunity to complete the procedure up to that point multiple times and then complete the remainder of the procedure with all those individual samples. The volumes used in the second half of the procedure allow more samples to be processed simultaneously, so the partially processed samples could be collected in the freezer up to the maximum number for the second half of the procedure. In most cases, eight samples at a time were processed before we completed the procedure with 16–32 separate samples collected from two to four repetitions of the first half of the procedure.
32 TRAP analysis of 46 Pelargonium accessions:
A sample image of the TRAP amplification is shown in Figure 2.1. This image
was generated by a fixed primer (QHF6H21L) designed against a sunflower EST and an
arbitrary primer Sa12–700. Of the 46 samples shown in this gel image, most of those in
the left half of the gel are similar accessions from the same species (P.xhortorum), and
those on the right half of the gel include additional Pelargonium species. This image demonstrates that many fragments were amplified in a single run and that most of the accessions could be differentiated on the basis of those fragments. The images were scored by two separate researchers to generate binary data sets by scoring the presence/absence of individual bands across the 46 accessions for further analysis with
NTSYSpc2.
To compare the consistency of the polymorphic patterns generated by different primer combinations with these 46 accessions, two dendrograms were constructed (Fig.
2.2 and 2.3). One was based on 148 bands amplified by the primer combination w3a-700
(QHF6H21L + Sa12-700) (Fig. 2.2), and the second was based on 160 bands amplified by two primer combinations, w1-700 (QHA21B09a + Sa12-700) and w1-800
(QHA21B09a + Ga5-800) (Fig. 2.3). The population being analyzed was intentionally diverse in order to evaluate the applicability of this procedure to as much of the population as possible. This had the added effect of producing some polymorphism in every marker scored. In the data set that combined w1-700 and w1-800 data, 107 of the
160 markers were present in more than one individual. In the other data set, every scorable marker was present in multiple individuals. The dissimilarity coefficient for the first branch point of both dendrograms was 92% (see the scales in Fig. 2.2 and 2.3),
33 indicating that the TRAP markers easily differentiated the most diverse Pelargonium
accessions. In addition, most of the accessions that clustered together were
differentiated. For some accessions, the dissimilarity coefficient is 8% on at least one of
the two trees, implying that those accessions are very similar. Defining how similar
accessions must be to be considered redundant will be an important step to applying this
technique to sort the OPGC collection. Similarities between the two dendrograms
suggest that analysis of the whole collection will be possible.
Because the two dendrograms were produced from the same accessions, similar
groupings of the accessions could be expected. In fact, many similarities between the
two dendrograms were observed. All 24 P.xhortorum accessions grouped together with
the exception of 548, which was the last individual in the P.xhortorum section on one tree
(Fig. 2.2) and mixed in with the scented accessions on the other tree (Fig. 2.3). Within
the P.xhortorum cluster, individuals with some shared parentage (based on pedigrees from Richard Craig) clustered together. For example, accessions 222 and 223 are both derived from cultivars Stadtbern and Wilhelm Langguth, and on both dendrograms they group together. The dendrogram from the w3a-700 primer set also clusters accession 234
nearby; it has parentage from ‘Stadtbern’. Accession 505, with parentage traced to
‘Wilhelm Langguth’, is separate from that cluster in both trees even though it shares
some parentage. Both dendrograms have accessions related to ‘Honseler’s Glorie Rot’ or
‘Jean Billes’ clustered somewhat closely (see related accessions identified in Table 2.1).
Only accession 256 escaped from that cluster on the dendrogram from the w3a-700
34 primer set, but it clustered with the rest of the group (238, 240, 244, 284, 285, 338) on the other dendrogram. Accession 256, a descendant of a colchicine induced tetraploid, is also the only tetraploid to cluster in the primarily diploid section of the dendrogram from the w3a-700 primer set.
The other 22 accessions included 18 accessions that were divided into distinct clusters related to the species of scented cultivars. Those clusters were consistent in both dendrograms, demonstrating that the polymorphic patterns generated by different primer combinations represent the same phylogenetic relationships. One of these clusters contained primarily P. × fragrans accessions, which were divided into two subsections on both dendrograms. There was also a pair of P. crispum accessions that clustered together; one cluster included the two P. caylae accessions; and another cluster had three
P. graveolens accessions. The highly similar results in both dendrograms suggest that the unidentified plants may represent the same species in each respective cluster. For example, the ‘Cloves’ cultivar is grouped close to P.xfragrans accessions in both dendrograms. These results suggest that TRAP will be helpful with the classification of the other unidentified accessions in our collection.
Three of the 46 samples (accessions 512, 535, and 542) did not amplify well as they consistently had only a few faint bands. The only P. inquinans accession (264) in this data set only had sufficient amplification to be considered a practical part of the w3a–700 dendrogram. In the TRAP reactions with the w1-700 and w1-800 primer sets, insufficient amplification produced faint and limited bands from 264 just like the other three. The first three were removed from the analysis in both data sets, and 264 was
35 removed from the w1–700/w1–800 data set. Because so many of our accessions were
hybrids, some overlap is visible in Fig. 2.3. One of the P. × hortorum accessions (548)
is in the part of the dendrogram filled with scented species. This is not too surprising
because 548 is also the very last accession to cluster with the P. × hortorum accessions in
Fig. 2.2, and hybrids are also present in the scented accessions. An analysis of how well the dendrograms produced from those data sets represent the data used to generate them was completed through the cophenetic correlation. Because both dendrograms had r values approaching 1, they were both closely fit to their data sets.
The first five cycles of the TRAP PCR reaction are less stringent, so there does not have to be a perfect match for a fragment to be produced with the fixed primer. The
abundant fragments produced in combination between the fixed primer and one of the
fluorescently labeled arbitrary primers provide sufficient data to differentiate the
individuals in a population with only one or two reactions. This is a significant
advantage when the population to be analyzed includes many individuals. In addition,
the results can be even more useful when the target of the fixed sequence is a gene
relevant to the population. The results using a fixed gene targeting a specific gene of
interest in combination with a segregating population contain some fragments that can be
used as markers for that gene of interest. For example, the telomere-specific primers Hu
(2006) used to map the ends of linkage groups. This gives TRAP a significant advantage
over other methods for mapping purposes because there is an increased chance (over
methods using random markers) of finding markers for the gene targeted by the fixed
primer.
36 In summary, we have optimized the DNA extraction method and established the
TRAP protocols for Pelargonium species. The two dendrograms generated from two different data sets were similar, suggesting that only a few TRAP amplifications could be enough to generate a sufficient number of markers to classify the OPGC collection of
Pelargonium species. The grouping of these Pelargonium accessions was consistent with the species of the individuals. Because OPGC has many accessions of Pelargonium species to be classified, TRAP will be a useful tool for efficiently accomplishing this task. Using TRAP results from the entire OPGC Pelargonium population, TRAP’s effectiveness in species delineation could be compared with the classic Pelargonium systematics of accessions of identified species. TRAP results could then provide an initial categorization for unknown/unidentified accessions clustering with accessions of identified species. In addition, TRAP’s potential for identifying markers that are specific to the gene of interest should allow screening of the population with primers for horticulturally relevant traits, such as pest and disease resistance or differences in essential oil composition.
37
Figure 2.1: A sample TRAP image generated by the primer combination w3a-700 (QHF6H21L + Sa12-700). Most of the samples on the left half of this image are from P. ×hortorum while the right half represents several different species. (From left to right: 218, 219, 221, 222, 223, 226, 234, 238, 240, 244, 256, 258, 264, 265, 273, 282, 283, 284, 285, 287, 337, 338, 346, 417, 495, 503, 505, 511, 512, 514, 516, 518, 519, 523, 524, 526, 527, 529, 531, 535, 542, 543, 548, 549, 550, 551) The DNA size standard of Li-Cor (50 to 700 bp) is loaded in the last lane on the right.
38
Figure 2.2: Dendrogram from the w3a-700 primer set. P. ×hortorum accessions are
encased in a box on the dendrogram. Within that box, dotted circles indicate diploids and solid circles indicate tetraploids; in the list, the same plants are boxed according to shared parentage. Remaining clusters of matching species are circled on the dendrogram. The branch probabilities are indicated in a box for each branch.
Cophenetic correlation: r =-0.91163
39 218 P.×hortorum 85.7 221 P.×hortorum 219 P.×hortorum 92.9 273 P.×hortorum 222 P.×hortorum 92.9 223 P.×hortorum 226 P.×hortorum 75.0 285 P.×hortorum 89.3 57.1 338 P.×hortorum 234 P.×hortorum 238 P.×hortorum ‘Juliette’ syn ‘Risque’ 85.7 240 P.×hortorum 244 P.×hortorum 89.3 64.3 284 P.×hortorum 'Jubilee' 264 P. inquinans 265 P.×hortorum 92.9 67.9 337 P.×hortorum 92.9 282 P.×hortorum 71.4 287 P.×hortorum 503 P.×hortorum ‘Petals’ 75.0 283 P.×hortorum 92.9 505 P.×hortorum ‘Wilhelm Langguth’ 82.1 256 P.×hortorum 85.7 495 P.×hortorum ‘Frank Headley’ 548 P.×hortorum ‘Mme. Salleron’ 89.3 258 P. caylae 96.4 417 P.caylae 89.3 514 P.denticulatum ‘Balsam’ 100 519 P. graveolens ‘Peppermint Rose’ 92.9 516 P. graveolens ‘Chocolate Mint’ 92.9 524 P. graveolens ‘Snowflurry’ 92.9 551 P. dasycaule 346 P.×fragrans 100 511 P.×fragrans ‘Apple’ 100 518 P. sp. ‘Cloves’ 100 100 526 P.×fragrans ‘Snowy Nutmeg’ 529 P.×fragrans ‘Golden Nutmeg’ 100 100 543 P.×fragrans ‘Old Spice’ 527 P. sp. ‘Lavender Lad’ 100 96.4 549 P.sidoides 96.4 550 P.cotyledonis 523 P. crispum ‘Lemon’ 100 531 P.crispum ‘French Lace’
40
Figure 2.3: Dendrogram from w1-700 and w1-800 primers sets. P. ×hortorum accessions are encased in a box on the dendrogram. Clusters, ploidy levels and branch probabilities are indicated in the same manner as in Figure 2.2. Note: Accession 264 is not included because it did not have sufficient data, and an extra box is drawn around the
P. ×hortorum accession that is separated from the others. Cophenetic correlation: r =-
0.89712
41
218 P.×hortorum 100 219 P.×hortorum 66.7 221 P.×hortorum 222 P.×hortorum 223 P.×hortorum 238 P.×hortorum ‘Juliette’ syn ‘Risque’ 100 66.7 287 P.×hortorum 100 338 P.×hortorum 495 P.×hortorum ‘Frank Headley’ 66.7 226 P.×hortorum 100 240 P.×hortorum 265 P.×hortorum 100 337 P.×hortorum 100 244 P.×hortorum 100 273 P.×hortorum 66.7 284 P.×hortorum 'Jubilee' 100 285 P.×hortorum 100 256 P.×hortorum 66.7 282 P.×hortorum 100 283 P.×hortorum 100 505 P.×hortorum ‘Wilhelm Langguth’ 503 P.×hortorum ‘Petals’ 234 P.×hortorum 346 P.×fragrans 100 100 100 511 P.×fragrans ‘Apple’ 100 518 P. sp. ‘Cloves’ 543 P.×fragrans ‘Old Spice’ 100 514 P.denticulatum ‘Balsam’ 100 551 P. dasycaule 100 100 527 P. sp. ‘Lavender Lad’ 100 100 549 P.sidoides 526 P.×fragrans ‘Snowy Nutmeg’ 100 100 529 P.×fragrans ‘Golden Nutmeg’ 548 P.×hortorum ‘Mme. Salleron’ 100 258 P. caylae 100 417 P.caylae 550 P.cotyledonis 516 P. graveolens ‘Chocolate Mint’ 100 519 P. graveolens ‘Peppermint Rose’ 100 524 P. graveolens ‘Snowflurry’ 100 523 P. crispum ‘Lemon’ 100 531 P.crispum ‘French Lace’
42
Figure 2.4: Photographs of Pelargonium accessions. Zonal P. ×hortorum (A, left to right, top: 222, 256, 223, middle: 238, 338, 221, 284, bottom: 234, 218, 219), continuing clockwise to the next picture P. caylae (B, top: 258, bottom: 417), followed by variegated
P. ×hortorum (C, top: 495, 503, bottom: 505, 548), plants grouped with P. ×fragrans (D, top: 518, 527, bottom: 549, 550), followed P. ×fragrans (E, top: 511, 543, 529, bottom:
346, 542, 526), P. graveolens and closely grouped (F, top: 514, 516, middle: 524, bottom:
551, 519), P. crispum (G: 523, 531), and finally in the middle P. inquinans (H: 264)
43 A
G
F
H B
E D C
Figure 2.4
44 CHAPTER 3
IDENTIFYING THE MOST DIVERSE SUBSET OF OPGC’S PELARGONIUM
COLLECTION
CHAPTER 3: Identifying the most diverse subset of OPGC’s Pelargonium collection
Abstract:
The task of reducing a large collection of germplasm, while maintaining as much of the diversity as possible, is challenging. The Pelargonium collection at the Ornamental Plant
Germplasm Center (OPGC) contains about 800 accessions, the majority of which have no species designation or pedigree. Those accessions without sufficient background information need to be categorized according to their relationships with other accessions, so that OPGC’s Pelargonium collection may be reduced by 75% while maintaining as much of the genetic diversity as possible. In this study, Target Region Amplification
Polymorphism (TRAP) was selected as a molecular screening tool for evaluation of
OPGC’s Pelargonium collection. Two targeted primers, QHF6H12L and TeloRA, without a link to a functional Pelargonium gene were used so as to avoid biasing the analysis with genes under functional constraint. Accessions belonging to section
Ciconum were separated from those in other taxonomic sections in a neighbor-joining tree. A Bayesian analysis of the “non-Ciconum” part of the collection clustered the accessions by taxonomic section. A separate Bayesian analysis of potentially duplicated accessions within section Ciconum confirmed the identity of about half of the potentially
45 duplicated accessions. Related accessions within section Ciconum were analyzed with both the previous primers, and new targeted primers (ArthR3, ArthR19, LRR4 and
EthylrecpR4) designed based on insect and disease resistance genes. Although the new primers did not yield any TRAP fragments that could be used to screen for resistance among the Pelargonium accessions, the dendrogram produced using the data from primers based on resistance genes matched the known relationships among the accessions. The dendrogram from the primers targeting resistance genes was more resolved than the dendrogram produced using data from the first two primers. In conclusion, genetic analyses using TRAP markers have shown that 103 accessions represent most of the diversity in OPGC’s collection from sections other than Ciconum, whereas the accessions within section Ciconum include many closely related cultivars and hybrids.
