phD thesis Esfeld

phD thesis Esfeld
The use of low-copy nuclear genes in the
radiation of the
Macaronesian Crassulaceae
Sempervivoideae –
Phylogeny and evolutionary processes
Dissertation
Korinna Esfeld
Born in Lutherstadt Wittenberg
2009
Dissertation
submitted to the
Combined Faculties for the Natural Sciences and for Mathematics
of the Ruperto-Carola University of Heidelberg, Germany
for the degree of
Doctor of Natural Sciences
presented by
Diplom-Biologin Korinna Esfeld
born in: Lutherstadt Wittenberg
Oral-examination: .......................................
i
The use of low-copy nuclear genes in the radiation of the
Macaronesian Crassulaceae Sempervivoideae –
Phylogeny and evolutionary processes
Referees:
Prof. Dr. Marcus Koch
Prof. Dr. Claudia Erbar
ii
... für meine Familie und meine Freunde,
die geduldig daran geglaubt haben!
Am Ende ist alles gut und wenn es
nicht gut ist, dann ist das nicht das Ende....
iii
Contents
Contents
Contents................................................................................................................................ iv
Figures ...................................................................................................................................v
Tables .................................................................................................................................. vii
0.1. Summary...................................................................................................................... viii
0.2. Zusammenfassung ......................................................................................................... ix
1. Introduction ....................................................................................................................... 1
2. Materials and Methods .....................................................................................................21
2.1. Study species ............................................................................................................21
2.2. Molecular methods ....................................................................................................26
2.2.1. RNA material ......................................................................................................26
2.2.2. DNA material ......................................................................................................26
2.2.3. Amplification of the low-copy nuclear genes........................................................27
2.2.4. Cloning of the low-copy nuclear genes................................................................30
2.2.5. Colony PCR ........................................................................................................31
2.2.6. Amplification of nrITS and cpDNA regions ..........................................................32
2.2.7. Sequencing.........................................................................................................32
2.2.8. Sequence analysis..............................................................................................32
2.2.9. Definition of the gene regions .............................................................................33
2.2.10. Improvements of the datasets ...........................................................................33
2.2.11. Partition Homogeneity Tests .............................................................................34
2.2.12. Phylogenetic reconstructions ............................................................................34
2.2.13. Blast analyses...................................................................................................35
2.2.14. Neighbor-joining reconstructions for homologs of PEPC, AP1, and AP3...........35
2.2.15. Species phylogeny............................................................................................36
2.2.16. Nucleotide differences, replacements, and amino acid substitutions.................36
2.2.17. Relative Rate Tests...........................................................................................37
2.2.18. Selection pressure ............................................................................................37
3. Results .............................................................................................................................39
3.1. Datasets ....................................................................................................................39
3.2. Phylogenetic reconstructions .....................................................................................45
3.3. Blast and Neighbor-joining analyses..........................................................................64
3.4. Species phylogeny based on nrITS ...........................................................................66
3.5. Gene duplications......................................................................................................67
3.6. Nucleotide differences, replacements, and amino acid substitutions..........................69
3.7. Relative Rate Tests ...................................................................................................73
3.8. Ka/Ks-values .............................................................................................................74
3.9. Selection pressure.....................................................................................................76
4. Discussion........................................................................................................................78
4.1. Phylogenetic reconstructions .....................................................................................78
4.2. Gene duplications......................................................................................................86
4.3. Selection pressure.....................................................................................................97
4.4. Regulatory versus structural genes..........................................................................103
5. Summary........................................................................................................................108
Literature ............................................................................................................................110
Abbreviations .....................................................................................................................125
Overview of scientific contributions.....................................................................................127
Appendices ........................................................................................................................128
Acknowledgment ................................................................................................................150
iv
Figures
Figures
Fig. 1: Map of Macaronesia and the Canary Islands after Mort et al. (2002). Taken from
the website http://www.eiu.edu/~bio_data/posters/2002/poster_016.htm............................... 2
Fig. 2: Maximum parsimony phylogram of the MCS species based on cpDNA/nrITS data
(from Mort et al. 2002). .......................................................................................................... 9
Fig. 3: MIKC-like structure of the plant MADS-box proteins (adapted after Purugganan
et al. 1995). ..........................................................................................................................15
Fig. 4: The extended ABCDE model adapted after Erbar (2007)..........................................17
Fig. 5: Schematic exon and intron structure of the MCS_PEPC gene sequences. Exons
are shown as boxes and introns as lines. The relative length of the respective parts is
given in table 4. ....................................................................................................................40
Fig. 6: Schematic exon and intron structure of the MCS_AP1 gene sequences. Exons
are shown as boxes and introns as lines. The relative length of the respective parts is
given in table 6. ....................................................................................................................42
Fig. 7: Schematic exon and intron structure of the MCS_AP3 gene sequences. Exons
are shown as boxes and introns as lines. After the last exon box the 3´-UTR is indicated
as line. The relative length of the respective parts is given in table 8....................................44
Fig. 8: BI phylogram based on the full-length MCS_PEPC data. Posterior probabilities
are given at the nodes. .........................................................................................................47
Fig. 9: BI phylogram based on the MCS_PEPC exon data. Posterior probabilities are
given at the nodes. ...............................................................................................................49
Fig. 10: ML phylogram based on the MCS_PEPC intron data. Bootstrap support is given
at the nodes. ........................................................................................................................51
Fig. 11: ML phylogram based on the MCS_AP1 full-length data. Bootstrap support is
given at the nodes. ...............................................................................................................53
Fig. 12: ML phylogram based on the MCS_AP1 exon data. Bootstrap support is given
at the nodes. ........................................................................................................................55
Fig. 13: BI phylogram based on the MCS_AP1 intron data. Posterior probabilities are
given at the nodes. ...............................................................................................................57
Fig. 14: ML phylogram based on the MCS_AP3 full-length data. The 3´-UTR region is
excluded. Bootstrap support is given at the nodes................................................................59
Fig. 15: BI phylogram based on the MCS_AP3 exon data. Posterior probabilities are
given at the nodes. ...............................................................................................................61
Fig. 16: BI phylogram based on the MCS_AP3 intron data. Posterior probabilities are
given at the nodes. ...............................................................................................................63
Fig. 17: BI phylogram based on the improved nrITS dataset. Own sequences are
marked with numbers. Posterior probabilities are given at the nodes. ..................................67
Fig. 18-28: Distribution of Ka/Ks-values for MCS_PEPC, MCS_AP1, and MCS_AP3
genes for the MCS and the Sedum sister- as well as outgroup species. Indication of the
copies follows the phylograms in fig. 8-16. ...........................................................................75
Fig. 29: ML phylogram based on the MCS_PEPC full-length data. Bootstrap support is
given at the nodes. .............................................................................................................132
Fig. 30: ML phylogram based on the MCS_PEPC exon data. Bootstrap support is given
at the nodes. ......................................................................................................................133
Fig. 31: BI phylogram based on the MCS_PEPC intron data. Posterior probabilities are
given at the nodes. .............................................................................................................134
Fig. 32: BI phylogram based on the MCS_PEPC MCS intron data (Sedum sequences
were excluded). Posterior probabilities are given at the nodes. ..........................................135
Fig. 33: ML phylogram based on the MCS_PEPC MCS intron data (Sedum sequences
were excluded). Bootstrap support is given at the nodes....................................................136
Fig. 34: BI phylogram based on the MCS_AP1 full-length data. Posterior probabilities
are given at the nodes. .......................................................................................................137
Fig. 35: BI phylogram based on the MCS_AP1 exon data. Posterior probabilities are
given at the nodes. .............................................................................................................138
v
Figures
Fig. 36: ML phylogram based on the MCS_AP1 intron data. Bootstrap support is given
at the nodes. ......................................................................................................................139
Fig. 37: BI phylogram based on the MCS_AP1 MCS intron data (Sedum sequences
were excluded). Posterior probabilities are given at the nodes. ..........................................140
Fig. 38: ML phylogram based on the MCS_AP1 MCS intron data (Sedum sequences
were excluded). Bootstrap support is given at the nodes....................................................141
Fig. 39: BI phylogram based on the MCS_AP3 full-length data. The 3´-UTR-region is
excluded. Posterior probabilities are given at the nodes. ....................................................142
Fig. 40: ML phylogram based on the MCS_AP3 exon data. Bootstrap support is given
at the nodes. ......................................................................................................................143
Fig. 41: ML phylogram based on the MCS_AP3 intron data. Bootstrap support is given
at the nodes. ......................................................................................................................144
Fig. 42: BI phylogram based on the MCS_AP3 MCS intron data (Sedum sequences
were excluded). Posterior probabilities are given at the nodes. ..........................................145
Fig. 43: ML phylogram based on the MCS_AP3 MCS intron data (Sedum sequences
were excluded). Bootstrap support is given at the nodes....................................................146
Fig. 44: BI phylogram based on exon data of MCS_AP1 including all amplified
sequences (chimers, sequences with unique splice sites or introns, and with premature
stops caused by frameshift mutations). Posterior probabilities are given at the nodes........147
Fig. 45: NJ phylogram of the enlarged MCS_PEPC dataset. Bootstrap support is given
at the nodes. ......................................................................................................................148
Fig. 46: Ultrametric tree. Posterior probabilities are given at the nodes. Asterisks mark
the positions of potential alternative duplication events. .....................................................149
vi
Tables
Tables
Table 1: Overview of the main characteristics of the studied species...................................26
Table 2: Overview of the components used in PCR reactions for the different
polymerases.........................................................................................................................30
Table 3: Number of obtained total and reduced cloned sequences for MCS_PEPC. ...........39
Table 4: Exon and intron positions and lengths of the MCS_PEPC sequences....................40
Table 5: Number of obtained total and reduced cloned sequences for MCS_AP1. ..............41
Table 6: Exon and intron positions and lengths of the MCS_AP1 sequences. .....................42
Table 7: Number of obtained total and reduced cloned sequences for MCS_AP3. ..............43
Table 8: Exon and intron positions and lengths for MCS_AP3 sequences. ..........................44
Table 9: Selected model (AIC criterion) and results of the BI analyzes for different
datasets. Default settings of MrBayes (four chains, sample every 100th generation,
random starting tree) were used...........................................................................................45
Table 10: Number and percentage of nucleotide differences, replacements, and quite
different amino acids (aa) between species-specific sequences of MCS_PEPC
distinguished into orthologous and paralogous sequences...................................................71
Table 11: Number and percentage of nucleotide differences, replacements, and quite
different amino acids (aa) between species-specific sequences of MCS_AP1
distinguished into orthologous and paralogous sequences...................................................72
Table 12: Number and percentage of nucleotide differences, replacements, and quite
different amino acids (aa) between species-specific sequences of MCS_AP3
distinguished into orthologous and paralogous sequences...................................................73
Table 13: Mean Ka/Ks-values. Indication of the copies follows the phylograms in
fig. 8-16. ...............................................................................................................................74
Table 14: Comparisons of Ka/Ks-values for the respective genes for 1) copy A and B
of the MCS species, 2) copies of the MCS vs. sister- and outgroup (OG) species,
respectively, and 3) regulatory vs. structural genes based on t-tests (n.s. = not
significant, * significant (p < 0.05) and ** highly significant (p < 0.01)). .................................76
Table 15: Comparison of the infrageneric classification of Aeonium: Lems (1960),
Liu (1989), Mes (1995), and Mort et al. (2002). Studied species are indicated in bold. .......129
Table 16: Location of the collection sites of the studied species. Indicated are area and
name of the collection site as well as coordinates in Degrees Minutes and Seconds. ........130
Table 17: Primers used to amplify the respective gene regions. Indicated are name,
gene region, sequence, and application. ............................................................................131
vii
Summary
0.1. Summary
Speciation and evolution of species are two of the most exciting topics in biology.