Introduction
The species of Pelargonium in the Ornamental Plant Germplasm Center’s
(OPGC) collection are primarily in sections Ciconum (10 species at OPGC), Pelargonium
(10 species at OPGC) and Reniformia (4 species at OPGC. Section Ciconum includes ivy-leaved and zonal geraniums (Oliver and Van der Walt, 1984; Van der Walt, 1992;
Gibby et al., 1990). Section Pelargonium is likely related to sections Glaucophyllum and
Campylia (Van der Walt, 1992), both of which have at least one representative cultivar or hybrid at OPGC (Table 3.1). In their review of the history of P. × domesticum Loehrlein and Craig (2001) presented likely contributors to the hybrid species, including several species from sections Pelargonium and Glaucophyllum.
46 In addition to accessions with species or pedigree information that identified them as a part of a particular section of Pelargonium, the collection at OPGC also includes many accessions with no pedigree. Without this background information the task of reducing the collection to 25% of its original size is quite challenging. One way to accomplish the purge would have involved randomly removing accessions. However, random selection of accessions to purge could result in retention of closely related and over represented cultivars, and the loss of the more diverse representatives from other sections. A high throughput molecular analysis was required to group non-pedigreed accessions with related accessions, and also to identify redundant accessions.
A preliminary assessment of TRAP markers on 46 accessions from OPGC’s collection revealed the utility of this method for categorization of the collection at both the species and cultivar level (Fig. 2.2 and 2.3). The cultivars were clustered by species and with other accessions that shared the same parentage. Figure 2.4 shows pictures of some of those accessions grouped according to the TRAP results from the test run
(Palumbo et al., 2007). This study will reveal the categorization of OPGC’s entire collection of Pelargonium accessions into taxonomic sections by TRAP markers, and compare the dendrograms with previous phylogenies that have grouped species according to taxonomic sections (James et al. 2004; Bakker et al. 2000 and 2004). Additionally, this study will examine clustering patterns among section Ciconum accessions with known relationships using TRAP primers based on resistance genes.
47 Materials and Methods
Plant material and DNA extraction:
In scaling up the procedure to the whole collection, high through-put Qiagen kits
were selected for the DNA extraction, and all stages of the reaction (tissue collection –
PCR) were completed in 96-well plates. DNA from 775 Pelargonium accessions of the
~800 accessions stored at OPGC in 2004 was extracted using Qiagen (96 well) plant
DNAeasy kits. OPGC’s accessions included species from many accessions of the genus
as well as an extensive representation of accessions from within section Ciconum. Some of the latter accessions had known pedigrees, which identified the species or cultivars used in their breeding, and others were only identified as being members of the genus
Pelargonium. The 123 accessions with species identified in their pedigrees, other than P.
× hortorum accessions without a cultivar name, are listed in Table 3.1.
Some of the pedigree information available suggested that there were some duplicate accessions within OPGC’s collection, either from repeated cultivar names or sports and virus-induced cultivars. Those potential duplicates are listed in Table 3.2.
Among them, 85 accessions were selected that represented 26 groups of potential duplicates from within section Ciconum, closely related accessions, and all identified section Ciconum species. Those accessions were analyzed as one data set using the data collected during the analysis of the entire collection. A separate dataset of section
Ciconum accessions was based on 88 selections that included groups of accessions with related pedigrees (Fig. 3.1), accessions with known resistance or susceptibility traits
(Table 3.3), and the same section Ciconum species as the duplicate set.
48 TRAP primers
The primers selected for TRAP analysis of the entire collection were based on
Jinguo Hu’s experience that these primers, which were developed for other projects in his
laboratory (Hu et al., 2005), produced scorable TRAP markers for most plant families.
These primers were intended to estimate relationships between accessions rather than the
evolution of a specific gene of interest. The two targeted primers were QHF6H12L,
targeting a sunflower EST homologous to the BEL1-like homeobox protein and already
tested on Pelargonium accessions in Palumbo et al. (2007), and TeloRA, targeting the
telomere region with an extra “A” to avoid repetition. The arbitrary primers for
QHF6H12L were Trap03-700 (CGTAGCGCGTCAATTATG) and Trap13-800
(GCGCGATGATAAATTATC), and the arbitrary primers for TeloRA were Sa12-700
(TTCTAGGTAATCCAACAACA) and Ga5-800 (GGAACCAAACACATGAAGA).
New targeted primers were designed for the final analysis of section Ciconum,
which matched approximately 20 bp (with a melting point close to 50 °C) from three
resistance-related sequences: a fatty acid desaturase, a TIR-NBS-LRR gene, and an
ethylene receptor. Two targeted primers were designed based on the sequence for the ∆9
14:0-acyl carrier protein fatty acid desaturase gene, which is involved in the production
of the anacardic acid that causes arthropod resistance in Pelargoniums (Schultz et al.,
1996): ArthR3-AAAATATTACGGAGTACCCT, and ArthR19-
TACACCAGTTACTCCAACTT. Two other targeted primers were designed as back-up
to the arthropod resistance gene, which also had potential for disease resistance. LRR4-
ACGTAATCTGCAAGTAAAAG, based on the a sequence Borhan et al. (2004) reported
49 for an Arabidopsis thaliana TIR-NBS-LRR (leucine rich repeat) gene, was selected
because the rust resistance it provides to Arabidopsis might be relevant to some rust
resistant Pelargonium accessions. EthylrecpR4-ATAAGAGTAGCCAAAAGGTT,
based on a conserved region between two ethylene receptor homologs from Pelargonium
xhortorum (Dervinis et al., 2000), was selected because of its relevance to petal
abscission. ArthR3 and EthylrecepR4 were each matched with the pair of labeled
primers: ga3.800-TCATCTCAAACCATCTACAC and sa4.700-
TTCTTCTTCCCTGGACACAAA from Dr. Jinguo Hu. ArthR19 and LRR4 were each
matched with the pair of labeled primers: odd8.800-CACAAGTCGCTGAGAAGG and
odd15.700-GCGAGGATGCTACTGGTT also from Dr. Jinguo Hu.
TRAP amplification:
The PCR reaction was conducted using 96-well plates holding 15 µL per well,
~25 ng DNA, 1.5 µL Qiagen 10x buffer, 1.0 µL 25mM MgCl2, 0.5 µL 10mM dNTPs, 0.3 pmol of the 700nm fluorescently labeled arbitrary primer, 0.5 pmol of the 800nm fluorescently labeled arbitrary primer, 3 pmol fixed primer, and 1.5 units of Taq polymerase. The reaction ran at 94°C for 2 min; 5 cycles of 94°C for 45 s, 35°C for 45 s,
72°C for 1 min; 35 cycles of 94°C for 45 s, 50°C for 45 s, 72°C for 1 min; and finally
72°C for 7 min. At the completion of the analysis 7 µL loading dye was added to each well. TRAP amplification was the same in all analyses except that the final analysis of section Ciconum with new primers contained Qiagen 10x+CL buffer instead of the standard buffer. The new buffer eliminated the need for a gel loading buffer, and so
50 reduced the appearance of “smile” distortions in the gel image from differences in dye between ladder and samples. Also, there was a DNA-free blank sample included as a control for these new primers to ensure that the fragments were not a result of primer dimers.
Gel and data analysis:
The PCR products were analyzed using digital images from LI-COR sequencers, as described in Palumbo et al. (2007), except that only the first 60-90 bands larger than
50 bp were included. That range has the most readily scored markers, since the “smile” effect becomes increasingly apparent in longer fragments. Fragments were scored in a binary interpretation as present (1) or absent (0) using Crosschecker (Buntjer, 1999), and any unclear fragments were scored as missing data. The data was converted into nexus format, and analyzed using PAUP (Swofford, 2002). A distance-based analysis was performed using the neighbor joining method and the Nei-Li coefficient. Jackknife values were calculated from 1000reps. For the final dendrogram ties were broken randomly, and both the higher probability branches and any lower probability branches that did not change the structure of the tree were included.
The resulting neighbor joining tree was used to assign non-pedigreed accessions to section Ciconum or to a non-Ciconum group. In order to achieve better resolution of the relationships between taxonomic sections, all accessions not in section Ciconum (Fig.
3.2) were analyzed as a separate subset along with a small group of known section
Ciconum plants. Bayesian analysis was selected because of the statistics available for the convergence of multiple runs. The analysis of the entire collection analysis was
51 completed using MrBayes (Huelsenbeck and Ronquist, 2001; Ronquist and Huelsenbeck,
2003; Altekar et al., 2004) for three simultaneous runs with 10 chains each, attempting 25 swaps per generation, for 20 million generations saving every 100th tree. Analysis of the potential duplicates was completed in two runs after 24 million generations, and related accessions from section Ciconum were analyzed based on data from the first primer set in three runs of 14 million generations with all other parameters unchanged. The data from the new primers for the related accessions from section Ciconum was analyzed by 44 million generations of Bayesian analysis in two simultaneous runs with all other parameters the same. Data from the new primers was collected in the same way as with the first primer set, except that the data was scored as “present” for any band regardless of intensity. That way missing data was only recorded where the appearance of the gel image appeared ambiguous from anomalies in the gel.
In all of the Bayesian analyses, the number of trees to skip, or “burn-in,” from the beginning of the runs was determined by the point at which the average standard deviation of split frequencies leveled off at the end of the run. The parameters were summarized for the selected burn-in rate using the sump function graph of the log probability to ensure the burn-in was high enough. Finally, the trees saved after the burn- in point were summarized in a consensus tree showing the percent of the trees with each branch and averaged branch-lengths.
The success of categorization for the unknown accessions was estimated by comparison to the categorization of known accessions. Pelargonium species were compared to phylogenies from other molecular analyses that grouped individual
52 examples of each species according to taxonomic sections (James et al. 2004; Bakker et al. 2000 and 2004). In addition, accessions with recorded breeding histories were used to help determine the accuracy of clustering accessions at the cultivar level.
53
Table 3.1: OPGC Pelargonium accessions with known pedigree.- listed by taxonomic sections (Synonyms and Sections are 54 from the USDA-ARS Germplasm Resource Information Network: GRIN www.ars-grin.gov)
ID Section Species Cultivar at0559 Ciconum acetosum L'Hér br0558 Ciconum burtonii (burtoniae L. Bolus =steno petalum Ehrh.) st0566 Ciconum stenopetalumEhrh. cy0258 Ciconum caylae Humbert 'Caylae' cy0417 Ciconum caylae Humbert ft0455 Ciconum frutetorum R.A. Dyer ft0612 Ciconum frutetorum R.A. Dyer in0264 Ciconum inquinans (L.) L'Hér. in0560 Ciconum inquinans (L.) L'Hér. pt0383 Ciconum peltatum (L.) L'Hér. pt0407 Ciconum peltatum (L.) L'Hér. pt0410 Ciconum peltatum (L.) L'Hér. pt0411 Ciconum peltatum (L.) L'Hér. pt0800 Ciconum xhortorum L.H. Baileyx peltatum (L.) L'Hér. 'Madeline Crozy'
55 pt0403 Ciconum xhortorum L.H. Baileyx peltatum (L.) L'Hér. pt0415 Ciconum xhortorum L.H. Baileyx peltatum (L.) L'Hér. pt0387 Ciconum (xhortorumL.H. Bailey xp eltatum (L.) L'Hér.) x xhortorum L.H. Bailey pt0373 Ciconum xhortorum L.H. Bailey x (xhortorumL.H. Bailey xp eltatum (L.) L'Hér.) pt0385 Ciconum xhortorum L.H. Bailey x (xhortorumL.H. Bailey xp eltatum (L.) L'Hér.) pt0418 Ciconum xhortorum L.H. Bailey x (xhortorumL.H. Bailey xp eltatum (L.) L'Hér.) pt0374 Ciconum xhortorum L.H.Bailey x {xhortorumL.H. Bailey x (xhortorum L.H. Bailey xp eltatum (L.) L'Hér.)} pt0402 Ciconum xhortorum L.H.Bailey x {xhortorumL.H. Bailey x (xhortorum L.H. Bailey xp eltatum (L.) L'Hér.)} pt0404 Ciconum xhortorum L.H.Bailey x {xhortorumL.H. Bailey x (xhortorum L.H. Bailey xp eltatum (L.) L'Hér.)} pt0413 Ciconum xhortorum L.H.Bailey x {xhortorumL.H. Bailey x (xhortorum L.H. Bailey xp eltatum (L.) L'Hér.)} pt0372 Ciconum (xhortorumL.H. Bailey x xhortorumL.H. Bailey) x (xhortorum L.H. Bailey xp eltatum (L.) L'Hér.) ht0394 Ciconum xhortorum L.H. Bailey(inquinans (L.) L'Hér. x zonale (L.) L'Hér.) 'Ben Franklin' ht0389 Ciconum xhortorum L.H. Bailey(inquinans (L.) L'Hér. x zonale (L.) L'Hér.) 'Ben Franklin' sport Table 3.1 (continued on text page)
Table 3.1: (continued)
ID Section Species Cultivar ht0230 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér. x zonale (L.) L'Hér.) 'Calypso' ht0340 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Dixieland' ht0416 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Dolly' ht0419 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Fox' ht0495 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Frank Headley' ht0369 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Gail' ht0395 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Hildegaard' ht0284 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Jubilee' ht0238 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Juliette' syn 'Risque' ht0548 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Madame Salleron' ht0279 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Misty' ht0366 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Nittany Lion Red' 56 ht0392 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'North Star' ht0235 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Paris' ht0347 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Pascal' ht0503 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Petals' ht0334 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Siren' ht0398 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Stadtbern' ht0505 Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 'Wilhelm Langguth' ht#### Ciconum xhortorum L.H. Bailey (inquinans (L.) L'Hér.x zonale (L.) L'Hér.) 80 accessions mo0564 Ciconum monstrosumL'Hér. = xhortorumL.H. Bailey fl0491 Ciconum xfloribunda x hortorumx (floribundas patented as xhortorum L.H. Bailey) kw0567 Ciconum xkewense R. A. Dyer (zonale (L.) L'Hér. x inquinans(L.) L'Hér.) zn0263 Ciconum zonale (L.) L'Hér. zn0799 Ciconum zonale (L.) L'Hér. ec0557 Cortusina echinatum Curtis fu0565 Glaucophyllum fruticosum (Cav.) Willd. (continued on next page)
Table 3.1: (continued)
ID Section Species Cultivar rp0580 Hoarea rapaceum (L.) L'Hér. 'Mrs. Kingsley' ct0550 Isopetalum cotyledonis (L.) L'Hér. te0556 Jenkinsonia tetragonum (L. f.) L'Hér. ds0551 Otidia dasycaule Haw. = crithmifolium Sm. fe0563 Otidia ferrolaceum (ferulaceum (Burm. F.) Willd. = carnosum (L.) L'Hér.) ac0517 Pelargonium acerifolium L'Hér = cucullatum subsp. strigifolium Volschenk 'Citronella' ad0534 Pelargonium adcifolium (acerifolium L'Hér = cucullatum subsp. strigifolium Volschenk) 'Snowflake' cp0538 Pelargonium capitatum L'Hér 'Attar Of Roses' cr0531 Pelargonium crispum (P. J. Bergius) L'Hér 'French Lace' cr0523 Pelargonium crispum (P. J. Bergius) L'Hér 'Lemon' cr0536 Pelargonium crispum (P. J. Bergius) L'Hér 'Lemon Crispum' dt0514 Pelargonium denticulatum Jacq. 'Balsam' fc0598 Pelargonium filicifolium hort. = denticulatum 'Filicifolium' Jacq. 'Fern Leaf' 57 gl0597 Pelargonium glutinosum (Jacq.) L'Hér. 'Pheasant's Food' qu0582 Pelargonium quercifolium (L. f.) L'Hér. 'Fair Ellen' qu0507 Pelargonium quercifolium (L. f.) L'Hér. 'Staghorn Oak' rd0599 Pelargonium radens H.E. Moore 'Crowfoot' sc0602 Pelargonium scabrum (L.) L'Hér. 'Apricot' gr0516 Pelargonium graveolens L'Hér. 'Chocolate Mint' gr0533 Pelargonium graveolens L'Hér. 'Mint-Scented Rose' gr0530 Pelargonium graveolens L'Hér. 'Old Fashioned Rose' gr0519 Pelargonium graveolens L'Hér. 'Peppermint Rose' gr0524 Pelargonium graveolens L'Hér. 'Snowflurry' tb0581 Pelargonium terebinthinaceum (Cav.) Desf = graveolens L'Hér. 'Little Gem' hy0591 Pelargonium tomentosum Jacq. x graveolens L'Hér. 'Joy Lucille' tm0521 Pelargonium tomentosum Jacq. 'Peppermint' fr0511 Reniformia fragrans Willd. 'Apple' fr0529 Reniformia fragrans Willd. 'Golden Nutmeg' fr0346 Reniformia fragrans Willd. 'Juicy Fruit' fr0542 Reniformia fragrans Willd. 'Nutmeg' (continued on next page)
Table 3.1: (continued)
ID Section Species Cultivar fr0543 Reniformia fragrans Willd. 'Old Spice' fr0526 Reniformia fragrans Willd. 'Snowy Nutmeg' io0553 Reniformia ionidiflorum (Eckl. & Zeyh.) Steud. re0552 Reniformia reniforme Curtis sd0549 Reniformia sidoides DC. hy0584 denticulatumJacq. x xdomesticum L.H. Bailey Glaucop( hyllum, Pelargonium, et al.) 'Clorinda' gt1382 gratum Willd. 'Cinnamon' hy0547 graveolens L'Hér. x echinatumCurtis (Pelargonium, Cortusina) 'Blandfordianum Roseum' tr0520 torento (= nervosum 'Torento' Sweet Campylia?) 'Ginger' nr0528 Campylia xnervosum Sweet 'Lime' sr0525 Campylia xscarboroviae Sweet 'Strawberry' ga0561 Campylia xglaucifolium Sweet (gibbosum (L.) L'Hér. x lobatum (Burm. F.) L'Hér =Poly actium) 58 ml0522 Campylia xmelissimum Sweet (crispum ( P.J. Bergius) L'Hér. x graveolens L'Hér. =Pelar gonium) xdomesticum L.H. Bailey grandiflo( rum (Andrews) Willd., cucullatum (L.) L'Hér., et al) dm0764 = Glaucophyllum, Pelargonium, et al. 'Allure' xdomesticum L.H. Bailey grandiflo( rum (Andrews) Willd., cucullatum (L.) L'Hér., et al) dm0767 = Glaucophyllum, Pelargonium, et al. 'Camelot' xdomesticum L.H. Bailey grandiflo( rum (Andrews) Willd., cucullatum (L.) L'Hér., et al) dm0769 = Glaucophyllum, Pelargonium, et al. 'Crystal' xdomesticum L.H. Bailey grandiflo( rum (Andrews) Willd., cucullatum (L.) L'Hér., et al) dm0770 = Glaucophyllum, Pelargonium, et al. 'Dandy' xdomesticum L.H. Bailey grandiflo( rum (Andrews) Willd., cucullatum (L.) L'Hér., et al) dm0771 = Glaucophyllum, Pelargonium, et al. 'Dandy' xdomesticum L.H. Bailey grandiflo( rum (Andrews) Willd., cucullatum (L.) L'Hér., et al) dm0772 = Glaucophyllum, Pelargonium, et al. 'Dapper Burgandy' xdomesticum L.H. Bailey grandiflo( rum (Andrews) Willd., cucullatum (L.) L'Hér., et al) dm0773 = Glaucophyllum, Pelargonium, et al. 'Debutante' (continued on next page)
Table 3.1: (continued)
ID Section Species Cultivar xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = dm0774 Glaucophyllum, Pelargonium, et al. 'Debutante' dm0775 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Emperor' dm0776 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Empress' syn. 'Clarissa' dm0777 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Enchantment' dm0778 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Excalibur' dm0779 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Fantasy' dm0780 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Fascination' dm0781 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Fascination' dm0784 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Lois' dm0787 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Maiden Sunrise' dm0788 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Majestic' 59 dm0789 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Monarch' dm0790 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Royal Velvet' dm0791 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Sandra' dm0792 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Splendor' dm0793 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Symphony' dm0794 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Tiara' dm0795 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Tiara' dm0796 xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = G 'Virginia' xdomesticum L.H. Bailey grandifloru( m (Andrews) Willd., cucullatum (L.) L'Hér., et al) = dm0783 Glaucophyllum, Pelargonium, et al.
Table 3.2: Potential duplicates within OPGC’s Pelargonium collection. Duplication predictions are based on cultivar names, 60
or pedigree (sports or virus-induced cultivars) are listed with explanations of what they may duplicate.