Radiations, with their wide morphological and physiological variety, provide a
promising tool to understand speciation and diversity of species. Numerous studies
have revealed that the high morphological diversity of radiated species is not
represented at the molecular level. Neutral markers like rDNA nrITS and chloroplast
(cp) DNA evolve slowly compared to speciation in radiations and thus, may not
provide enough information to resolve phylogenetic relationships. In contrast, lowcopy nuclear genes evolve faster and may help to resolve relationships. This is
supported by the hypothesis that accelerated changes in regulatory genes, as
opposed to structural genes, can explain the evolution of species.
To contribute to this ongoing discussion, the radiation of the Macaronesian
Crassulaceae Sempervivoideae (MCS) was studied. The polyploid species of the
MCS are mainly distributed on the Canary Islands and comprise more than 70
species in three genera (Aeonium, Aichryson, and Monanthes) that display a huge
morphological (e.g., flower color, number of floral organs, growth-form) and
physiological (e.g., CAM activity) variety.
Two regulatory genes, homologs of the floral homoeotic genes APETALA1 and
APETALA3, and the structural gene encoding for phosphoenolpyruvate carboxylase
(PEPC) were analyzed with respect to the following aims: 1) to evaluate the use of
the low-copy nuclear genes to reconstruct phylogenies and to compare genealogies
with the species phylogeny; 2) to estimate the impact of the studied genes in the
speciation process and elucidate differences between the roles of regulatory and
structural genes; 3) to determine if gene duplications occurred and to distinguish
duplicates into orthologs and paralogs, and 4) to calculate the selection pressure
(Ka/Ks-values) acting on the respective gene copies.
The three analyzed low-copy nuclear genes both support and contradict the
phylogenetic relationships inferred by other markers. The selection acting on the
studied low-copy genes is in contrast with the neutral evolution of nrITS and cpDNA
markers and may explain observed differences. In particular APETALA3 seems to be
a promising marker for resolving species relationships.
In addition, the studied genes may have had an influence in speciation since
individually they exhibit accelerated Ka/Ks-values compared to mean Ka/Ks-values
estimated for regulatory and structural genes. Their Ka/Ks-values are also much
higher than those obtained for other genes in studies with comparable experimental
designs. Accelerated evolutionary rates were estimated for the regulatory genes as
opposed to the structural gene PEPC. However, summarizing all observations, the
impact of these genes may be limited. Further study is recommended to evaluate
their true impact.
For all studied genes duplications were observed and emphasize the greatest
challenge of working with low-copy nuclear genes – the differentiation of orthologs
and paralogs. The observed duplication pattern suggests that the gene duplications
are the result of polyploidization, a phenomenon to which the island colonization of
the MCS species was connected previously.
In addition, all gene copies were under purifying selection pressure, even if the
estimated Ka/Ks-values for the respective copies varied. Rate differences were
estimated for PEPC and APETALA3; the latter also showed significant differences in
the Ka/Ks-values comparing copy A and copy B. For APETALA1 similar evolutionary
rates and highest Ka/Ks-values were found.
Altogether, this thesis offers a promising approach to study speciation and evolution
in the radiation of the MCS and is a valuable basis for further studies.
viii
Zusammenfassung
0.2. Zusammenfassung
Artbildung und Evolution gehören zu den spannendsten Themen der Biologie. Für
deren Verständnis bieten Radiationen mit ihrer morphologischen und physiologischen Variation eine wichtige Grundlage. Zahlreiche Studien zeigten dabei, dass
die morphologische Diversität radiierter Arten nicht mit ihrer molekularen Diversität
korreliert. Neutrale molekulare Marker (rDNA nrITS, Chloroplasten-DNA) evolvieren
im Vergleich zur Artbildung in Radiationen langsam und eignen sich demzufolge
zumeist nicht, um die verwandtschaftlichen Beziehungen wiederzugeben.
Kodierende Kerngene jedoch, die in einer geringen Kopiezahl vorliegenden,
evolvieren schneller und könnten Verwandtschaftsbeziehungen radiierter Arten
widerspiegeln. Zudem wird ein größerer Einfluss von regulatorischen Genen im
Vergleich zu Strukturgenen auf die Artbildung diskutiert.
Um zu dieser Diskussion beizutragen, wurde die Radiation der Makaronesischen
Semperviven (MCS) untersucht. Diese hauptsächlich auf den Kanaren verbreitete
Artengruppe umfasst mehr als 70 polyploide Arten in den Gattungen Aeonium,
Aichryson und Monanthes. Die Arten zeichnen sich durch hohe morphologische (z.B.
Blütenfarbe, Anzahl der Blütenorgane, Wuchsform) und physiologische (z.B.
Unterschiede in CAM-Aktivität) Variation aus. Zwei regulatorische Gene, Homologe
der Blütenmorphologiegene APETALA1 und APETALA3 sowie das für die
Phosphoenolpyruvat-Carboxylase kodierende Strukturgen (PEPC) wurden auf ihren
Einfluss bezüglich der Artbildung hin untersucht. Die Ziele waren dabei 1) die
Eignung der untersuchten Kerngene für phylogenetische Analysen zu bestimmen
sowie Genbäume und Artbaum zu vergleichen, 2) den Einfluss der Gene auf den
Artbildungsprozess zu bestimmen und Unterschiede zwischen Regulator- und
Strukturgenen zu determinieren, 3) Genduplikationen, unterschieden in Orthologe
und Paraloge, nachzuweisen und 4) den auf die jeweiligen Genkopien wirkenden
Selektionsdruck (Ka/Ks-Werte) zu ermitteln.
Die analysierten Kerngene spiegeln nur teilweise die Artphylogenie wider und Unterschiede zwischen Genbäumen und dem Artbaum wurden beobachtet. Hierbei spielt
die auf den Kerngenen wirkende Selektion eine wesentliche Rolle, da diese im
Gegensatz zu der Evolution von nrITS und Chloroplasten-DNA Markern nicht neutral
ist. Speziell APETALA3 ist jedoch ein vielversprechender phylogenetischer Marker.
Ein Einfluss der analysierten Gene auf die Artbildung ist denkbar. Ihre durchschnittlichen Ka/Ks-Werte liegen über denen anderer Regulator- und Strukturgene und sind
höher als Ka/Ks-Durchschnittswerte anderer Gene, die in ähnlichen Studien ermittelt
wurden. Auch konnte eine höhere Evolutionsrate der Regulatorgene im Vergleich zu
dem Strukturgen ermittelt werden. Insgesamt ist der Einfluss jedoch wohl limitiert und
weitere Untersuchungen notwendig, um den tatsächlichen Einfluss zu ermitteln.
Genverdopplungen, die bei allen Genen beobachtet wurden, verdeutlichen, dass die
Unterscheidung in Orthologe und Paraloge eine wesentliche Herausforderung bei der
Arbeit mit Kerngenen darstellt. Das dabei beobachtete Muster legt nahe, dass die
Verdopplungen auf Polyploidisierung, die mit der Inselbesiedlung gekoppelt war,
zurückzuführen sind.
Alle Genkopien unterliegen unterschiedlich starker reinigender Selektion. Unterschiedliche Evolutionsraten wurden bei PEPC sowie bei APETALA3 gefunden, wo
zudem signifikant unterschiedliche Ka/Ks-Werte für die Kopien beobachtet wurden.
APETALA1 Genkopien haben die höchsten Ka/Ks-Werte, zeigen jedoch hierbei
sowie bei der Evolutionsrate keine signifikanten Unterschiede.
Die vorliegende Arbeit bietet einen sehr vielversprechenden Ansatz für die
Untersuchung von Artbildung und Evolution in der Radiation der MCS und bildet eine
wichtige Basis für weiterführende Untersuchungen.
ix
Introduction
1. Introduction
The spectacular diversity of plant and animal species is an ongoing marvel for human
beings. Speciation processes and evolution fascinate scientists from all fields and
raise numerous questions and discussions. For example, traditionally adaptation and
speciation was discussed due to fixation of many genes with small effects (see, e.g.,
Schemske and Bradshaw 1999). However, extensive studies were done recently to
define major forces acting in speciation and examples are known where few traits
and genes highly influence evolution. One of the key morphological differences
between maize and teosinte is encoded by the teosinte-branched-1 locus (for
reference see Ford 2002). Innovation of nectar spurs lead to the extensive radiation
of Aquilegia since this key character allowed specialization to different pollinators, a
well-known speciation mechanism (Hodges and Arnold 1994a, Hodges 1997, Whittall
and Hodges 2007). In the genus Mimulus, a single homozygous mutation in a gene
of the flavonoid pathway causes a shift in flower color resulting in different pollinator
preferences (Schemske and Bradshaw 1999, see also Durbin et al. 2003). Regarding
these studies, it seems that a small number of genes may be responsible for the
evolution of major differences in growth-form and flower morphology of plants.
Central to speciation seems to be the accelerated evolution of regulatory genes,
which has been proposed as a main explanation for the morphological development
of species (King and Wilson 1975, Remington and Purugganan 2002, Durbin et al.
2003, Purugganan and Robichaux 2005).
Radiations, with their wide morphological and ecological variety, offer a promising
tool to answer questions about speciation, species diversity, and evolution (Whittall et
al. 2006). They are common on islands where extraordinary speciation is favored due
to high diversity of habitats and lack of competition after colonization (MacArthur and
Wilson 1967, Arnedo et al. 1996). Therefore, islands are seen as natural laboratories
and model systems to study speciation (e.g., Francisco-Ortega et al. 1996, Baldwin
et al. 1998). In the past, most studies concerning plant and animal evolution focused
on the Pacific islands of Galapagos, Hawaii or Juan Fernandez that are several
thousand kilometers away from their respective continent (Baldwin et al. 1998, see
also Jorgensen and Olesen 2001). Recently, also Atlantic islands have become the
focus of interest. Macaronesia, comprising the Azores, Canary Islands, Cape Verde,
Madeira, and Salvages (fig. 1), serves as an ideal system to understand the origin
1
Introduction
and evolution of island biota and the consequences of colonization and isolation. The
seven main islands and several islets of the Canaries (fig. 1) are of particular interest
due to their geographical proximity to the African continent. They are situated in the
northeast Atlantic Ocean between 27°37’ and 29°25’N and 13°20’ and 18°10’W.
Fuerteventura is closest to the continent, approximately 110 km away, and also La
Palma is situated only 460 km away from the northwest African mainland (Carracedo
1994, Juan et al. 2000). Despite known sea mountains close to the sea level
between the Canaries and the mainland of Portugal (Francisco-Ortega et al. 2000), a
connection with the continent was never given (Andrus et al. 2004 and references
therein, Sanmartín et al. 2008).
Fig. 1: Map of Macaronesia and the Canary Islands after Mort et al. (2002). Taken from the website
http://www.eiu.edu/~bio_data/posters/2002/poster_016.htm.
The Canary Islands are of volcanic origin and provide a broad range of geological
ages. Nearly ordered in a line, El Hierro is the youngest and westernmost island and
Fuerteventura the oldest and easternmost. Ages of the seven main islands are 20.7
My for Fuerteventura, 15.5 My for Lanzarote, 13.9-16 My for Gran Canaria, 11.6 My
for Tenerife, 10-12.5 My for La Gomera, 1.5-2 My for La Palma, and > 0.7 My for El
Hierro (Carracedo 1994, Arnedo et al. 1996, Kim et al. 1996).
The combination of important factors such as geology, trade winds, elevation, and
inclination has lead to distinct vegetation zones on the Canary Islands providing the
fundament for evolutionary processes (Lems 1960, Baldwin et al. 1998). The humid
and cool northeastern trade winds lead to two main climatic regions: one humid due
2
Introduction
to the influence of these trade winds and the other not and therefore more arid.
Ecological zones range from coastal desert and lowland scrub to humid laurel forest,
pine forest, and alpine desert (Francisco-Ortega et al. 1996, Jorgensen and
Frydenberg 1999).
Additionally volcanic activity has provided unstable conditions, which highly
influences fast and divergent speciation by founder events, genetic bottlenecks, and
genetic drift (Kull 1982, Nyffeler 1995, Jorgensen and Olesen 2001). Also inter-island
dispersal and colonization of similar ecological zones on different islands has
influenced speciation. Thus, the Canary Islands exhibit well-known examples of large
radiated groups: animal species like geckos, several genera of Coleoptera, e.g.,
Nesotes or the spider genus Dysdera, and plant genera such as Aeonium,
Argyranthemum, Echium, Sideritis, and Sonchus (Kull 1982, Lösch 1990, Arnedo et
al. 1996, Böhle et al. 1996, Francisco-Ortega et al. 1996, Kim et al. 1996, Barber et
al. 2000, Juan et al. 2000, Emerson 2002, Mort et al. 2002).
One of the most famous of the radiated plant groups are the Macaronesian
Crassulaceae Sempervivoideae (MCS). This group comprises approximately 70
species within the three genera Aeonium Webb & Berthel. (including the former
separated genus Greenovia Webb & Berthel.), Monanthes Haw., and Aichryson
Webb & Berthel. and is a well supported group of the Crassulaceae (Berger 1930,
Praeger 1932, Lems 1960, Liu 1989, Nyffeler 1992, Mes 1995, Mort et al. 2002).
The family of the Crassulaceae is monophyletic, part of the Saxifragales clade, and
comprises six subfamilies, 35 genera, and 1500 species (Berger 1930, Liu 1989, van
Ham and ‘t Hart 1998, Fishbein et al. 2001, Mort et al. 2001). The species in this
group are morphologically diverse. Common features are succulent leaves and
pentamerous, radially symmetrical flowers with one or two whorls of stamens (van
Ham and ‘t Hart 1998, Mort et al. 2001). Crassulaceae are mainly herbaceous but for
several genera like Aeonium, Cotyledon, Crassula, Kalanchoe, and Sedum woody
species are known (Mes et al. 1996). Members of the family predominately inhabit
semiarid to arid and mountainous habitats (Mes et al. 1996, van Ham and ‘t Hart
1998). They are adapted to low water supply by their succulent leaves and their
special CO2 fixation pathway, the Crassulacean Acid Metabolism (CAM; Lösch
1990). Crassulaceae are distributed worldwide but predominately found in subtropical
and temperate zones (van Ham and ‘t Hart 1998). The origin of the family lies in
3
Introduction
Southern Africa or in the Mediterranean area with centers of diversity in Mexico,
South Africa, Himalaya, the Mediterranean region, and Macaronesia (van Ham and ‘t
Hart 1998, Mort et al. 2001).
Within the Crassulaceae Berger (1930) defined six subfamilies and 33 genera based
on the number and arrangement of floral parts, degree of sympetaly, and phyllotaxis,
but five of his described subfamilies are polyphyletic (van Ham and ‘t Hart 1998, Mort
et al. 2001). He defined two lineages, a “Crassula lineage” and a “Sedum lineage”,
mainly separated by geography (van Ham and ‘t Hart 1998, Mort et al. 2001). This
well-supported split at the base of the Crassulaceae nowadays defines the
separation into two subfamilies: Crassuloideae and Sedoideae (van Ham and ‘t Hart
1998, Mort et al. 2001). The Crassuloideae contains only the Crassula clade whereas
the Sedoideae comprises the remaining subfamilies (van Ham and ‘t Hart 1998).
Four clades could be found within the Sedoideae: Acre, Aeonium, Kalanchoe, and
Leucosedum. While evidence also exists for separate Telephium and Sempervivum
clades, the genus Sedum is highly polyphyletic (Mort et al. 2001).
Berger (1930) also defined a Sempervivoideae clade, classifying species with
polymerous flowers. It comprises the genera Sempervivum, Aeonium, Aichryson,
Greenovia, and Monanthes. Another classification of the Sempervivoideae was done
by Praeger (1932) who recognized these five sections, 32 species, one variety, and
25 hybrids (Liu 1989). Nevertheless, the Sempervivoideae are not monophyletic and
a separation between the genus Sempervivum and the monophyletic MCS,
comprising the remaining four genera, has been proposed (Mes et al. 1996, Mort et
al. 2001).
As indicated by their name, the species of the Macaronesian Crassulaceae
Sempervivoideae are nearly endemic in Macaronesia and mainly restricted to the
Canary Islands (Mort et al. 2001). Nevertheless, Aeonium leucoblepharum and A.
stuessyi are known in eastern Africa (Liu 1989). This disjunct distribution pattern
brought up a controversial debate about the origin of the group (Mes 1995, Mes et al.
1996). Traditionally species of Aeonium were regarded as descendants of African
progenitors of Tertiary resembling (Bramwell 1976, Kim et al. 1996), but the
expansion of the Sahara desert resulted in extinction of these ancestral species. Only
a few Aeonium species survived in northern Africa and western Morocco, dispersed
4
Introduction
to the Canary Islands and gave rise to the present diversity through island speciation
and adaptive radiation (Berger 1930, Lems 1960, Liu 1989, Lösch 1990).
Recently, molecular markers have revealed another picture. It now seems
indisputable that the MCS arose from a single colonization event from northern Africa
and diversified on the Canary Islands into the present monophyletic species group
(Mes 1995, Mes et al. 1996, Mort et al. 2001). The large shrubby Aeonium species of
East Africa are the result of dispersal events from Macaronesia back to the African
continent (Mes et al. 1996).
Following the debate of the origin of the clade, there were contradicting views about
the woodiness of the MCS species. For species of Atlantic islands woodiness was
traditionally seen as relict status (Lems 1960, Bramwell 1976, Liu 1989) conflicting
the theory of Carlquist (1962, 1974). Carlquist proposed that the large and woody
taxa of otherwise predominately herbaceous species on oceanic islands evolved
through an increased secondary “insular woodiness” (see Jorgensen and Frydenberg
1999).
Molecular and physiological data support Carlquist’s theory and suggest that the
MCS derived from an herbaceous ancestor (Pilon-Smits et al. 1992, Mort et al. 2002)
which is also supported by the terminal position of the woody African taxa in
phylogenetic reconstructions (Mes et al. 1996). For species of the Pacific islands
woodiness was always considered as being derived on islands by rapid evolution
from ancestral herbaceous progenitors (Baldwin et al. 1998). Recent studies
confirmed this evolutionary pathway as well as derived woodiness for species of the
Atlantic islands, e.g., in the genera Aeonium, Echium or Sonchus (Böhle et al. 1996,
Kim et al. 1996, Mes at al. 1996).
Based on karyological evidence, Uhl (1961) discussed the relationship between
North African Sedum species and the MCS. Based on this information, Mes (1995)
subsequently described the three herbaceous Sedum species, S. jaccardianum, S.
modestum, and S. surculosum (Sedum series Monanthoidea) as well as S.
caeruleum and S. pubescens, as basal to the MCS clade. This suggestion was
confirmed by several studies using different nuclear DNA (nrDNA; the internal
transcribed spacer region = nrITS) and chloroplast DNA (cpDNA) molecular markers
(Mes and ‘t Hart 1994, Mes et al. 1996, van Ham and ‘t Hart 1998).