OPGC # Species Cultivar potential duplicates similiarity or parentage dm0794 xdomesticum 'Tiara' 'Tiara' #795, #794 dm0795 xdomesticum 'Tiara' 'Tiara' #795, #795 ht0398 xhortorum 'Stadtbern' Stadtbern #398, #1546 sp1546 sp. 'Stadtbern' Stadtbern #398, #1546 sp1363 sp. 'Picotee' Picotee #1363, #1730 sp1730 sp. 'Picotee' Picotee #1363, #1730 sp1362 sp. 'Peacock' Peacock #588, #1362 sp0588 sp. 'Peacock' Peacock #588, #1362 Juliette syn Risque #238, #339, 80-191-3 x 80-207-18 (Honselor's Glorie sp1098 sp. 'Risque' #1098 Rot) 'Juliette' syn Juliette syn Risque #238, #339, 80-191-3 x 80-207-18 (Honselor's Glorie ht0238 xhortorum 'Risque' #1098 Rot) 'Juliette' syn Juliette syn Risque #238, #339, 80-191-3 x 80-207-18 (Honselor's Glorie sp0339 sp. 'Risque' #1098 Rot) sp0259 sp. 'Jubliee' Jubilee #284, #259, #341 'Jean Billes' x 'Honseler's Glorie Lila' 61 ht0284 xhortorum 'Jubliee' Jubilee #284, #259, #341 'Jean Billes' x 'Honseler's Glorie Lila' ht0341 xhortorum 'Jubliee' Jubilee #284, #259, #341 'Jean Billes' x 'Honseler's Glorie Lila' sp1729 sp. 'Freckles' Freckles #1063, #1729 sp1063 sp. 'Freckles' Freckles #1063, #1729 dm0780 xdomesticum 'Fascination' Fascination #781, #780 87-92-1 x 90-12-15 dm0781 xdomesticum 'Fascination' Fascination #781, #780 87-92-1 x 90-12-15 dm0770 xdomesticum 'Dandy' Dandy 90-30-16 #770, #771 90-70-20 x 90-93-20 dm0771 xdomesticum 'Dandy' Dandy 90-30-16 #770, #771 90-70-20 x 90-93-20 sp1224 sp. 'Candy' Candy #1224, #1163 sp1163 sp. 'Candy' Candy #1224, #1163 ht0230 xhortorum 'Calypso' Calypso #1050, #230 Graeffin Mariza sp1050 sp. 'Calypso' Calypso #1050, #230 Graeffin Mariza sp0262 sp. 95-6-1 95-6-1 #262, #799 zn0799 zonale 95-6-1 95-6-1 #262, #799 Table 3.2 (continued on next page)
Table 3.2: (con’t)
OPGC # Species Cultivar potential duplicates similiarity or parentage ht0240 xhortorum 203 (203) #240, #365 'Honseler's Glorie Rot' x 'Karminball' sp0365 sp. 203 (203) #240, #366 'Honseler's Glorie Rot' x 'Karminball' 'Honseler's Glorie Rot' x 'Karminball' syn. ht0241 xhortorum 208 (208) #241, #351 'Glacier Carmen' (80-210-27) 'Honseler's Glorie Rot' x 'Karminball' syn. sp0351 sp. 208 (208) #241, #351 'Glacier Carmen' (80-210-27) ht0449 xhortorum 93-6-11 93-6-11 #449, #491 'Marilyn' x 'Ben Franklin' (Sport#389) xfloribunda x fl0491 xhortorum 93-6-11 (Light Variegation) 93-6-11 #449, #491 'Marilyn' x 'Ben Franklin' (Sport#389) org. BF Sport 'Ben Ben Franklin #394, #1441, 'Wilhelm Langguth' (#505) x 'Snowmass' ht0389 xhortorum Franklin' #389(sport) (#1440) Ben Franklin #394, #1441, 'Wilhelm Langguth' (#505) x 'Snowmass' ht0394 xhortorum 'Ben Franklin' #389(sport) (#1440) 62 Ben Franklin #394, #1441, 'Wilhelm Langguth' (#505) x 'Snowmass' sp1441 sp. 'Ben Franklin' #389(sport) (#1440) ht0505 sp. 'Wilhelm Langguth' sp1440 sp. Snowmass Purple Seedling 279 x 'Dapper Burgundy' sp1101 sp. Royal Velvet Royal Velvet #790, #1101 (#772)(syn. 'Burghi') Purple Seedling 279 x 'Dapper Burgundy' dm0790 xdomesticum 'Royal Velvet' Royal Velvet #790, #1101 (#772)(syn. 'Burghi') dm0772 xdomesticum 'Dapper Burgandy' (VPP) ht0279 xhortorum 'Misty' Misty #279, #343, #1147 80-194-29 (#232) x 80-210-45 sp0343 sp. 'Misty' Misty #279, #343, #1147 80-194-29 (#232) x 80-210-45 sp1147 sp. 'Misty' Misty #279, #343, #1147 80-194-29 (#232) x 80-210-45 ht0232 xhortorum 80-194-29 78-166-5 GX 557-5 (continued on next page)
Table 3.2: (con’t)
OPGC # Species Cultivar potential duplicates similiarity or parentage sp1268 sp. 'Monarch' Monarch #789, #1268 Oglevee No. 425 x 'Hazel Ripple' ('Hazel Ripple' = sport of seedling: Red dm0789 xdomesticum 'Monarch' Monarch #789, #1268 Star #1097 and P. formatosum) sp1097 sp. 'Red Star' sp1043 sp. 'Afterglow' similiarity to Red Star #1097 sp1061 sp. 'Francis James' similiarity to Red Star #1097 sp1100 sp. 'Satellite' similiarity to Red Star #1097 dm0774 xdomesticum 'Debutante' Debutante #773, #774 'Allure' (#764) x 'Fantasy' (#779) dm0773 xdomesticum 'Debutante' Debutante #773, #774 'Allure' (#764) x 'Fantasy' (#779) dm0764 xdomesticum 'Allure' 'White Glory' x 'Melissa' dm0779 xdomesticum 'Fantasy' 'Miss Cherry Vale' x 'Virginia' (#796) dm0796 xdomesticum 'Virginia' sp1494 sp. Roller's Pioneer Harvard #1484 with crocodile virus 63 sp1484 sp. Harvard similiarity to Burgundy Beauty #1479 sp1457 sp. Roller's David similiarity to Harvard #1484 sp1479 sp. Burgundy Beauty similiarity to Harvard #1485 sp1529 sp. Queen of Hearts light pink version of Queen of Hearts sp1532 sp. Ten of Hearts version #1529 sp1473 sp. Stitches Sport of Ten of Hearts #1532 sp1487 sp. Memories Sport of Sybill Holmes #1505 sp1505 sp. Sybil Holmes sp1454 sp. Beauty of Diane Rouletta sport #1460 sp1462 sp. Watercolor light pink Rouletta sport #1460 sp1460 sp. Rouletta sp1461 sp. Spotlite Sally pink form of Rouletta #1460 sp0605 sp. 'Pink Pandora' very light pink Pandora sport #1444 sp0604 sp. 'Red Pandora' bright scarlet Pandora sport #1444 sp1444 sp. Pandora (continued on next page)
Table 3.2: (con’t) OPGC # Species Cultivar potential duplicates similiarity or parentage sp1181 sp. Pink Ice Orion sport? #1177 sp1177 sp. Orion sp1070 sp. Halo Chimera sport of Orange Supreme #1547 much like Mr. Wren #1080 sp1080 sp. Mister Wren sp1547 sp. Orange Supreme A selection from Orange Ricard sp1143 sp. Black Mrs. Cox Mrs. Cox sport #574 sp0574 sp. 'Mrs. Cox' sp0608 sp. 'Greengold' similiarity to Sunspot #609 sp0609 sp. 'Sunspot' Little Darling sport #607 sp0496 sp. 'Frosty' variegated Little Darling sport #607 sp0607 sp. 'Little Darling' ('Kleiner Liebling') sp0499 sp. 'Variegated Little Darling' variegated Little Darling#607 sp1486 sp. Lady Lexington sp1496 sp. Rose Lady of Lexington pink Lady Lexington sport #1486 64 sp0504 sp. 'Platinum' silver leaf Frank Headley sport #495 ht0495 sp. 'Frank Headley' sp0501 sp. 'Flower Of Spring' ('Jayne Varness') sp1123 sp. Attraction Flower of Spring sport #501 sp1449 sp. Kaliedescope Contrast sport #573 sp0573 sp. Contrast' sp0578 sp. 'Red Capri' dark red Capri form #585 Dark red form of Capri sp0585 sp. 'Capri' sp1493 sp. Roller's Pathfinder mesh veined form of Salmon Queen #1498 sp1498 sp. Salmon Queen sp0579 sp. 'Old Scarlet Unique' sp0577 sp. 'White Unique' white Old Scarlet Unique form #579 sp1731 sp. Orange Glow (was another 1082) Orangeade #1082 sp1082 sp. Orangeade Orange Glow #1731 hy0591 hybrid 'Joy Lucille' P. tomentosum x P. graveolens sp0587 sp. 'Variegated Joy Lucille' variegated Joy Lucille #591
Table 3.3: OPGC accessions from section Ciconum with pedigree and/or known resistance/susceptibility traits. 65
Accessions are organized by resistance, as well as parentage, into potentially related groups.
OPGC # Species Cultivar duplicate Parentage resistance 71-17-7(#265) x 85-26-6 F1 Hybrid- ht0350 xhortorum 87-6-1 Resistant (17-7) x Susceptible (10-1) Arthropod Resistant ht0357 xhortorum 87-24-4 71-18-6 rthAropod Resistant ht0336 xhortorum 87-21-1 85-27-5 Inbred Parent (15-4) Arthropod Susceptible sp1238 sp. Frills Spidermite prone sp1277 sp. Orange Marmalade Spidermite prone sp1288 sp. Rebecca Spidermite prone sp1391 sp. Rosette Spidermite prone sp1402 sp. Golden Bird's Egg Spidermite prone sp1404 sp. Magenta Bird's Egg Spidermite prone sp1405 sp. Plenty Spidermite prone requires protection from high sp1310 sp. Gold Dust heat and direct sun Sunbelt Hot Pink sp1379 sp. (formerly Marian) performs in warmth & humidity 66 Sunbelt Rose sp1380 sp. (formerly Hazel) performs in warmth & humidity sp1553 sp. Sunbelt Coral performs in warmth & humidity Sunbelt Salmon sp1554 sp. (Formerly Waltztime) performs in warmth & humidity Sunbelt Dark Red sp1555 sp. (Formerly Toreador) performs in warmth & humidity sp1521 sp. Duster excellent heat tolerance sp1050 sp. Calypso Graeffin Mariza heat tolerant sp1336 sp. Diana Palmer might revert to old Fiat var. of the 20's and 30's ht0340 xhortorum 'Dixieland' 'Bruni' x 'Fiat' heat resistant ht0334 xhortorum 719 'Siren' 'Bruni' ht0393 xhortorum 86-150-6 'Bruni' ht0236 xhortorum 81-340-2 'Bruni' x 'Yours Truly' sp1417 sp. Unk. Carnation Sport Sport of Yours Truly (continued on next page) Table 3.3
Table 3.3 (continued)
OPGC # Species Cultivar duplicate Parentage resistance ht0347 sp. 'Pascal' Presumed P. xhortorum x P. peltatum pt0800xhortorum x peltatum 'Madeline Crozy' xhortorum x (xhortorum x pt0385peltatum) 84-41-4 'Jean Billes' x 'Madeline Crozy'(#800) rust xhortorum x (xhortorum x 'Honseler's Glorie Lila' x 'Madeline pt0373peltatum) 84-40-2 Crozy'(#800) rust 'Honseler's Glorie Lila' x 'Salmon pt0403xhortorum x peltatum 84-88-1 Queen'(#1498) rust
(xhortorum x pt0387peltatum) x xhortorum 84-53-4 'Pascal' (#347) x 'Honseler's Glorie Rot' rust 67 xhortorum x (xhortorum x pt0418peltatum) (1306) syn. 85-9-2 ('Veronica' x 'Yale' #1508) x 'Veronica' rust ht0399 sp. 85-8-1 'Veronica' x ('Veronica' x 'Yale'(#1508)) rust sp0382 sp. 84-58-3 (P. xfloribunda) x (P. xhortorum) rust sp0335 sp. 85-10-1 rust sp0363 sp. 85-5-2 rust sp0386 sp. 84-59-5 rust sp0401 sp. 84-60-2 rust ht0376 xhortorum 84-82-1 'Mrs. Florence Block' x 84-45-1(#415) rust sp1525 sp. Grenchen ivy/ zonal cross ht0235 xhortorum 'Paris' syn. 80-208-2 'Honselor's Gloria Lila' ht0250 xhortorum 86-93-4 'Jean Billes' x 'Honseler'sGlori e Lila' ht0284 xhortorum 'Jubilee' syn. (175) 83-16-3 'Jean Billes' x 'Honseler'sGlori e Lila' ht0256 xhortorum 86-106-15 'Jean Billes' (continued on next page)
Table 3.3 (continued)
OPGC # Species Cultivar duplicate Parentage resistance sp1440 sp. Snowmass ht0394 xhortorum 'Ben Franklin' Ben Franklin 'Wilhelm Langguth' (#505) x 'Snowmass' (#1440) sp1441 sp. 'Ben Franklin' Ben Franklin 'Wilhelm Langguth' (#505) x 'Snowmass' (#1440) ht0505 sp. 'Wilhelm Langguth' ht0222 xhortorum 79-32-5 'Statbern' (#1546?) x 'Wilhelm Langguth' (#505) ht0223 xhortorum 79-33-6 ('Statbern' (#1546?) x 'Wilhelm Langguth' (#505)) x 'Berlin' ht0234 xhortorum 81-18-2 'Stadtbern' (#1546) ht0398 xhortorum 'Stadtbern' Stadtbern ht0405 xhortorum 78-139-7E GX 505-4-4 (Chris Kramer) ht0803 xhortorum 83-202-1 78-139-7 (#405) x 78-58-2 ht0377 xhortorum 80-500-2 78-139-7 (#405) x 'Honseler's Glorie Rot' ht0241 xhortorum (208) syn. 81-344-3 'Honseler's Glorie Rot' x 'Karminball' syn. 'Glacier Carmen' (80-210-27) ht0238 xhortorum 'Juliette' syn 'Risque' Juliette syn Risque 80-191-3 x 80-207-18 (Honselor's Glorie Rot) 68 ht01098 sp. Risque Juliette syn Risque 80-191-3 x 80-207-18 (Honselor's Glorie Rot) ht0366 xhortorum 'Nittany Lion Red' Florist Mixture from Ferry Morse ht0364 xhortorum 88-51-10 Seedling of 'Nittany Lion Red' (#366) ht0479 xhortorum 98-37-6 Ben Franklin'(#394) x 'Nittany Lion Red'(#366) ht0473 xhortorum 98-17-4 'Ben Franklin'(#394) x 79-32-5(#222) ht0449 xhortorum 93-6-11 'Marilyn' x 'Ben Franklin' (Sport#389) ht0476 xhortorum 98-35-7 85-101-4A(#390) x 'Marilyn' xhortorum x ft0474 frutetorum 98-3-9 85-101-4A(#390) x Pelargonium frutetorum sp1164 sp. Dusty Rose Seedling out of Heidi (#1169) sp0224 sp. 80-164-1 80-164 sp0225 sp. 80-164-2 80-164 sp0504 sp. 'Platinum' Frank Headley sport Frank Headley "Silver leaf" sport ht0495 sp. 'Frank Headley' Frank Headley sport (continued on next pagw)
Table 3.3 (continued)
OPGC # Species Cultivar duplicate Parentage resistance at0559 acetosum br0558 burtonii Progenator of Ivy's cy0258 caylae 'Caylae' cy0417 caylae Steu 2198 ec0557 echinatum ft0455 frutetorum Steu 754 ft0612 frutetorum gt1382 gratum Cinnamon in0264 inquinans 86-22-1 syn. G 630 Pelargonium inquinans seedling in0560 inquinans mo0564 monstrosum pt0410 peltatum 97-28-1 pt0383 peltatum 97-28-3 (P. gravolens x P. echinatum) 'Jill' 69 pt0411 peltatum 97-30-3 sp1498 peltatum Salmon Queen sp1508 peltatum Yale zn0263 zonale 86-24-1 G 632 P. zonale Seedling from Self-pollination zn0799 zonale 95-6-1
Figure 3.1: Diagrams of relationships between Ciconum selections based on records
from OPGC’s database. Italicized names correspond to records that only identify the
species of an accession’s parent. Names are used for cultivars that are listed in breeding
histories, but which are not included in OPGC’s collection. “X” designates an unknown parent from cases where the parent was identified only with a breeding accession number.