5
Introduction
The respective Sedum species are diploid with exception of the tetraploid S.
surculosum (for reference see Mes and ‘t Hart 1994). The chromosome base
number of the species is variable: for S. caeruleum x = 12-13, for S. pubescens x =
11, and for S. jaccardianum, S. modestum, and S. surculosum x = 8.
The species of Aeonium and Monanthes have a strict chromosome base number of x
= 18 and for the species of Aichryson x = 15, 16, and 17 is reported. All MCS species
are exclusively polyploid and thus, colonization of the Canaries was accompanied by
polyploidization (Mes 1995, Mes et al. 1996, Mort et al. 2001). The increase in
chromosome number must have occurred either in North Africa in an extinct
ancestral species or on the Canary Islands (Mes et al. 1996). Colonization connected
with polyploidization and without further change of the chromosome base number is
a well known phenomenon for radiations and island speciation (Mes et al. 1996 and
references therein). However, within the MCS subsequent increase in chromosome
numbers (tetraploid up to hexaploid species, e.g., A. arboreum, A. haworthii, A.
leucoblepharum, A. simsii, A. urbicum, Ai. pachycaulon, M. anagensis, and M.
polyphylla) are known (Mes 1995).
Beside polyploidization, hybridization is assumed to play a minor role in the MCS,
and only few species of hybrid origin are known (Mes and ‘t Hart 1996, Mes et al.
1997, van Ham and ‘t Hart 1998, Jorgensen and Frydenberg 1999, Mort et al. 2002).
In general, island species and species of radiated clades are considered as highly
cross compatible with few or no internal barriers for crossing (Whitkus 1998,
Jorgensen and Olesen 2001). Even if, e.g., most Aeonium species have very
restricted habitat requirements and show ecological and topographic isolation (Kull
1982, Jorgensen and Frydenberg 1999), almost all species of Aeonium, Aichryson,
and Monanthes can hybridize when occurring in sympatry and with overlapping
flowering times (Lems 1960, Jorgensen and Olesen 2001). Hybrids are viable and
fertile even if parental species seem to be fitter and survive better (Liu 1989, Nyffeler
1992). It is possible that the rarity of hybrids in the field and the lack of extensive
hybrid zones are caused by the absence of suitable habitats for hybrid species (Lems
1960). In addition, a pollinator study has shown that infraspecific pollen transfer
prevails and interspecific pollen transfer is limited in nature (Esfeld et al. 2009).
6
Introduction
Whereas the division in the three genera Aeonium, Aichryson, and Monanthes is
widely accepted (Mes 1995, Mort et al. 2002), relationships within the group and
especially within the genera and clades are rather unresolved and highly debated.
For
classification
and
phylogenetic
reconstruction
biochemical,
cytological,
molecular, morphological, and physiological data have been used (Lems 1960, Uhl
1961, Tenhunen et al. 1982, Liu 1989, Lösch 1990, Nyffeler 1992, Pilon-Smits et al.
1992, Mes 1995, Nyffeler 1995, Stevens 1995, Mes and ‘t Hart 1996, Mes et al.
1996, 1997, van Ham and ‘t Hart 1998, Jorgensen and Frydenberg 1999, Mort et al.
2001, 2002, Fairfield et al. 2004, Mort et al. 2007) and important aspects of the three
genera are summarized below.
The genus Aeonium Webb & Berthel. (including Greenovia Webb & Berthel.) is
formed of more than 40 species and is the largest radiation both on the Canary
Islands and within the MCS. The species of this genus display a huge morphological
and physiological variety. Their growth-forms range from small rosettes herbs to
large woody plants including subshrubs, shrubs, candelabrum, and rosettes trees.
Flower color varies between white, green, yellow, and rose nuances to dark red for A.
nobile. The species colonize habitats in all vegetation zones of the Canary Islands
and habitats are connected to physiological adaptation, e.g., strength of the C3 and
CAM activity (Pilon-Smits et al. 1992). The species are mainly distributed as single
island endemics on all Canary Islands. Only six species are known outside this
centre of diversity; they are distributed on Madeira (2 species), Cape Verde (1), in
Morocco (1), and Eastern Africa (2; Mes 1995, Jorgensen and Frydenberg 1999).
Historically, the genus Greenovia was separated from the genus Aeonium according
to the derived number of flower organs, the absence of nectariferous scales, and
different placentation (Berger 1930, Liu 1989, Mes 1995). Greenovia species are
characterized by small, hapaxanth rosettes and highly polymerous (16-35merous)
flowers (Mes and ‘t Hart 1996). However, molecular data confirm the inclusion of the
four Greenovia species in the genus Aeonium (Mes 1995, Mort et al. 2002).
Additionally, this is supported by the chromosome base number and because
hybridization is possible between species of Greenovia and Aeonium (Praeger 1932,
see reference in Mort et al. 2002).
While classifying the Aeonium clade Lems (1960) largely followed the classification of
Praeger (1932). He focused on several growth characteristics such as lignification
7
Introduction
and length of the stem, branching pattern, inflorescence size, and pattern, type, and
size of leaves. Section Holochrysa comprises crassicaulous, monocarpic shrubs
with 8-12merous yellow flowers. In the sect. Megalonium A. nobile, a monocarpic
shrub with 7-8merous dark red flowers, is the only species. Section Urbica
comprises several crassicaulous or woody shrubs or subshrubs with white or pink 69merous flowers. Within sect. Canariensia branched or monocarpic rosette plants
were found with yellow 7-13merous flowers. Finally, sect. Goochia is characterized
by dwarf shrubs and stoloniferous rosette plants with yellow and 7-16merous flowers.
Liu (1989) focused on typical morphological characters (e.g., plant height, branching
type, surface reticulation of the stem, length, width, and thickness of leaves,
inflorescence position, and number of flower organs) for phenetic and cladistic
analyses. In total, he used 39 characters to define seven sections, 31 taxa, and six
varieties. Section Petrothamnium is represented by small twiggy subshrubs with
yellow flowers (pinkish in A. goochiae). Section Chrysocome contains perennial
terrestrial twiggy subshrubs without nectariferous glands and yellow flowers. In sect.
Patinaria species are biennial to perennial herbs that form rosettes and have pale
yellow to nearly white flowers. Section Aeonium comprises perennial terrestrial
subshrubs with yellow flowers. The monotypic sect. Megalonium contains again only
A. nobile. To sect. Pittonium belong perennial terrestrial subshrubs with yellow
flowers and sect. Leuconium comprises perennial terrestrial subshrubs that have
white petals, often with pink streaks on central regions.
Studies based on molecular markers were, e.g., done by Mes (1995), Mes and ‘t Hart
(1996), Jorgensen and Frydenberg (1999), and Mort et al. (2002). Mes (1995) used
chloroplast and nuclear spacer sequences to study phylogenetic relationships and
inferred
nine
sections:
Aeonium,
Canariensia,
Chrysocome,
Goochiae,
Greenovia, Leuconium, Patinaria, Petrothamnium, and Pittonium. Mort et al.
(2002) provided the most extensive study based on nrITS and cpDNA markers in the
genera of MCS and described four Aeonium subclades. The infrageneric
classification is summarized in table 15 (appendix). Figure 2 displays the
phylogenetic relationships of the MCS species based on combined cpDNA and nrITS
data of Mort et al. (2002).
8
Introduction
Fig. 2: Maximum parsimony phylogram of the MCS species based on cpDNA/nrITS data (from Mort et
al. 2002).
The monophyletic genus Aichryson Webb & Berthel. contains about 13 species
and is basal and sister to Aeonium and the perennial Monanthes species (Mort et al.
2002, Fairfield et al. 2004; fig. 2). The genus can be differentiated into six major
clades and two lineages that correspond to habit and growth-form. One consists of
the woody, perennial species Ai. tortuosum and Ai. bethencourtianum; the second
comprises the herbaceous, annual members of the genus (Fairfield et al. 2004). The
woody species are CAM species whereas the species of the second group mainly
use the C3 gas-exchange pathway (Lösch 1990). Species are generally highly
branched, predominately hapaxanth, and up to 40 cm high (Mes 1995). They are
9
Introduction
characterized by 6-12merous flowers with digitate nectariferous glands (Liu 1989,
Fairfield et al. 2004).
Aichryson species colonize moist, shady habitats and are especially common in the
laurel forest belt. Ten species are endemic to the Canary Islands whereas three, Ai.
villosum, Ai. divaricatum, and Ai. dumosum, occur on Madeira and the Azores
(Fairfield et al. 2004). At least one species of Aichryson can be found on each of the
seven main Canary Islands. The centre of diversity is La Palma and only one species
is
found
on
Lanzarote.
Three
species
are
single-island
endemics,
Ai.
bethencourtianum on Fuerteventura, Ai. palmense on La Palma, and Ai.
porphyrogennetos on Gran Canaria. Fuerteventura and Lanzarote were the first
colonized by species of Aichryson, and Madeira was subsequently populated from
these two easternmost islands. Only once again this colonization pattern was found
(genus Crambe; Francisco-Ortega et al. 2002) and contrasts biogeographic
implications of other genera where a close biogeographic affinity between Madeira
and the five western Canary Islands was detected (Fairfield et al. 2004).
Up to 13 species are known for the genus Monanthes Haw.. They are classified into
three sections: Monanthes, Monanthoidea, and Sedoidea (Nyffeler 1992, Mes et al.
1997). They show a high level of different growth-forms comprising the dwarf annual
herb M. icterica, perennial branched or unbranched herbaceous rosettes, and small
branched shrublets. The large nectariferous scales are the most characteristic
feature of Monanthes. Being the showiest part of the flowers, they take over the
attracting function. Petals are very small, narrow, and decurved. Flowers are in
general 6-8merous but the number of flower organs is not constant and can even
vary within a single plant. Flower color is greenish or brownish and often variously
variegated with red (Nyffeler 1992). Pollinators are flies that are attracted by
segregated nectar and an intense musty scent in the evening hours (Nyffeler 1992).
Species are self-compatible although temporal and spatial separation of male and
female organs prevents self-pollination (Nyffeler 1992, 1995).
The species are distributed on all Canary Islands and the Salvage Islands with
highest diversity on Tenerife. Monanthes anagensis and M. minima are single island
endemics and only M. laxiflora and M. polyphylla are represented on more than two
islands (Nyffeler 1992). Species of Monanthes colonize rather mesic habitats with a
regular water supply. They grow in crevices of rocks and cliffs protected from direct
10
Introduction
sunlight in north- or northeast-exposed locations. Thus, in physiological adaptation
the C3 gas-exchange prevails even if weak CAM activity is also known (Nyffeler
1992). In addition, the characteristic bladder cell-idioblasts at the margin of leaves
are probably important for water storage (Nyffeler 1992, 1995).
Because different species with partly overlapping flowering times are often found in
sympatry, hybridization is frequent and indicates lack of genetic isolation. Hybrids
show intermediate habits, are restricted in their distribution, and usually occur only in
small numbers (Nyffeler 1995). An exception is M. icterica for which no hybridization
event is known (Mes et al. 1997). Monanthes muralis is of hybrid origin and the only
known allotetraploid species (Mes et al. 1997). Further tetraploid species such as M.
laxiflora, M. pallens, and M. polyphylla are known and M. anagensis is most probably
hexaploid (2n = 108; Mes et al. 1997). In general, chromosomes are very small and
the only exception is M. icterica which has large chromosomes and a chromosome
base number of x = 10 compared to x = 18 for all other Monanthes species (Nyffeler
1992, Mes et al. 1997).
Based on the wide range of growth-forms, flower morphology, and habitats, the MCS
were traditionally seen as an excellent example of an adaptive radiation. Lems
(1960) even compared the species of Aeonium with the Darwin finches. However,
molecular markers reveal a different picture. Inter-island colonization between similar
ecological zones – together with adaptation and hybridization – is the main force
driving speciation in the MCS (Jorgensen and Frydenberg 1999, Mort et al. 2002).
Inter-island colonization is also known for other Macaronesian genera like
Argyranthemum, Bystropogon, Crambe, and Sonchus (Francisco-Ortega et al. 1996,
Kim et al. 1996, Juan et al. 2000, Francisco-Ortega et al. 2001, 2002, Trusty et al.
2005, Sanmartín et al. 2008). Exceptions to this rule are the two genera Micromeria
(Meimberg et al. 2006) and Sideritis (Barber et al. 2000) in which adaptive radiation
is the main explanation for speciation. For lineages of the Pacific archipelagos
adaptive radiation is a common phenomenon and driving force for speciation with the
outstanding example of the Hawaiian silversword alliance (HSA; Baldwin and
Robichaux 1995, see also Barber et al. 2000).
Phylogenetic studies were done in many island species and radiations to
understand relationships among and within genera or species, define their origin,
11
Introduction
reveal phylogeographic and biogeographic patterns, and gain insight into speciation
processes. In particular molecular markers have broadened the knowledge and shed
new light on processes influencing radiation events (for review, e.g., Baldwin et al.
1998 or Emerson 2002).
Resolving relationships in recently diverged taxa is a huge problem in molecular
systematics (Syring et al. 2005). Most common molecular markers, such as nrITS
and cpDNA, are thought to be selectively nearly neutral and evolve slowly relative to
speciation in radiations (Baldwin et al. 1998, Sang 2002, Small et al. 2004, Whittall et
al. 2006). Thus, phylogenetic relationships were highly unresolved due to low genetic
divergence and a lack of fixation of synapomorphic mutations (Mes et al. 1996,
Baldwin et al. 1998). Subsequently, there is an ongoing debate that changes in
coding genes could verify the diversity of radiated lineages and provide valuable
tools for phylogenetic reconstructions (Baldwin et al. 1998, Sang 2002, Small et al.
2004).
Nuclear DNA (nrITS) and cpDNA markers are widely taken for phylogenetic
reconstructions. Universal primers are available which can be used over a wide
range of taxa (White et al. 1990, Taberlet et al. 1991, Blattner 1999). High copy
numbers facilitate amplification and sequencing can be done without prior cloning.
Low-copy nuclear genes are not extensively used until now (Bailey and Doyle
1999). Their main disadvantages are the development of species, gene or copy
specific primers, cloning and intensive sequencing (Sang 2002, Small et al. 2004).
Numerous nuclear coding genes exist in gene families and as multiple copies due to
gene or genome duplications (Bailey and Doyle 1999, Small et al. 2004). Speciation
events result in orthologous gene copies and duplication of genes within one species
in paralogous copies (Litt and Irish 2003). Thus, orthologs and paralogs are
frequently distributed in species and need to be distinguished since phylogenetic
analyses depend strictly on comparisons of orthologous gene copies to deduce
robust and correct relationships (Bailey and Doyle 1999, Litt and Irish 2003).
Advantages of nuclear coding genes are that they provide a nearly unlimited source
of additional, independent, unlinked, and bi-parentally inherited phylogenetic
information (Sang 2002, Small et al. 2004, Syring et al. 2005). They are alternative
markers to explore hybridization and polyploidization and to resolve contrasting
signals between nrITS and cpDNA data (Bailey and Doyle 1999, Small et al. 2004).
12
Introduction
In most cases, nuclear coding genes show higher evolutionary rates than neutral
evolving DNA regions and are therefore of special interest to resolve relationships at
low taxonomic levels and within radiations (reviewed in Small et al. 2004). In addition,
nuclear coding genes can be divided into four different parts that evolve differently.
First the 5´-untranslated region (UTR) which comprises conserved promoter
elements that are responsible for gene regulation. Highly variable parts may be used
for phylogenetic reconstructions (Small et al. 2004). Second the conserved exon
regions which contain the protein coding information. Within the 1st and 2nd codon
positions a nucleotide change results in a nonsynonymous amino acid replacement.
Nucleotide changes at the synonymous 3rd codon position result in the same amino
acid and the nucleotides at this position typically diverge at similar rates as noncoding regions (Nei and Gojobori 1986, Purugganan et al. 1995, Nei and Kumar
2000, Small et al. 2004). Alignments of exon sequences are in most cases easy,
especially at the amino acid level, even between genera or families. Thirdly and in
contrast, introns are much more variable. This is particularly true for variations at the
sequence level whereas the length of introns is often important for correct splicing. In
several cases introns also contain important regulatory elements that are conserved
(Small et al. 2004). Aligning intron regions is not trivial and in most cases only
possible between closely related species (see, e.g., Fortune et al. 2007 or Zhang et
al. 2008). The last region, the 3´-UTR, controls mRNA processing and the poly-A tail
and is highly variable even among species of the same genus (Small et al. 2004,
Whittall et al. 2006).
Molecular marker studies have surprisingly revealed that the high morphological
divergence found in radiated lineages does not correlate with the molecular diversity
of these species (see, e.g., Baldwin et al. 1998 or Purugganan and Robichaux 2005).
King and Wilson (1975) postulated that changes in regulatory genes rather than
changes in structural genes are the main factor for morphological variation of
species. Structural genes encode proteins that directly fulfill their task in the
organism. On the other hand, regulatory genes encode transcription factors that
regulate the expression of other genes and therefore play a central role in eukaryotic
development (Purugganan and Robichaux 2005). For example, in the adaptive
radiation of the HSA accelerated substitution rates for homologs of the regulatory
genes APETALA1 (ASAP1) and APETALA3 (ASAP3) compared to the structural
13
Introduction
gene CHLOROPHYLL A/B BINDING PROTEIN9 (ASCAB9) could be detected
(Barrier et al. 2001).
Contributing to this ongoing debate on the emphasis of these two gene classes, the
radiation of the MCS was studied. Physiological as well as morphological differences
within the MCS indicate that several genes may have had impact on the speciation.
Physiological differences are well known in adaptation to dry habitats and a limited
availability of water. In response, CO2 fixation is shifted from C3 to CAM fixation in the
MCS species (Lösch 1990). Here, fixation of CO2 results first in phosphoenolpyruvate
(PEP) that is converted to malic acid and stored in the vacuole during night. During
daytime malate is decarboxylated and CO2 re-assimilated via ribulose-1,5bisphosphate carboxylase (RUBISCO). The fixation is characterized by changing pHvalues during day and night and triggered by the enzyme PEP carboxylase (PEPC;
Cushman and Bohnert 1999). PEPC is a multifunctional enzyme and ubiquitous in
the plant kingdom with different isoforms and tissue-specific or specific physiological
roles. This structural gene is encoded by a small multigene family and several gene
copies per taxon are known (Gehrig et al. 1995). Protein alignments of the PEPC
coding genes show highly conserved motifs. It is likely that these encode domains
which are involved in the activity and regulation of the enzyme (Gehrig et al. 1995,
Chollet 1996, Cushman and Bohnert 1999).
Considerable differences in activity and strength of CAM and C3 for the species of the
MCS were reported by Tenhunen et al. (1982). Lösch (1990) classified the species in
strong, weak, and intermediate CAM and C3 fixation types. Pilon-Smits et al. (1992)
and Mort et al. (2007) refined these results and concentrated mainly on the impact of
CAM in the evolution of the MCS. Thus, the structural gene which encodes for PEPC,
afterwards named MCS_PEPC, is a good candidate for playing an important role in
the speciation process.
Other candidate genes with a high impact on speciation might be found involved in
reproduction. Mutations in genes that regulate floral color, reward or flower
architecture can accelerate the diversification process and cause rapid isolation by
influencing differential pollinator visitation (e.g., Baker and Baker 1983, Hoballah et
al. 2007, Whittall and Hodges 2007).
14
Introduction
Flower development is controlled by a large number of homeotic genes. Since the
identity of floral organs strictly depends on the activity of these genes they may have
an extraordinary influence in speciation (Theissen 2005, Erbar 2007). Floral
homeotic genes are regulatory genes and almost all of them belong to the MADSbox gene family (Becker et al. 2000, Erbar 2007). As the name indicates, the MADS
(MCM1 from budding yeast, AGAMOUS from Arabidopsis, DEFICIENS from
snapdragon, and SRF from human) gene family exists in animals, fungi, and plants
and control diverse developmental processes (Becker et al. 2000). In plants, most
MADS-box genes display floral-specific expression. However, some are expressed in
vegetative tissues and control, e.g., flowering time, inflorescence development and
structure, leaf development, or determine cell specification (Purugganan et al. 1995,
Baum 1998, Vergara-Silva et al. 2000, Saedler et al. 2001).