70
x Veronica 1508 x 1440 505 1546 398
399 418 234 1441 366 394 222 x Berlin x frutetorum x Marilyn x 364 479 473 223
1328 474 476 449 hortorum peltatum
435 612 Karminbal Honselor’s 241 Glorie Rot 347 x 405 387 803 377 238 1098 Fiat Bruni Your’s Truely
1336 340 334 393 236 1417
Jean 1498 Biles Mrs. Honselor’s Florence x 800 403 Glorie Lila 250 284 256 385 Block 376
235 373 Figure 3.1
71 Results
TRAP analysis was completed on 775 Pelargonium accessions at OPGC, and 301
markers were recorded from four primer pairs in two multiplexed reactions. The
neighbor joining analysis generated by PAUP revealed a division of the collection into
two large groups: section Ciconum and non-Ciconum accessions. The largest group
(more than 650 accessions) included all of the accessions from section Ciconum with
many accessions that had no pedigree. There was much more diversity in the second
group, which included the accessions from every other taxonomic section represented at
OPGC. In order to observe more details regarding the latter group, a separate Bayesian
analysis was performed. Likewise, several separate Bayesian analyses were completed
on the section Ciconum group in order to test the suspected phylogenetic relationships
between the accessions. The results of TRAP analyses in both groups are presented
separately below.
Phylogenetic analysis of non-Ciconum accessions
Figure 3.2 shows a subdivision of the neighbor joining tree generated for the entire Pelargonium collection. None of the accessions on this part of the tree belong to section Ciconum except for the two P. caylae accessions (cy0417 and cy0258), and possibly the unknown accession that clustered with those two (sp0452). This separation of P. caylae from the rest of its section is consistent with other molecular analyses (James
et al. 2004; Bakker et al. 2000 and 2004). Clusters of accessions that match known
background information are indicated by brackets to the right of the tree.
72 First the P. × domesticum cluster includes one pair of potentially duplicated
accessions (dm0794 and dm0795 in a box) which was identified as a perfect match in
Figure 3.2. Three other pairs were within the P. × domesticum cluster in Figure 3.2
(indicated by solid, or doubled, grey curves), but they were also equally clustered with
other P. × domesticum accessions. In contrast, three pairs of potential duplicates (the first of each pair is in the non-Ciconum group: dm0789 and sp1268; dm0790 and sp1101; sp0588 and sp1362) were divided between the non-Ciconum group and section Ciconum.
Accessions with shared parentage in the P. × domesticum cluster are indicated by dotted lines.
The sister branch to the P. × domesticum cluster shown in Fig. 3.2 does not include any accessions with a species designation, but one accession (sp0577) was included as a potential duplicate based on its description as a white form of accession sp0579. Neither this pair nor the two other potentially duplicate pairs based on similar forms that were included in Figure 3.2 (solid grey curves), were found to match perfectly based on neighbor joining analysis of the TRAP data. However, one of those three pairs
(sp0587 and hy0591) clustered within one of two large section Pelargonium clusters.
Another sister branch to the P. × domesticum cluster shown in Fig. 3.2 includes three accessions of one section Pelargonium species, P. crispum, clustered in the same branch as three section Campylia hybrids. On a separate branch section Pelargonium is divided into four clusters. Two have many accessions with species designation in section
Pelargonium, and the other two clusters only have a few accessions with known species designations. A small cluster of just three accessions (sp0585, hy0584, sp0579) was a sister cluster to the cluster combining all the section Pelargonium and P. × domesticum
73 accessions. One last large cluster included two branches of section Reniformia. One of the section Reniformia branches included all of the P. × fragrans accessions, and several accessions without a species designation. The other branch included three other section
Reniformia species, and the only accessions from sections Isopetalum and Jenkinsonia.
Although the neighbor joining tree (Fig. 3.2) included clusters that were consistent with taxonomic sections, there were some exceptions that required further examination. First the accessions of P. caylae (a member of section Ciconum) clustered with two section Otidia accessions (ds0551 and fe0563), one section Campylia accession
(ga0561), and one section Cortusina accession (ec0557). Also, most of the other section
Campylia accessions were in the same cluster as P. crispum accessions. Finally, several clusters had too many accessions without pedigrees or species designations to be assigned to a taxonomic section.
Bayesian analysis
In order to define the clusters more clearly, Bayesian analysis was applied to a subset of Pelargonium accessions including of all the accessions from Figure 3.2 and several representative species from section Ciconum (for comparison to P. caylae).
Bayesian analysis compares multiple runs that start from different starting trees, and saves sample trees as the runs progressively find more likely trees. Statistical evaluations performed during the analysis to identify when the multiple runs have converged into similar trees, and help to determine the number of sample trees that should be excluded from the consensus tree. The purpose of excluding (or burning-in) trees from the beginning of the run is to eliminate trees formed before the runs converge.
74 The average standard deviation of split frequencies (AvgSD) was approximately
0.027 after 19 million generations with a 50 percent burn-in, and it remained at approximately that value until the run ended at 20 million generations. Therefore the summary used a burn-in equivalent to 50 percent of the trees saved after 19 million generations to produce a consensus tree from 95,000 trees. The summary of parameters with that burn-in showed that the log probability had stabilized. The parameters show that the three runs of this analysis seem to have converged, and the consensus tree with the same burn-in rate is shown in Figure 3.3. Most of the accessions in Figure 3.3 were divided between clusters of accessions related to section Pelargonium, accessions in section Reniformia, and accessions in section Ciconum.
Pelargonium
The first main group on the tree (labeled “Pelargonium” in Figure 3.3) contained all of the accessions related to section Pelargonium. This cluster was further divided into clusters containing related species: a large cluster of multiple species including P. graveolens, a cluster of P. × domesticum, a cluster combining P. crispum and section
Campylia hybrids, a small cluster including P. tomentosum, and a single branch with the only section Glaucophyllum species.
The first cluster in section Pelargonium could be divided into four main subclusters (numbered 1-4 in Fig. 3.3) as well as two separate branches pairing individual accessions and several independent branches. This cluster also includes the three potential duplicates (marked by solid grey curves) that failed to match in section
Pelargonium of Figure 3.2, and these potential duplicates also did not match in Figure
3.3. One of the separate branches in this cluster paired only two non-pedigreed
75 accessions (sp0589 and sp0614), and so it was uninformative regarding those two
accessions. A second branch (enclosed in the box for section Hoarea) included the only
representative of section Hoarea (rp0580), with a non-pedigreed accession (sp0588).
The independent branches included three non-pedigreed accessions (sp0761, sp0596 and sp0577), a hybrid of P. × graveolens and P. × echinatum (hy0547), and a single example of P. terebinthinaceum (tb0581). P. terebinthinaceum is a species which has since been reclassified as a part of P. graveolens (GRIN www.ars-grin.gov).
Aside from P. terebinthinaceum and the hybrid previously described, all of the accessions with a known relationship to P. graveolens (gr0524, gr0533, gr0516, gr0519, and gr0530) clustered in subcluster number one (Fig. 3.3). Other members of this sub cluster (Fig. 3.9) included the two accessions of P. cucullatum (ad0534 and ac0517), one of the two accessions of P. denticulatum (dt0514), P. capitatum (cp0538), and P. × melissimum (ml0522), which is listed on GRIN (the USDA-ARS Germplasm Resource
Information Network: www.ars-grin.gov)as both a member of Campylia and originating as a hybrid of P. graveolens and P. crispum. Four of the accessions with no background
information were also included as part of this subcluster.
All of the next subdivisions contained more nonpedigreed accessions than
accessions with known species designations. Subdivision 2 of this cluster also included
P. scabrum (sc0602), one of the two P. quercifolium accessions (qu0507) and a hybrid of
P. denticulatum and P. × domesticum (hy0584). The other P. quercifolium accession
(qu507) was in subdivision 3 with the other accession of P. denticulatum (fc0598) and the only accession of P. glutinosum (gl0597). The last subdivision (4) included P. radens
(rd0599) as well as a hybrid of P. tomentosum and P. graveolens (hy0591).
76 The second cluster in section Pelargonium included all the accessions of the
hybrid species, P. × domesticum, which putatively includes species from section
Pelargonium in its background. As already described, the sole representative of the other
major contributing section listed on GRIN, Glaucophyllum (fu0565), was on a separate
branch set within the section Pelargonium cluster. Three of the four pairs of potentially
duplicated accessions that clustered with this hybrid species in Figure 3.2 clustered as a
perfect match in Figure 3.3 (solid grey curves). Along with the identified P. ×
domesticum accessions (labeled by that species name in Figure 3.3), a sister cluster
included seven accessions without identified species, but identified as regal geraniums by
their breeder (Fig.3.6).
The last main cluster in section Pelargonium combined two section Campylia
species with one section Pelargonium species. The section Campylia species at the base
of the cluster was P. × nervosum (tr0520 and nr0528), which was listed on GRIN as a
hybrid of unknown origin (Fig. 3.8). The section Pelargonium species, P. crispum
(cr0536, cr0523 and cr0531), was the next group in this clade, followed by P. ×
scarboroviae (sr0525) from section Campylia. Four non-pedigreed accessions were also included at the end of this clade. Also included in the group related to section
Pelargonium were two branches with only one or two accessions (Fig. 3.7): P. tomentosum (tm0521), a non-pedigreed accession (sp1384), and the only accession from
section Glaucophyllum (fu0565).
77 Reniformia
The cluster labeled “Reniformia” in Figure 3.3 included all accessions from section Reniformia along with some accessions with no previous species information.
The first division of that cluster separated the known members of P. × fragrans from other species in this section: P. sidiodes (sd0549), P. reniformia (re0552) and
P.ionidiflorum (io0553). The same two groups were produced by the neighbor-joining tree (Fig. 3.2) with high jackknife values (over 85%). All accessions in the P. × fragrans group had soft heart shaped leaves (Fig. 3.4), while the other species were slightly different. Three of the four other accessions had leaves very similar to the P. × fragrans group, but a lighter color than most of the P. × fragrans group. The morphology of the
P. sidiodes accession (sd0549) differed from that of the rest of the section (Fig. 3.5).
78
Figure 3.2: Non-Ciconum section of the Pelargonium dendrogram according to the neighbor joining tree. Jackknife values are based on 1000 replicates. The entire tree is
too large to show on one page, so only the non-Ciconum group is shown here. The only
exceptions are the two accessions of P. caylae and possibly the non-pedigreed accession
in the same cluster. Potential duplicates are connected by solid (or doubled) lines;
parents by dotted lines. Accessions are labeled with IDs that match those in Table 3.1.
79
P. × domesticum
P. crispum
Section Campylia sp1384
Section Pelargonium
sp1383 sp1382 sp1385
Section Pelargonium
P. caylae
Section Reniformia
Section Section Reniformia Ciconum Figure 3.2 80
Figure 3.3: Dendrogram from Bayesian analysis of all Pelargonium accessions from
outside section Ciconum and a few representatives of that section. Solid (or doubled) lines connect potential duplicates, and dotted lines connect parents to offspring. Sections
are labeled with arrows, boxes, or text at the base of a branch, and accessions are labeled
with IDs that match those in Table 3.1.
81
Figure 3.3 1 Campylia
2 P. denticulatum × P × domesticum 3
sp1385 4 sp1383 sp1382 Hoarea P. × graveolens × P. × echinatum
Pelargonium (and related groups)
P. × domesticum
P. crispum P. × nervosum Campylia sp1384 Glaucophyllum
Reniformia P. × fragrans
sp1323 .
Ciconum Ciconum – P. caylae Campylia Isopetalum Otidia Cortusina Jenkinsonia
82
Figure 3.4: P. fragrans accessions and three accessions with no pedigree that all clustered together in Figure 3.3.