Plants possess type II MADS-box genes. The encoded proteins are of ~260 amino
acids and characterized by the MIKC structure comprising the four conserved
sequence regions: MADS-box domain, short Intervening region (I-region), Keratinlike domain (K-domain), and C-terminal region (Purugganan et al. 1995, Kramer and
Hall 2005; fig. 3).
Fig. 3: MIKC-like structure of the plant MADS-box proteins (adapted after Purugganan et al. 1995).
The highly conserved MADS-box domain comprises ~57 amino acids and may act as
a sequence specific DNA-binding motif (Coen and Meyerowitz 1991, Saedler et al.
2001). The I-region, which separates the MADS- and K-domain, shows considerable
sequence variability (Purugganan et al. 1995, Saedler et al. 2001, Kramer and Hall
2005). According to Becker et al. (2000) this region is important for selective
formation of DNA-binding heterodimers. The K-domain comprises ~70 conserved
amino acids and might mediate protein-protein interactions (Coen and Meyerowitz
1991, Vergara-Silva et al. 2000, Saedler et al. 2001). The C-terminal region is most
variable in sequence and length. It encodes a putative transactivation domain which
is involved in the formation of heterodimers (Kramer et al. 1998, Becker et al. 2000,
Vergara-Silva et al. 2000). MADS-box, I-region and the first part of the K-domain are
known as the core region and are necessary for DNA-binding and dimerization
15
Introduction
activity. The non-core region comprises the other part of the K-box and the Cterminal region and is important for strengthening the dimerization activities (Zhang
et al. 2008).
MADS-box genes were first characterized in the model species Arabidopsis thaliana
and Antirrhinum majus (for reference see, e.g., Erbar 2007) and defined as early and
late acting genes in plant development (Purugganan et al. 1995). Multiple gene
duplications have occurred and suggest that gene diversification via gene duplication
plays an important role in this gene family (Aagaard et al. 2005 and references
therein). At least three monophyletic MADS-box gene groups could be defined whose
members share similar expression patterns and functional roles. Many species
possess
more
than
one
locus
from
the
three
groups
AGAMOUS,
APETALA3/PISTILLATA, and APETALA1/AGL9 (Purugganan et al. 1995). The
AGAMOUS-group consists of 10 genes such as AGAMOUS and PLENA which are
inferred in the development of stamens and carpels (Kramer and Hall 2005). The
APETALA3/PISTILLATA-group
comprises
10
genes
that
are
important
for
development of petals and stamens. The group can be subdivided into two distinct
monophyletic subclades: the first comprises the orthologous genes APETALA3 (AP3,
from A. thaliana) and DEFICIENS (DEF, A. majus); the second PISTILLATA (PI, A.
thaliana) and GLOBOSA (GLO, A. majus). Within one species AP3 and PI as well as
DEF and GLO are paralogs and both are required for gene activity. They form
heterodimers consisting of AP3/PI (for A. thaliana) and DEF/GLO (A. majus) that bind
to their own promoters to set-up self-regulation (Purugganan et al. 1995, Kramer et
al. 1998, Saedler et al. 2001). Homologs of AP3 also show an influence on the size
of petals and stamens (Juenger et al. 2000, Lawton-Rauh et al. 2000). The third
group, APETALA1/AGL9, contains 14 genes in three distinct monophyletic gene
clades. The first clade, named APETALA1, consists of, e.g., APETALA1 (AP1),
SQUAMOSA (SQUA), CAULIFLOWER (CAL) or AGL8. The second clade AGL9
comprises AGL9, TM5, FBP2, while the third clade AGL6/AGL13 is represented by
AGL6 and AGL13. Some of these genes control floral meristem identity (AP1 and
CAL) whereas others (e.g., AP1) are important actors in floral organ identity of sepals
and petals (Purugganan et al. 1995, Berbel et al. 2001). Furthermore, studies of
quantitative trait loci (QTL) reported links between the AP1 locus and flowering time
as well as inflorescence branching patterns (Mandel and Yanofsky 1995).
16
Introduction
Mechanisms controlling flower development and floral organ identity are highly
conserved in evolution (Coen and Meyerowitz 1991). Flowers are organized in four
concentric whorls of organs. The two outermost whorls comprise sepals and petals
whereas the inner whorls consist of stamen and carpels as reproductive organs
(Coen and Meyerowitz 1991). Floral architecture is determined by the overlapping
activities of regulatory MADS-box genes and summarized in the ABC(DE) model
(Coen and Meyerowitz 1991, Theissen et al. 2000, Theissen 2001, Lohmann and
Weigel 2002, Erbar 2007; fig. 4).
A-function genes alone determine the organ identity of sepals whereas A- and Bfunction genes together determine petals. The B- and C-function genes are important
for stamen organ identity while the C-function genes alone are responsible for
carpels (Coen and Meyerowitz 1991). The classic ABC model was enlarged to the
ABCDE model in 2001 when genes with D- and E-function were described (VergaraSilva et al. 2000, Theissen 2001). D-function genes are important for the origin of
ovules whereas E-function genes are crucial co-factors for the ABCD-function genes
(Kramer and Hall 2005; fig. 4).
Fig. 4: The extended ABCDE model adapted after Erbar (2007).
In the present study, homologs of the A-function gene AP1 and the B-function gene
AP3, here afterwards named MCS_AP1 and MCS_AP3, were selected to analyze
their influence in the evolution of the MCS.
Since the selected genes exist in multigene families and the studied species are
polyploid (e.g., Mes 1995, Lawton-Rauh 2003), gene duplications are expected to
occur frequently in the MCS.
17
Introduction
Gene duplications may provide the raw material for evolution (Ohno 1970, Baum
1998, Lynch and Conery 2000, Zhang 2003). They are common, frequent, and
ongoing in organisms and without them, the plasticity of a genome or species in
adaptation to changing environments would be limited (Lynch et al. 2001, LawtonRauh et al. 2003, Zhang 2003). They may arise via unequal crossing over,
retroposition, or chromosome as well as genome duplications (Zhang 2003). Unequal
crossing over usually generates tandem gene duplications. The duplicated region
contains parts of a gene, an entire gene or several genes including the introns
(Zhang 2003). If genes are duplicated by retroposition, messenger RNA (mRNA) is
retrotranscribed to complementary DNA (cDNA) and then inserted into the genome.
This process is connected with the loss of introns and regulatory sequences, the
presence of poly A tracts and of flanking short direct repeats. The duplicated gene
copies are usually unlinked to the original gene and often become pseudogenes
immediately because they lack necessary transcription elements like promoters and
regulatory sequences (Zhang 2003). Genome duplications are especially well known
for
plant
species.
70-80%
of
the
angiosperm
species
have
undergone
polyploidization at some point in their history (Moore and Purugganan 2005). The
consequence of polyploidization is the duplication of the whole genome and thus, all
genes are immediately doubled, which is in sharp contrast to single gene
duplications.
Several scenarios and consequences could be assumed for duplicated genes: 1)
nonfunctionalization, pseudogenization, and subsequently gene loss, 2) redundant
maintenance, 3) subfunctionalization, and 4) neofunctionalization (Lynch and Conery
2000). According to Zhang (2003) many duplicated genes get lost. Furthermore,
since gene duplications lead to functional redundancy, it is often not advantageous to
have two identical copies (Baum 1998, Zhang 2003). It is likely that one gene copy
acquires mutations that disrupt the structure and function of the copy which gradually
becomes a pseudogene, whereas the other copy is involved in the proper function of
the gene. Pseudogenization is believed to occur in the first million years after
duplication if there are no other evolutionary processes acting on the duplicated gene
copies (Lynch and Conery 2000, Zhang 2003).
If the presence of duplicates in the genome is beneficial gene conversion via
concerted evolution or strong purifying selection can prevent duplicates from
diverging or getting lost. The maintenance of two gene copies becomes more likely
18
Introduction
when duplicates differ in some aspects of their functions (Duarte et al. 2006). This
could be achieved through subfunctionalization where each gene copy takes part in
the function of their ancestor gene. One important form, which seems to be a rule
rather than an exception, is division of gene expression after duplication like for
AP1/CAL/SQUA (see Duarte et al. 2006 or Shan et al. 2007). Another form can occur
at the protein level where partitioning the tasks of the ancestral gene takes place and
each copy is responsible for a unique set of subfunctions (Lynch et al. 2001, Zhang
2003). An example would be the AP3/PI homologs (Kramer et al. 1998).
One of the most important outcomes of gene duplication may be the origin of a novel
function by relaxed purifying selection or positive selection (Zhang 2003). Relaxed
purifying selection allows random mutations to become fixed in one gene copy and in
a changing environment might induce an altered gene function. In contrast, if positive
selection is involved in the development of a new gene function, accelerated fixation
of advantageous mutation may occur (Zhang 2003). Moore and Purugganan (2003)
showed that positive selection plays a key role in preserving gene copies and can act
at early stages to maintain duplicates.
Selection pressure acting on genes might be different and may vary between
retained orthologs and paralogs (see, e.g., Small and Wendel 2002 or Wang et al.
2007). In general, selection pressure acting on a gene could be distinguished into
neutral, purifying, and positive selection and base on nucleotide differences.
Nucleotide differences in the coding region have two different consequences. Either
nucleotide alteration resulted in the same amino acid in the translated protein
(synonymous or silent substitution) or in amino acid replacements (nonsynonymous
substitutions). The latter case could have dramatic effects on the deduced proteins
depending if very similar or quite different amino acids replace the original one
(Hughes et al. 2000).
Based on the number of synonymous and nonsynonymous substitutions selection
can be calculated. The selection pressure (ω) is defined as the ratio of Ka/Ks, where
Ka denotes the pairwise nonsynonymous substitution per nonsynonymous site and
Ks represents the number of synonymous substitution per synonymous site (Hurst
2002). If the ratio is nearly 1, neutral selection is expected indicating that chance
alone determines whether a mutation will become fixed or not. Additionally, the
likelihood of fixation in nonsynonymous substitutions is as high as those in
19
Introduction
synonymous substitutions. When the ratio of Ka/Ks is less than 1, purifying selection
is assumed that means selection against nonsynonymous nucleotide substitutions.
Selection eliminates deleterious mutations and maintains the function of the protein.
If the ratio of Ka/Ks is larger than 1, positive selection occurs where nucleotide
substitutions are favored, fixed and changes in the amino acid composition lead to
changes in the protein (Nekrutenko et al. 2001, Hurst 2002). In addition, if the value
of Ka is significantly higher than that of Ks, the most probable assumption is that the
gene has undergone adaptive evolution (Eyre-Walker 2006). So far, positive
selection has been detected for genes involved in sexual reproduction (e.g., gamete
recognition genes or self recognition genes in plants), for host-parasite interaction
genes (e.g., plant resistance genes (R-genes), plant chitinases), for genes which
encode for enzymes involved in energy metabolism (e.g., the pancreatic
ribonuclease genes in leaf-eating monkeys), and for genes involved in adaptation to
specific environments (e.g., regulatory genes involved in plant morphology; Bishop et
al. 2000, Ford 2002, Barrier et al. 2003, Zhang 2003). Thus, the influence of genes,
gene classes or gene copies in the speciation process could be estimated by
calculating the selection pressure acting on a gene, gene region or gene copy.
Summarizing these paragraphs the following questions arise:
1.) Do the studied low-copy nuclear genes reflect the species relationships as
inferred from other markers and are thus valuable tools to deduce and reconstruct
phylogenetic relationships of the MCS and Sedum species?
2.) Can gene duplications and therefore different functional or orthologous and
paralogous gene copies be detected for the MCS and/or Sedum species?
3.) If so, do gene copies evolve with same or different evolutionary rates and does
the selection pressure acting on different gene copies vary?
4.) Do the selected regulatory or structural genes have different impact on the
speciation process and what might have triggered the evolution of the MCS species?
The aim of this thesis is to find answers to these questions and to elucidate the
difficult phylogenetic relationships within the MCS.
20
Materials and Methods
2. Materials and Methods
2.1. Study species
In the following section the studied species will be shortly introduced. After the
species name sections referring to Lems (1960), Liu (1989), and Mes (1995), and the
clade number corresponding to the combined cpDNA/nrITS phylogram of Mort et al.
(2002) are mentioned for the Aeonium species, if possible (compare table 15 and fig.
2). Sections for the Monanthes species refer to Nyffeler (1992). For a better overview
the most important characters are given in table 1. Information for the species are
summarized from Lems (1960), Maire (1976), Kull (1982), Liu (1989), Lösch (1990),
Nyffeler (1992), Hohenester and Welss (1993), Nyffeler (1995), Mes et al. (1997),
Schönfelder and Schönfelder (1997), Mort et al. (2002), and Fairfield et al. (2004).
Aeonium aureum (C.Sm. ex Hornem.) T.Mes (sect. Greenovia [Mes], clade 2) is a
perennial terrestrial subshrub, 30-45 cm high and the most common species of the
former genus Greenovia. Plants form rosettes and always daughter rosettes, leaves
are glaucous and glabrous, and stems are unbranched with a lax branched
inflorescence. Flowering time is between March and April and the golden yellow
flowers are 20-35merous without nectaries. The species is endemic to the Canary
Islands and grows on occasionally moist rocks, roofs, and on rocky surfaces in the
pine region at 400-2000 m elevation.
Aeonium canariense Webb & Berthel. (sect. Canariensia, Patinaria, Canariensia,
clade 1) is a perennial terrestrial pilose herb which forms large rosettes. Stems are
unbranched and up to 45 cm high, and the inflorescence up to 70 cm. Petals, 8-10,
are very pale yellow-green to nearly white; flowers produce nectar and bloom
between April and August. The species is found in the laurel forest on rocks, soil
banks, and cliffs in fairly dry habitats up to 1300 m in the north of Tenerife.
Aeonium cuneatum Webb & Berthel. (sect. Canariensia, Patinaria, Canariensia) is
a typical representative of the Canarian laurel forest and is morphologically similar to
A. canariense. It is a perennial terrestrial, but sometimes also epiphytic, unbranched
herb which forms rosettes. Leaves are glabrous and glaucous. The inflorescence is
18-60 cm long; flowers are 8-9merous, bright or golden yellow and segregate nectar.
21
Materials and Methods
The species blooms between April and June and colonizes fairly moist habitats at
elevations of 500-950 m. It is found on rocks and soil banks among bushes and
occasionally on trees in the eastern and western part of Tenerife in the Anaga and
Teno mountains.
Aeonium goochiae Webb & Berthel. (sect. Goochiae, Petrothamnium, Goochiae,
clade 2) is a perennial terrestrial viscid subshrub with rosettes and stems up to 40
cm, which are densely branched. Flowers are 7-8merous and the flower color is very
pale yellow, rose or nearly white with a central portion of pink. The inflorescence
comprises 10-45 flowers, which produce nectar and bloom between February and
June. The plant is common on rocks, walls, and cliffs usually in the shadow of trees
or rocks in fairly moist habitats. It is an endemic species of the humid north coast of
La Palma at 100-700 m elevation.
Aeonium nobile (Praeger) Praeger (sect. Megalonium, Megalonium, Leuconium,
clade 4) is a monocarpic perennial terrestrial subshrub with stems up to 60 cm and
large succulent leaves. It is rarely branching and the large inflorescence is flat-topped
to broadly dome-shaped. Its dark red flowers are unique and the result of numerous
reddish stripes on the 7-9 pale yellow petals. Flowers produce nectar and bloom
between March and July. The species is endemic to La Palma where it can be found
on oldest rock formations in dry slopes, soil banks, and cliffs in soil pockets and
crevices, from the sea level up to 800 m.
Aeonium rubrolineatum Svent. (sect. Holochrysa, Aeonium, Aeonium, clade 3) is a
perennial terrestrial subshrub. Its stems are erect, branches often in groups, and the
inflorescence is dense. Flowers are 9-11merous and yellow to pale yellow with
reddish variegations caused by reddish veins or reddish bases and margins. Nectar
is produced; plants bloom between May and November and loose all leaves during
the flowering period. It colonizes soil banks and cliffs from 800-1200 m and is
especially common in the SW sector of La Gomera.
Aeonium saundersii Bolle (sect. Goochiae, Petrothamnium, Petrothamnium, clade
2) is a perennial terrestrial subshrub and unique in its balsam odor. Stems are up to
25 cm high and inflorescences have 5-70 flowers. These are 12-16merous, pale
22
Materials and Methods
yellow and do not segregate nectar. Flowering time is between April and June and
the species is common on vertical rocks both in sun and shade. It is found in cervices
where it grows rapidly, branching out and undergo a summer rest period with leaves
forming rounded bud-like structures. It is endemic to La Gomera where it is found in
the east at 150-800 m elevation.
Aeonium smithii Webb & Berthel. (sect. Goochiae, Chrysocome, Chrysocome) is a
perennial terrestrial unbranched herb with stems up to 60 cm. Plants are hirsute with
multi-cellular trichomes. Number of sepals and petals is 8-12, mostly 10. The yellow
flowers do not produce nectar and bloom between March and October. The species
is common on rocks and cliffs in the pine forest zone and Cañadas between 1502150 m. It is found in the south and east of Tenerife.
Aichryson laxum (Haw.) Bramwell is an annual or biennial herb, pilose and often
reddish. Stems are 15-30 cm high and regularly branched. Flowers are 8-12merous
and bright yellow. The flowering time of the species is between March and June. The
species colonizes moist crevices and walls and is occasionally epiphytic and endemic
to the Canary Islands.
Aichryson pachycaulon Bolle is an annual or biennial herb. Stems are thick, erect,
up to 65 cm, and plants are glabrous. Flowers are pale to bright yellow, 7-8merous
and bloom between October and April. Aichryson pachycaulon is classified into five
distinct species or subspecies which cluster into three different clades (Fairfield et al.
2004). They show a range of morphological and cytological variation although they
are generally characterized by the lack of pubescence. Each subspecies is a single
island endemic and Ai. pachycaulon subsp. immaculatum occurs on Tenerife. The
species colonizes mesic habitats like rocks in the laurel forest of the Canary Islands.
Monanthes anagensis Praeger (sect. Sedoidea) is a slightly woody, diffusely
branched shrublet up to 30 cm high. Leaves are ordered alternative and never
covered with wax. The terminal inflorescence is regularly branched with 3-18 flowers.
Flowers are 6-8merous, pale yellow, occasionally with brown-red stripes and with
nectaries. The species is restricted to higher elevations of the Anaga Mountains on
Tenerife from 600-900 m and locally common. It mostly occurs on moist rocks and
23
Materials and Methods
cliffs and sometimes in open and more arid places. It is a characteristic taxon of the
laurel forest zone but partly occurring in the lowland xerophytic zone as well.
Monanthes icterica (Webb ex Bolle) Praeger (sect. Monanthoidea) is a dwarf
annual herb, reaches up to 6 cm height, is rather unbranched with alternate leaves
and glabrous. It is the only known annual species of the otherwise perennial
Monanthes. The inflorescence is terminal, regularly branched with 3-7 flowers. These
are pale yellow, often slightly brown-red striped, and 6-7merous with nectaries. The
species is distributed on Tenerife and La Gomera on ledges and in crevices of rocks
and cliffs from 100-900 m.
The habit of M. icterica resembles species of the genus Aichryson but the enlarged
nectaries, the entirely glabrous axes and leaves, and the bladder idioblasts place this
species clearly in Monanthes.
Sedum caeruleum L. is an herbaceous glabrous green plant with spots of red. The
sepals often have a black spot at the base and the flowers are 5-9merous, normally
7merous. The single terminal inflorescence is not compact; flowers are star-shaped
and azure or rarely white. They have very small white nectaries and flowers between
March and June. Plants colonize stone fields, boulders, clearance with stones, and
sometime banks of small rivers from sea level up to middle mountains. It is
distributed in well watered regions in Southern Europe especially in Italy, Corsica,
Sardinia, Sicily, and Malta.
Sedum jaccardianum Maire is an herbaceous sticky plant. The petals are outside
citron yellow and inside yellow with an orange zone in the upper part. The number of
floral organs range between 6 and 10 and nectaries are very small and whitish. The
species flowers between May and July and colonizes calcareous stones from 1600 m
up to 2800 m. Sedum jaccardianum is endemic in Northern Africa.