83
Figure 3.5: Four accessions of section Reniformia that were separated from the P. fragrans accessions in Figure 3.3. Top left: P. ionidiflorum, top right: P. reniforme, bottom left: P. sidoides, bottom right: sp0554
84
Figure 3.6: Two groups of P. × domesticum accessions. Accessions from Dr. Craig are on the left, while additional accessions without previously identified species are on the right. The leaves of the new accessions are consistent with their being part of P. × domesticum.
85
Figure 3.7: The three accessions in section Pelargonium that are not part of a larger cluster in Figure 3.3. P. tomentosum (tm0521) is on the left with small leaves, and the accession on the bottom right is sp1384 which was the only accession in the cluster with P. tomentosum. The only section Glaucophyllum accession, P. fruiticosum (fu0565), is on the top right.
86
Figure 3.8: The differences in leaf shape and plant form found in P. crispum and P. × nervosum accessions. The top two pictures show P. crispum, and the bottom two show P. × nervosum.
87
Figure 3.9: Variation in appearance among accessions in section Pelargonium. The top image shows accessions from a branch within the larger P. graveolens cluster (cluster 1 of figure 3.3) in which all five accessions had similar leaves in spite of coming from different species (one is from P. graveolens and two are from P. cucullatum). The bottom picture shows examples of the leaf shapes scattered throughout section Pelargonium.
88
Figure 3.9
89 Phylogenetic analysis of section Ciconum
Although P. caylae is a member of section Ciconum, on the neighbor joining tree in Figure 3.2 this species was in a cluster separate from all the other species of that section. The Bayesian analysis of the same data excluded the accessions from section
Ciconum that lacked species designations, and resulted in a tree that included two separate clusters for section Ciconum (Fig. 3.3). One cluster included the two P. caylae accessions (cy0417 and cy0258) as well as one accession without a species designation
(sp0452). In Figure 3.3 the cluster with P. caylae was equally separated from the rest of section Ciconum as the other sections at the bottom of the tree.
Figure 3.3 included only a few representative accessions from section Ciconum, but the species with more than one representative still clustered together. The two P. zonale accessions (zn0263 and zn0799) appeared to cluster together, as did the two P. stenopetalum accessions (br0558 and st0566). The two hybrid species (kw0567 and mo0564) combining P. inquinans and P. zonale are in between those two accessions on the tree. The P. inquinans (in0264 and in0560) and P. frutetorum (ft0612 and ft0455) accessions all clustered together along with one non-pedigreed accession. Because there were more than 650 closely related accessions assigned to section Ciconum by the neighbor joining analysis, two additional studies of that section were completed: 1) to determine whether potentially duplicated accessions could be confirmed as redundant germplasm and 2) to compare accessions with known pedigree using the TRAP results produced by the data from all the other studies and results produced using new primers based on resistance genes
90 Potential duplicates
Bayesian analysis was completed with 85 section Ciconum accessions
representing 26 sets of potentially duplicated accessions (Table 3.2) and some
representative accessions with known species designations. The consensus tree produced
at the end of the analysis was generated using the same diagnostics as the analysis of
“non-Ciconum” accessions. Convergence was demonstrated with an average standard
deviation of split frequencies of 0.0133, and a random pattern on the log probability
graph with a 50% burn-in rate. Approximately half of the potential duplicate sets in
section Ciconum included at least two accessions that could be confirmed as duplicates
(Fig. 3.10).
One example of a potential duplicate that did not match in Figure 3.10 was the pair that included zn0799 and sp0262. Although those two accessions did not cluster, zn0799 was one of two accessions (zn0263 was the other) that had a species designation of P. zonale. Unlike the potential duplicate pair, there was a cluster containing the species pair for P. zonale. Another potential duplicate set that did not match in Figure
3.10 included three accessions identified as ‘Ben Franklin’ (ht0389, ht0394 and sp1441), as well as a potentially duplicated pair that was a hybrid of ‘Ben Franklin’ (fl0491 and ht0449), and both parents of ‘Ben Franklin’ (sp1440 and ht0505). ‘Wilhelm Langguth’
(ht0505) was the only cultivar of that group that was not included in one of the two clusters containing the remainder of these accessions. The population of Pelargonium
accessions at OPGC contains many hybrid relationships such as the ‘Ben Franklin’
complex, which may reduce the resolution of the groups produced.
91 The potential duplicates, in addition to the two accessions from P. zonale, which matched in Figure 3.10 included: sp1487 and sp1505; sp1493 and sp1498; sp1454, sp1462 and sp1461; sp1457 and sp 1479; sp1473, sp1532 and sp1529; ht0240, ht0241, sp0365 and sp0351; ht0284 and ht0341; sp0605 and sp0604; sp0608, sp0609 and sp0607; sp1181 and sp1177; sp1070 and sp1080; and sp0495 and sp504. Of those 12 duplicate groups the following six clusters included the entire set of potential duplicates. First,
‘Memories’ (sp1487) and ‘Sybil Holmes’ (sp1505) form a cluster (#1 in Fig. 3.10), which is consistent with the identification of ‘Memories’ as a sport of ‘Sybil Holmes.’ Next, cluster 2 in Figure 3.10 includes ‘Roller’s Pathfinder’ (sp1493) and three duplicated data entries for ‘Salmon Queen’ (sp1498). This is also consistent with the description of
‘Roller’s Pathfinder’ as a mesh veined form of ‘Salmon Queen.’ The third cluster includes ‘Queen of Hearts’ (sp1529) and a light pink version, ‘Ten of Hearts’ (sp1532).
The same cluster also contained ‘Stitches’ (sp1473), a sport of ‘Ten of Hearts.’ The fourth cluster included four accessions (ht0240, ht0241, sp0365 and sp0351) that might be considered synonymous with ‘Glacier Carmen’ (not at OPGC) based on their parentage. Cluster 5 included ‘Orion’ (sp1177) and the cultivar putatively believed to be its sport, ‘Pink Ice’ (sp1181). The final cluster that included an entire set of potential duplicates was ‘Platinum’ (sp0504) and the cultivar of which sports it, ‘Frank Headley’
(ht0495).
Accessions with known pedigree
The 88 accessions in Table 3.3 were selected for a final evaluation of section
Ciconum, because of known pedigrees, species designations, or resistance to insects and diseases. The first TRAP analysis of these accessions was a Bayesian analysis of the
92 same data already presented with regards to other subsets of accessions (Fig. 3.11).
Consensus of this analysis was demonstrated by an average standard deviation of split frequencies of 0.0197, and a random pattern on the log probability graph with a 50% burn-in rate. The same set of accessions was also analyzed using targeted TRAP primers based on resistance genes, and 334 new markers were scored (Fig. 3.12). Convergence of the Bayesian analysis of the new dataset was demonstrated by an average standard deviation of split frequencies that remained at approximately 0.0321 for the last 4 million generations of the analysis, and a random pattern on the log probability graph.
Comparisons of the dendrograms produced by these two sets of data shows that, in some cases, the markers from primers that targeted resistance genes were more successful at clustering related cultivars.
The dendrogram from the first TRAP analysis of this subset, using the data from the previous analysis of the entire collection, is in Figure 3.11. The species in Figure
3.11, include the two P. caylae accessions (cy0417 and cy0258) again clustered together and separate from the rest of section Ciconum. Also, the two P. stenopetalum accessions
(br0558 and at0566) clustered together. One of those two, br0558, was listed in OPGC’s database as a progenitor of the ivy geraniums, but in this case it has clustered more according to its species than the hybrids it may be related to. Although there were two accessions included for each of P. zonale, P. inquinans and P. frutetorum, the only other species that was in a single cluster of this tree was P. peltatum (ivy geraniums – sp1498, sp1508, pt0383, pt0410 and pt0411). Most of the hybrids of ivy geraniums were not included in the cluster for that species in Figure 3.11. Also included on this tree were three pairs of potential duplicates, but only one of those pairs (ht0495 and sp0504)
93 clustered together. They also clustered on the potential duplicates tree. One group
includes sp0224 and sp0225, which are two seedlings from the same cross. The two
arthropod resistant accessions (ht0350 and ht0357) are also indicated in Figure 3.11, but
overall the relationships between cultivars in this section are not very well resolved.
In many cases, analysis of the same accessions using primers that target resistance
genes (Fig. 3.12) produced better resolution of the relationships within section Ciconum.
P. frutetorum was the only species in Figure 3.12 that did not cluster according to species designation. However, one of the two P. frutetorum accessions, ft0612, did cluster close to a hybrid of P. frutetorum (ft0474). In contrast to the dendrogram using the general data from Chapter 3, both P. zonale accessions (zn0799 and zn0263) clustered together.
The five P. peltatum accessions (pt410, pt0411, pt0383, sp1498 and sp1508) remained clustered together, but they were joined by many of the hybrids of those ivy geraniums into a larger related cluster (“Ivy” cluster in Fig. 3.12). The hybrids included ‘Pascal’
(ht0347 a presumed hybrid of P. × hortorum and P. peltatum), as well as ‘Madeline
Crozy’ (pt0800), and some unnamed hybrids. Of the eight hybrids of P. × hortorum and
P. peltatum, only pt0385, pt0373 and pt0418 are not included in the “ivy cluster” of
Figure 3.12. Also ‘Grechen’ (sp1525) is a hybrid of ivy and zonal cultivars and is included in the “ivy cluster.” The only accessions in the “ivy cluster” that are not already known (or suspected) as descendants of P. peltatum are: sp0386, sp0382, and sp0401.
All three of those accessions carry rust resistance, which was also identified in many hybrids of P. × hortorum and P. peltatum. One more species, P. inquinans (in0264 and
94 in0560) clustered into a larger branch that included several P. × hortorum accessions as
well. Also, of the four pairs of potential duplicates included in this tree, three pairs
(ht0495 and sp0504, ht0394 and sp1441; and ht0398 and sp1546) clustered as expected.
Discussion
The main finding of this TRAP analysis of OPGC’s entire Pelargonium collection was that most of the diversity in OPGC’s Pelargonium collection exists in sections other than section Ciconum. Within section Ciconum the accessions include redundant
accessions, and so most of the purge for this collection should come from section
Ciconum. This finding was based on dendrograms from both neighbor joining (Fig. 3.1)
and Bayesian (Fig. 3.10) analyses, and supported by further studies of subsets of
accessions from section Ciconum (Fig. 3.11 and Fig. 3.12).
The most notable exception to the clustering of species according to their
taxonomic section in Figures 3.2 and 3.3, is from section the Campylia. The Campylia accessions in this collection include hybrid species that in some cases are listed on GRIN both as hybrid species belonging in section Campylia and also as originating from hybrids of species in other sections (Table 3.1). As a result, OPGC’s accessions from section Campylia were scattered across the tree in Figure 3.3. Only one (ga0561) of the
Campylia accessions was separated from other accessions at the base of the tree. That accession was also listed as a hybrid of two species from section Polyactum, and so perhaps it should be the sole example of section Polyactum in this tree (Table 3.1).
Another Campylia accession, P. × melissimum (ml0522), clustered in subcluster 1 of the
95 first group of section Pelargonium, and was also identified as originating from a hybrid
of P. crispum and P. graveolens (Table 3.1). Three other Campylia hybrids were of
unknown origin, but were included in the same cluster as P. crispum (cr0536, cr0531 and
cr0523) from section Pelargonium. This clustering pattern suggested that P. crispum or a
related species might have been part of their hybrid origin. The two P. × nervosum accessions at the base of this cluster (tr0520 and nr0528) and the four unknown accessions at the end of the cluster (sp0515, sp0545, sp0541, sp0540) have the maple- leaved shape typical throughout section Pelargonium (Fig. 3.8). P. scarborvae (sr0525), one of the hybrids described as part of section Campylia on GRIN, grouped most closely with the P. crispum accessions and has a growth habit similar to that species, suggesting that this Campylia species is related to P. crispum. Also, clustering section Campylia as a subdivision within section Pelargonium is consistent a previous molecular phylogeny
(Bakker et al., 2004).
The other accessions in the non-Ciconum group were clustered according to their taxonomic section whenever there were multiple accessions available for the section.
The TRAP analyses divided the collection among three main sections with a few examples of other sections. First in section Pelargonium, 27 accessions with no previous species designation were added to the largest clade of the section. P. × domesticum, the second largest clade in section Pelargonium included a branch with seven more accessions with no previous species designation. OPGC’s collection of P. × domesticum accessions could be balanced by reducing to no more than 10 accessions, the very tightly clustered P. × domesticum group next to those previously unknown accessions (Fig. 3.6).
96 At the very least this group may be reduced by removing half of each pair of potential
duplicates that paired in this section. Three pairs clustered together, but the fourth
potential duplicate (dm0773 and dm0 774) was not quite as perfectly paired as the other
three.
Four additional accessions with no previous species designation shared a cluster
with P. caylae and section Campylia. In total 39 accessions were added to the section
Pelargonium group, and another seven were added to section Reniformia. Those 46 accessions along with the 70 accessions that were already identified as belonging to a section other than Ciconum, define a diverse group of OPGC’s Pelargonium accessions.
Elimination of about 15 redundant accessions of P. × domesticum results in a population of 103 accessions that are not in section Ciconum.
In order to maintain the most diversity in the collection, section Ciconum accessions need to be included. Since over 650 accessions were left in section Ciconum, the purge of that section will be very severe. Three accessions that should be retained were on a separate cluster (Fig. 3.2 and Fig 3.3) containing only P. caylae from section
Ciconum. This was not surprising, as P. caylae has been separated from the rest of section Ciconum in other molecular phylogenies (James et al. 2004; Bakker et al. 2000 and 2004), which have attributed this pattern to its physical differences and separate evolution on the island of Madagascar.
In order to evaluate the potential redundancy among the remaining accessions in section Ciconum, two additional studies were completed. First, a study of potential duplicates, and second, a study of related accessions with primers targeting resistance
97 genes. The first study confirmed that approximately half of the potentially duplicated
accessions were actually duplicated, and purging those duplicates can be the first step to
eliminating redundancy in the collection. In the second study there were more matches to
cultivar information in the clustering patterns produced with primers targeting resistance
genes, than with the more general primers used in the rest of the analyses.
Our new data used target primers based on Pelargonium sequences, and
specifically targeting the ∆9 14:0-acyl carrier protein fatty acid desaturase gene that
Schultz et al (1996) identified as being necessary for arthropod resistance. Schultz et al.
(1996) identified this gene as the single, dominant gene responsible for making the tricome exudates sticky and therefore toxic to insects. The product of this gene has been shown to have a broad range of insecticidal properties (Schultz et al., 2006), which makes it important for insect resistance. Insect resistance in general, is important to
Pelargoniums because of the many viruses that can be transmitted by insects, such as thrips (Chen and Williams, 2006). This point is especially important at OPGC where extra care is taken to avoid the spread of both pathogenic and ornamental viruses within the vegetatively propagated collection. Unfortunately, the two accessions with known resistance based on this gene, did not correspond with markers that could be used to screen the rest of the collection.
Conclusion
Based on this study the recommendation to OPGC is that they can maintain most of the Pelargonium collection’s diversity by keeping most of the accessions shown in
Figure 3.3, and then allowing selections from section Ciconum to fill up any remaining space. Additionally, they can reduce the number of accessions identified as P. ×
98 domesticum somewhat based on duplicated accessions, and the extra diversity shown by
the previously unknown accessions added to that group. Ultimately, reaching OPGC’s
goal of keeping as much diversity as possible in a population that is only 25% of its
initial size will require OPGC to reduce accessions from section Ciconum to about ten
percent of their holdings, because that section was initially so over represented in the
collection. (653 accessions on the neighbor joining tree.) The remaining accessions will
represent the most diversity possible, because section Ciconum includes so much redundant germplasm.
99
Figure 3.10: Dendrogram containing 26 sets of potentially duplicated accessions.
Approximately half the sets included at least two accessions clustering together. Boxes and solid (or doubled) lines show potential duplicates; dotted lines indicate parents
100
1
2
3 sp1342
‘Ben Franklin’
4
P. zonale
sp1323 sp1343
5
6
Figure 3.10
101
Figure 3.11: Bayesian analysis of Ciconum accessions selected for their resistance and breeding history. Arrows connect parents to children, grey lines connect siblings, orange indicates potential duplicates, and other colors indicate duplicated species
102
Arthropod resistant
P. peltatum
sp1324 sp1316 sp1325 sp1323 sp1306
sp1397 sp1370
sp1273
P. stenopetalum
P. caylae
Figure 3.11
103
Figure 3.12: TRAP analysis (targeting resistance genes) of Ciconum accessions selected for their resistance and breeding history. The dendrogram was produced by
Bayesian analysis of data with the new resistance gene based primers. Arrows connect parents to children, grey lines connect siblings, orange indicates potential duplicates, and other colors indicate species as indicated.
104 P. inquinans
“Ivy geraniums”
P. zonale P. caylae
Figure 3.12
105 CHAPTER 4
DISCUSSION AND GENERAL SUMMARY
CHAPTER 4: DISCUSSION AND GENERAL SUMMARY
Molecular markers can provide an efficient means of screening germplasm collections and have been used in many species. Molecular markers can be used to evaluate how well a germplasm collection represents a wild population, such as Borderea chouardii (Segarra-Moragues et al., 2005), artichokes(Sergio and Gianni, 2005) or peppers (Guzmán et al., 2005). Segarra-Moragues et al. (2005) determined that a sub population of B. chouardii needed more representation in a germplasm collection based on SSR markers. Sergio and Gianni (2005) used molecular markers to identify a ‘core’ population within a germplasm collection of artichokes, which could be important for future breeding. Guzmán et al. (2005) used AFLP markers to evaluate the diversity of pepper germplasm maintained in Guatemalan home gardens. When Sergio and Gianni
(2005) evaluated a germplasm collection of cardoons, they observed that characterizations of germplasm collections are enhanced by separate studies to divide main groups and to resolve relationships within those groups. Ornamental plants have not been the subject of as many genomic studies as crop plants, but molecular markers have the potential to improve cultivar identification and pedigree analysis of economically importance (Arús, 2000).
106 The Ornamental Plant Germplasm Center’s (OPGC) collection of approximately
800 Pelargonium accessions needed to be condensed to a more manageable core collection of approximately 200 accessions. This required the use of molecular markers, because many of the individual accessions were donated without much background information. Screening for insect and disease resistance in this collection was also important, because both pests can spread easily through the OPGC greenhouses. TRAP
(Target Region Amplification Polymorphism) analyses have the potential to provide markers that are linked to a gene of interest by using a targeted primer based on existing
DNA sequences. In this study, TRAP analyses of OPGC’s Pelargonium collection were completed to test the utility of TRAP markers in categorizing accessions from this genus
(Chapter 2), to categorize the non-pedigreed accessions into taxonomic sections (Chapter
3), and to observe relationships within section Ciconum (Chapter 3).
The purpose of Chapter 2 was to ensure that TRAP markers could be used to evaluate Pelargonium accessions. In order to complete the test of TRAP a DNA extraction technique was developed to compensate for the high levels of phenolic compounds present in some Pelargonium species. The Pelargonium DNA extraction technique was a modified form of existing DNA extraction techniques (Tel-zur et al.,
1999; Saghai-Maroof et al., 1984), and combined several washes to remove the phenolics with the addition of high levels of antiphenolic compounds. The DNA isolated from 46
Pelargonium accessions was used to evaluate the TRAP technique. Two sets of TRAP primers were sufficient to discriminate most of the accessions from each other. From one primer set 148 polymorphic markers were recorded, and there were 160 polymorphic markers recorded for the second primer set. The results from the two primer sets were
107 evaluated separately, and both data sets yielded similar clusters composed of accessions clustered by species or shared parentage. The similarities within clusters demonstrated that TRAP effectively characterized the accessions included in this evaluation.
The first TRAP analysis in Chapter 3 was applied to the entire collection, and resulted in 301 TRAP markers from four pairs of primers. Neighbor joining analysis of the entire dataset was used to separate the collection into two groups: section Ciconum and non-Ciconum accessions. Each group was analyzed further by Bayesian analysis.
The section Ciconum cluster on the neighbor joining analysis of the entire collection included over 650 accessions. Those 650 accessions included all accessions representing species from section Ciconum except for the two accessions from P. caylae, which were in the non-Ciconum group. All accessions in the non-Ciconum group were analyzed along with a few representative accessions from section Ciconum. The two P. caylae accessions and one non-pedigreed accession were in a sister cluster to the other representatives of section Ciconum. Most of the accessions that are not in section
Ciconum are in sections Pelargonium and Reniformia. The other sections are all represented by just a few individuals. TRAP analysis helped identify seven non- pedigreed accessions as members of section Reniformia, and assigned the 38 remaining non-pedigreed accessions in the non-Ciconum group to section Pelargonium.
Accessions from section Ciconum were included in two subsets, and Bayesian analysis of the same 301 markers was applied to those subsets. The first subset included
26 sets of potential duplicates, and the analysis confirmed that approximately half the sets contained at least one pair of duplicates. The second subset included accessions with known pedigrees or known resistance or susceptibility traits. In addition to analysis of
108 the 301 primers from the previous data set, a new data set was produced using new primers targeting resistance genes. The new primers were designed based on sequences for arthropod resistance (∆9 14:0-acyl carrier protein fatty acid desaturase gene: Schultz et al., 1996), rust resistance in Arabidopsis (TIR-NBS-LRR: Borhan et al., 2004) and a conserved region of an ethylene receptor homolog in P. × hortorum (Dervinis et al.,
2000). There were 334 markers scored from the four primer sets (eight pairs) based on the resistance genes. Bayesian analyses of the TRAP data from both the new and old primers showed that the primers targeting resistance genes yielded some clusters that matched the pedigrees better than the data from the first primers sets. Unfortunately, no
TRAP markers were identified linked to resistance, and so there is not enough information to screen the collection for resistance using these primers.
Using TRAP markers, we have extensively analyzed the Pelargonium germplasm collection at OPGC, and identified a core collection of Pelargonium accessions. There were 118 accessions identified as belonging to a section other than Ciconum; 46 accessions were identified from previously non-pedigreed accessions. OPGC can easily eliminate about 15 of the least diverse cluster of P. × domesticum, leaving the collection at a size of 103 accessions before they consider section Ciconum. Since that section includes most of the population, over 650 accessions, most of the accessions purged from the collection should be from that section. In all of the TRAP analyses, section Ciconum has two branches. The first is a smaller branch that includes just three accessions: two members of P. caylae and one unidentified accessions that matches the appearance of those two accessions. Accessions from that branch should probably be kept at OPGC since it is so significantly separated from the rest of section Ciconum, and has so few
109 representatives. Additionally, OPGC should consider keeping all 15 of the members of section Ciconum with known species designations other than P. × hortorum. That would assign 121 accessions to the collection leaving space for 79 additional representatives of section Ciconum. This section will have the highest representation, but it is also the section most frequently used for ornamental purposes.
110 LIST OF REFERENCES
LIST OF REFERENCES
Albers, F. and J.J.A. Van der Walt. 1992. Karyological and chemotaxonomical studies of the hybrid Pelargonium incarnatum x P. patulum (Geraniaceae) and its parents. Plant Systematics and Evolution 183: 161-167.
Albers, F., and J.J.A. Van der Walt, M. Gibby, D.E. Marschewski, R.A. Price and G. Du Preez. 1995. A biosystematic study of Pelargonium section Ligularia: 2. Reappraisal of section Chorisma. South African Journal of Botany 61(6):339- 346.
Albers, Focke, Mary Gibby and Mathilde Austman. 1992. A reappraisal of sect. Ligularia (Geraniaceae). Plant Systematics and Evolution 179: 257-276.
Altekar, G., S. Dwarkadas, J. P. Huelsenbeck, and F. Ronquist. 2004. Parallel Metropolis-coupled Markov chain Monte Carlo for Bayesian phylogenetic inference. Bioinformatics 20:407-415.
Alwala, S., A. Suman, J.A. Arro, J.C. Veremis, and C.A. Kimberg. 2006. Target region amplification polymorphism (TRAP) for assessing genetic diversity in sugarcane germplasm collections. Crop Sci. 46: 448–455.
Arús, P. 2000. Molecular Markers for ornamental breeding. Acta Horticulturae 508: 91-98.
Bakker, Freek T., Alastair Culham, Clive E. Pankhurst and Mary Gibby. 2000. Mitochondrial and Chloroplast DNA-based phylogeny of Pelargonium (Geraniaceae). American Journal of Botany 87(5): 727-734.
Becker, J. and F. Brawner. 1996. Scented geraniums: knowing, growing and enjoying scented Pelargoniums. Interweave Press, Loveland, CO.
Bentvelsen, Gé C.M. 1992. A new generation of ivies. Third International Geranium Conference Proceedings: 233-234.
111 Borhan, M.H., E.B. Holub, J.L. Beynon, K. Rozwadowski, and S.R. Rimmer. 2004. The arabidopsis TIR-NB-LRR gene RAC1 confers resistance to Albugo candida (white rust) and is dependent on EDS1 but not PAD4. Mol. Plant Microbe Interact. 17(7): 711-719.
Buntjer, J.B. 1999. Crosschecker, Ver. 2.91 Wageningen University and Research Center, Wageningen, The Netherlands.
Chen, J., J. Hu, B.A. Vick, and C.C. January. 2006. Molecular mapping of a nuclear male-sterility gene in sunflower (Helianthus annuus L.) using TRAP and SSR markers. Theor. Appl. Genet. 113: 122–127.
Chen, Yan and Kimberly A. Williams. 2006. Quantifying western flower thrips (Frankliniella occidentalis Pergande) (Thysanoptera: Thripidae) damage on ivy geranium (Pelargonium peltatum (L.) L’Her ex Ait.)(Geraniaceae Juss.) with Adobe® PhotoshopTM and Scion® image software. Journal of the Kansas Entomological Society 79(1): 83-87.
Coetzee, N. (née van Wyk), J.J.A. van der Walt and E.M. Marais. 1994. The identity of a natual hybrid of Pelargonium (Geraniaceae). South African Journal of Botany 60(1): 1-4.
Deng, Zhanao, Fahrettin Goktepe, Brent K. Harbaugh and Jinguo Hu. 2007. Assessment of genetic diversity and relationships among Caladium cultivars and species using molecular markers. Journal of the American Society for Horticultural Science 132(2): 219-229.
Dervinis, C., D.G. Clark, J.E. Barrett, and T.A. Nell. 2000. Effect of pollination and exogenous ethylene on accumulation of ETR1 homologue transcripts during flower petal abscission in geranium (Pelargonium x hortorum L.H. Bailey). Plant Mol. Biol. 42 (6): 847-856.
Gibby, M., F. Alders and B. Prinsloo. 1990. Karyological studies in Pelargonium sect. Ciconium, Dibrachya, and Jenkinsonia (Geraniaceae). Plant Systematics and Evolution 170: 151-159.
Gibby, M., S. Hinnah, E.M. Marais AND F. Albers. 1996. Cytological variation and evolution within Pelargonium section Hoarea (Geraniaceae). Plant Systematics and Evolution 203: 111-142
Grazzini, Richard. 1992. Molecular and biochemical protocols for the identification of genotypes. Proceedings of the Third International Geranium Conference: 91- 103.
112 Grazzini, R.A., P.R. Paul, T. Hage, D.L. Cox-Foster, J.I. Medford, R. Craig and R.O. Mumma. 1999. Tissue-specific fatty acid composition of glandular trichomes of mite-resistant and –susceptible Pelargonium xhortorum. Journal of Chemical Ecology 25(4): 955-968.
Guzmán, Félix Alberto, Helmer Ayala, César Azurdia, Myriam Cristina Dque and Carmen de Vicente. 2005. AFLP assessment of genetic diversity of Capsicum genetic resources in Guatemala: Home gardens as an option for conservation. Crop Sci. 45: 363-370.
Hassanein, A., E. Chevreau, and N. Dorion. 2005. Highly efficient transformation of zonal (Pelargonium xhortorum) and scented (P. capitatum) geraniums via Agrobacterium tumefaciens using leaf discs. Plant Sci. 169: 532–541.
Horn, Wolfgang. 1992. Genetics and crossability of Pelargoniums. Third International Geranium Conference Proceedings: 33-47.
Hu, J. 2006. Defining the sunflower (Helianthus annuus L.) linkage group ends with the Arabidopsis-type telomere sequence repeat-derived markers. Chromosome Res. 14: 535-548.
Hu, J. and B. Vick. 2003. Target region amplification polymorphism: a novel marker technique for plant genotyping. Plant Mol. Biol. Rpt. 21: 289–294.
Hu, J., O.E. Ochoa, M.J. Truco, and B.A. Vick. 2005. Application of the TRAP technique to lettuce (Lactuca sativa L.) genotyping. Euphytica 144: 225–235.
Hu, Jinguo, Gerald J. Seiler, Chao-Chien Jan and Brady A. Vick. 2003. Assessing genetic variability among sixteen perennial Helianthus species using PCR-based TRAP markers. Proceedings of the 25th Sunflower Research Workshop, Fargo, N.D.
Huelsenbeck, J. P. and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17:754-755.
James, Celia M., John A. Barrett, Stephen J. Russell and Mary Gibby. 2001. A rapid PCR method to establish the potential for paternal inheritance of chloroplasts in Pelargonium. Plant Molecular Biology Reporter 19: 163-167.
James, C.M., M. Gibby and J.A. Barrett. 2004. Molecular studies in Pelargonium (Geraniaceae). A taxonomic appraisal of section Ciconum and the origin of the ‘Zonal’ and ‘Ivy-leaved’ cultivars. Plant Systematics and Evolution 243: 131- 146.
113 Jerardo, A. 2006. Floriculture and nursery crops outlook, p. 1–25. Electronic Outlook Report from the Economic Research Service (USDA) FLO-05. 22 Sept. 2006.
Kuroiwa, Tsuneyoshi, Tamotsu Kawazu, Hidenobu Uchida, Toshiro Ohta and Haruko Kuroiwa. 1993. European Journal of Cell Biology 62: 307-313.
Li, G. and C.F. Quiros. 2001. Sequence-related amplification polymorphism (SRAP), a new marker system based on a simple PCR reaction: its application to mapping and gene tagging in Brassica. Theor. Appl. Genet. 103: 455–461.
Li, J., D.L. Klindworth, F. Shireen, X. Cai, J. Hu, and S.S. Xu. 2006. Molecular characterization and chromosome specific TRAP-marker development for Langdon durum D-genome disomic substitution lines. Genome 49(12): 1545- 1554.
Lis-Balchin, M., S.L. Hart, S.G. Deans, and E. Eaglesham. 1995. Potential agrochemical and medicinal usage of essential oils of Pelargonium species. J. Herbs Spices Med. Plants 3: 11–22.
Liu, Z.H., J.A. Anderson, J. Hu, T.L. Friesen, J.B. Rasmussen, and J.D. Faris. 2005. A wheat intervarietal genetic linkage map based on microsatellite and target region amplified polymorphism markers and its utility for detecting quantitative trait loci. Theor. Appl. Genet. 111: 782–794.
Loehrlein, Marietta M. and Richard Craig. 2001. History and culture of regal Pelargonium. HortTechnology 11(2): 289-296.
Marais, Elizabeth M. 1989. Delimitation and typification of the section Hoarea of Pelargonium (Geranium). South African Journal of Botany 55(2): 240-242.
Miklas, P.N., J. Hu, N.J. Grunwald, and K.M. Larson. 2006. Potential application of TRAP (targeted region amplified polymorphism) markers for mapping and tagging disease resistance traits in common bean. Crop Sci. 46: 910–916.
Mumma, Ralph O., Richard Craig, Diana Cox-Foster, June Medford, Herbert Grossman, Richard Grazzini, David Hesk, Donald Walters and Ellen Yerger. 1992. Biochemistry and genetics of small arthropod resistance. Third International Geranium Conference Proceedings: 173-184.
Olivier, Maria C. and J.J.A Van der Walt. 1984. The taxonomy of the Pelargonium peltatum (L.) L’Hérit. complex. Journal of South African Botany 50(1): 1-14.
114 Palumbo, Hong, Hu, Craig, Locke, Krause, Tay and Wang. 2007. Target Region Amplification Polymorphism (TRAP) as a Tool for Detecting Genetic Variation in the Genus Pelargonium. HortScience 42(5): 1118-1123
Ravindra, N.S., R.N. Kulkarni, M.C. Gayathri, and S. Ramesh. 2004. Somaclonal variation for some morphological traits, herb yield, essential oil content and essential oil composition in an Indian cultivar of rose-scented geranium. Plant Breed. 123: 84–86.
Renou, J.P., C. Aubry, M. Serveau and P. Jalouzot. 1997. Evaluation of the genetic variability in the genus Pelargonium using RAPD markers. Journal of Horticultural Science 72(2): 229-237.
Rohlf, J.F. 2000. NTsysPC, Ver. 2.11S. Applied Biostatistics Inc., Setauket, NY.
Ronquist, F. and J. P. Huelsenbeck. 2003. MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572-1574.
Saghai-Maroof, M.A., K.M. Soliman, R.A. Jorgensen, and R.W. Allard. 1984. Ribosomal DNA spacer-length polymorphisms in barley: Mendelian inheritance, chromosomal location, and population dynamics. Proc. Natl. Acad. Sci. U.S.A. 81:8014–8018.
Scheltema, A.G. and J.J.A van der Walt. 1990. Taxonomic revision of Pelargonium section Jenkinsonia (Geraniaceae) in southern Africa. South African Journal of Botany 56(3): 285-302.
Schultz, David J., Christian Olsen, Gary A. Cobbs, Neal J. Stolowich and Megan M. Parrott. 2006. Bioactivity of anacardic acid against Colorado potato beetle (Leptinotarsa decemlineata) Larvae. Journal of Agricultural and Food Chemistry 54: 7522-7529.
Schultz, David J., Edgar B. Cahoon, John Shanklin, Richard Craig, Diana L. Cox- Foster, Ralph O. Mumma and June I. Medford. 1996. Expression of a ∆9 14:0- acyl carrier protein fatty acid desaturase gene is necessary for the production of ω5 anacardic acids found in pest-resistant geranium (Pelargonium xhortorum). Proceedings of the National Academy of Sciences USA 93: 8771-8775.
Segarra-Moragues, José Gabriel, José María Iriondo and Pilar Catalán. 2005. Genetic fingerprinting of germplasm accessions as an aid for species conservation: A case study with Borderea chouardii (Diosoreaceae), one of the most critically endangered plants. Annals of Botany 96: 1283-1292.
Sergio, Lanteri and Baraccia Gianni. 2005. Molecular markers based analysis for crop germplasm preservation. The role of biotechnology for the characterization and
115 conservation of crop, forestry, animal and fishery genetic resources. International Workshop Villa Gualino, Turin, Italy.
Starman, Terri Woods and Shane Abbitt. 1997. Evaluating genetic relationships of Geranium using arbitrary signatures from amplification profiles. HortScience 32(7): 1288-1291.
Swofford, D. L. 2002. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4. Sinauer Associates, Sunderland, Massachusetts.
Tanksley, S.D. and S.R. McCouch. 1997. Seed banks and molecular maps: unlocking genetic potential from the wild. Science-New Ser. 277: 1063–1066.
Tay, D. 2003. The Ornamental Plant Germplasm Center-ranking priority genera for conservation. HortScience 38:678.
Tel-zur, N., S. Abbo, D. Myslabodsk, and Y. Mizrahi. 1999. Modified CTAB procedure for DNA isolation from epiphytic cacti of the genera Hylocereus and Selenicereus (Cactaceae). Plant Mol. Biol. Rpt. 17:249–254.
Van der Walt, J.J.A. (ed.). 1977. Pelargoniums of Southern Africa. Vol. 1. Purnell, Cape Town, South Africa.
Van der Walt, J.J. Andri. 1992. Discovering the world of Pelargoniums. Third International Geranium Conference Proceedings: 15-27.
Van der Walt, J.J.A., F. Albers, M. Gibby, D.E. Marschewski and R.A. Price. 1995. A biosystematic study of Pelargonium section Ligularia: 1. A new section Subsucculentia. South African Journal of Botany 61(6): 331-338.
Van der Walt, J.J.A., Focke Albers and Mary Gibby. 1990a. Delimitation of Pelargonium sect. Glaucophyllum (Geraniaceae). Plant Systematics and Evolution 171: 15-26.
Van der Walt, J.J.A. and F. Demarne. 1988. Pelargonium graveolens and P. radens: A comparison of their morphology and essential oils. South African Journal of Botany 54(6): 617-622.
Van der Walt, J.J.A. and J.P Roux. 1991. Taxonomy and phylogeny of Pelargonium section Campylia (Geraniaceae). South African Journal of Botany 57(5): 291- 293.
Van der Walt, J.J.A., H.J.T. Venter, R. Verhoeven and L.L. Dreyer. 1990b. The transfer of Erodium incarnatum to the genus Pelargonium (Geraniaceae). South African Journal of Botany 56(5): 560-564.
116 Van der Walt, J.J.A. and P.J. Vorster. 1988. Pelargoniums of Southern Africa, Vol. 3. In: J.N. Eloff (ed.). Annals of Kirstenbosch Botanic Gardens, Vol. 16. National Botanic Gardens, Kirstenbosch, South Africa.
Wang, T., S.S. Xu, M.O. Harris, J. Hu, L. Liu, and X. Cai. 2006. Genetic characterization and molecular mapping of Hessian fly resistance genes derived from Aegilops tauschii in synthetic wheat. Theor. Appl. Genet. 113: 611–618.
Winkelmann, T., K. Kaviani, and M. Serek. 2005. Development of a shoot regeneration protocol for genetic transformation in Pelargonium zonale and Pelargonium peltatum hybrids. Plant Cell. Tiss. Org. Cult. 80:33–42.
Yang, J., G. Bai, and G.E. Shaner. 2005. Novel quantitative trait loci (QTL) for Fusarium head blight resistance in wheat cultivar Chokwang. Theor. Appl. Genet. 111:1571-1579.
Yu, Jiwen, Shuxun Yu, Cairui Lu, Wu Wang, Shuli Fan, Meizhen Song, Zhongxu Lin, Xianlong Zhang and Jinfa Zhang. 2007. High-density linkage map of cultivated allotetraploid cotton based on SSR, TRAP, SRAP and AFLP markers. Journal of Integrative Plant Biology 49(5): 716-724.
117