Sedum modestum Ball is a very small herb with rosettes. Its leaves are glabrous
but hairs could be found at the back of the petals. Petals are golden yellow and later
white with purple spots. Flowers are star-like in shape, 5-7merous and have
nectaries. The inflorescence is terminal with one group of flowers and not compact.
Flowering time is between April and June. The species colonizes stone fields, soil
24
Materials and Methods
filled crevices, and stumps from the sea level up to 2200 m in semiarid and well
watered areas and is endemic in Northern Africa.
Sedum pubescens Vahl is an herbaceous sticky plant. Flowers are bright yellow
and have a star-like shape, number of floral organs is 5-6 and flowers have nectaries.
The inflorescence is large, not compact, with only one flower at the end and flowering
time is between May and July. Plants grow in forests, in low open scrublands with
many evergreen shrubs (Garrigue), in understorey, field of stones, and in crevices of
calcareous or silicate stones from 0-1300 m. The species occurs in semiarid and well
watered regions of Northern Africa and Southern Europe.
Sedum surculosum Coss. is an herbaceous perennial, glabrous plant with sessile
loose flat rosettes. Flowers are 5-7merous and have a star-like shape. They are pale
yellow with slight red-brown stripes and have small nectaries. Species flowers
between July and August and colonizes damp rocks and cliffs. Two different species
varieties can occur either on granitic and porphyritic rocks or on limestone (Nyffeler
1992). Sedum surculosum can be found along streams of high mountains from 24003800 m. The species is endemic to Northern Africa and distributed in Morocco in the
Great Atlas and Anti-Atlas.
The taxonomic history of S. surculosum is quite interesting. In 1873 J. Ball described
the species as Monanthes atlantica but in the same year a plant from the Moroccan
Atlas Mountains was described as S. surculosum by E. Cosson. Berger (1930)
retained it in the genus Monanthes whereas Praeger (1932) excluded it because of
its broad yellow petals, which are more Sedum or Aichryson-like, and S.
jaccardianum seems to connect S. surculosum with the Sedum species (Nyffeler
1992). Nevertheless, it was often still included in the genus Monanthes because of its
large nectariferous scales and resemblances in habit and flower morphology (see
Nyffeler 1992 or Mes et al. 1997). However molecular data strongly support the
inclusion in the genus Sedum because it shares with S. jaccardianum a unique 70
base pair deletion in the cpDNA trnL-trnF intergenic spacer (Mes and ‘t Hart 1994).
25
Materials and Methods
Table 1: Overview of the main characteristics of the studied species.
taxon
distribution
A. aureum
GC,T,H
A. canariense
T
A. cuneatum
T
A. goochiae
P
A. nobile
A. rubrolineatum
P
G
A. saundersii
A. smithii
Ai. laxum
Ai. pachycaulon
M. anagensis
M. icterica
S. caeruleum
S. jaccardianum
S. modestum
G
T
T,GC, P,H,G
T,GC, P,F,G
T
G,T
S Europe
MO
MO
S. pubescens
N Africa, S
Europe
MO
S. surculosum
CO2 fixation
type
C3 - weak
CAM
C3 - weak
CAM
C3 - weak
CAM
C3 (CAM
possible)
strong CAM
C3 (CAM
possible)
C3
strong CAM
C3
C3
C3
C3
C3
C3
C3
floral
organs
28-32
flower color
habit
habitat
herbaceous
moist
herbaceous
fairly dry
herbaceous
fairly moist
7-8
golden
yellow
pale yellow
green
golden
yellow
rose
subshrub
moist
7-9
9-11
dark red
yellow
subshrub
subshrub
dry
dry
12-16
8-12
6-12
7-8
6-8
6-7
5-9
6-10
5-7
subshrub
herbaceous
herbaceous
herbaceous
shrublet
herbaceous
herbaceous
herbaceous
herbaceous
mesic
dry
moist
mesic
moist
mesic
mesic/moist
mesic/moist
mesic/moist
herbaceous
mesic/moist
herbaceous
moist
8-10
8-9
C3
5-6
pale yellow
yellow
bright yellow
pale yellow
yellow
pale yellow
azure, rose
citron yellow
golden
yellow
bright yellow
C3
5-7
yellow
2.2. Molecular methods
2.2.1. RNA material
Extracted RNA was provided by Dr. M. Thiv (SMNS, Stuttgart). A combination of
fresh bud material of different stages of several Aeonium species was collected in the
Botanical Garden of the University Zurich and in the City Succulent Collection Zurich.
RNA was extracted following the standard protocol for the Concert™ Plant RNA
Reagent Kit (Invitrogen) and used for subsequent cDNA syntheses using Superscript
II reverse transcriptase (Invitrogen).
2.2.2. DNA material
Genomic DNA was extracted using the DNeasy Plant Mini Extraction Kit (Qiagen).
Living plants for most of the studied species were collected on Tenerife and La
Gomera (MCS) or in Morocco (Sedum; table 16; appendix). Plants were cultivated at
the State Museum of Natural History Stuttgart where vouchers of the species are also
deposited. Material of A. goochiae and A. nobile was provided by Prof. Dr. R. Lösch
(University Düsseldorf), material of S. caeruleum by Mrs. Lübenau-Nestle, and S.
26
Materials and Methods
pubescens by the City Succulent Collection Zurich (table 16; appendix). Leaves used
for DNA extraction were harvested, grinded with liquid nitrogen and DNA
subsequently extracted following the instructions of the supplier.
2.2.3. Amplification of the low-copy nuclear genes
Homologs of the floral homeotic regulatory genes AP1 and AP3 as well as of the
structural gene PEPC were amplified comprising the following steps: 1) cDNA
synthesis and subsequently the use of degenerated primers to amplify the respective
gene region from cDNA using GoTaq polymerase (Promega). 2) The obtained
sequences were aligned and blasted to check the sequence identity by similarity
comparisons using the National Center for Biotechnology Information (NCBI)
database. If the amplified sequences were confirmed as homologs of the target
genes, 3) specific primers for each gene were deduced and used for genomic DNA
amplification with proofreading enzymes (Pfu: Fermentas or Promega; Phusion:
NEB). Proofreading enzymes correct incorporation of nucleotides in 3´ to 5´ direction
and therefore reducing the number of PCR mistakes during template elongation.
All PCR reactions were performed in a final volume of 10 µl and components as
indicated in table 2. PCR reactions were done in a T-Gradient cycler (Biometra,
Göttingen) or GeneAmp 9700 PCR System (PE Biosystems, Foster City, CA).
Amplified PCR products were separated on 1% ethidium bromide stained agarose
gels to prove size and quality and were subsequently cloned and sequenced as
described below.
MCS_PEPC
Direct amplification from genomic DNA was possible using the primers PEPC-F and
PEPC-R (table 17; Gehrig et al. 1995). Initial PCR reactions were done using GoTaq
(Promega) under the following conditions: initial treatment of 95°C for 3 min., 33
cycles of [95°C 45 sec., 52.5°C 50 sec., 72°C 2.5 min.] and a post treatment of 72°C
for 7 min. before cooling.
For final reactions annealing temperature and elongation time were optimized and
amplification from genomic DNA done with Pfu polymerase (Fermentas; table 2)
under the following conditions: initial treatment of 95°C for 3 min., 33 cycles of [95°C
50 sec., 53°C 50 sec., 72°C 5 min.] and a post treatment of 72°C for 15 min. before
cooling. For several species Pfu polymerase (Promega; table 2) was used and the
27
Materials and Methods
PCR was performed under the following conditions: initial treatment of 95°C for 3
min., 35 cycles of [95°C 45 sec., 52.5°C 50 sec., 72°C 5 min.] and a post treatment of
72°C for 15 min. before cooling.
For the Sedum species further primer optimizations were done based on conserved
regions of the MCS species. For PCR reactions Phusion (NEB; table 2) was used,
the primer combination PEPC_Sed_for_2/PEPC_Sed_rev (table 17), and reactions
were performed under the following conditions: initial treatment of 98°C for 30 sec.,
30 cycles of [98°C 10 sec., 64°C 30 sec., 72°C 2 min.] and a post treatment of 72°C
for 10 min. before cooling.
MCS_AP1
cDNA synthesis with a specific poly(T)-AP1 primer (Litt and Irish 2003; table 17) was
done following the instruction of the supplier (Invitrogen). For the first PCR primer
AP1MDS3 (Litt and Irish 2003; table 17) and AP1-noT (table 17) were used and
reactions were performed under the following conditions: initial treatment of 94°C for
3 min., 35 cycles of [94°C 50 sec., 50°C 50 sec., 72°C 2 min.] and a post treatment of
72°C for 10 min. before cooling. Amplified fragments were separated on a 1%
ethidium bromide stained agarose gel, excised from the gel, and purified with GFX
gel purification columns (GE Healthcare). Purified products were 1:25 diluted and
used for nested PCR with the primer combination AP1MDS2/SQUAR (Litt and Irish
2003; table 17) as followed: initial treatment of 95°C for 3 min., 33 cycles of [95°C 50
sec., 55°C 50 sec., 72°C 2 min.], followed by post treatment of 72°C for 5 min. before
cooling.
Finally the specific primers AP1-11F and AP1-704R (table 17) were used for genomic
DNA amplification with Pfu (Promega; table 2) and the PCR program: initial treatment
95°C 3 min., 30 cycles [95°C 1 min., 54°C 50 sec., 72°C 5 min.] and post treatment
of 72°C for 15 min. before cooling.
For the Sedum species further primer optimizations were done based on conserved
regions of the MCS species. Primer combination AP1_Sed_for/AP1_Sed_rev (table
17) was used, Phusion (NEB; table 2), and PCR reactions were performed under the
following conditions: initial treatment of 98°C for 30 sec., 30 cycles of [98°C 10 sec.,
64°C 30 sec., 72°C 2 min.] and a post treatment of 72°C for 10 min. before cooling.
28
Materials and Methods
MCS_AP3
Two approaches were followed to develop specific primers for AP3 homologs. For
the first one cDNA synthesis was done with random primers as described by the
supplier. Subsequently, first PCR reactions were performed with each time 1.1 µl
cDNA and the primer combination ATG3/AP3-polydT (Kramer pers. communication;
table 17) as followed: initial treatment of 95°C for 1 min., 5 cycles of [95°C 30 sec.,
60°C 3 sec. with ramp of -0.2°C/sec. to 50°C, 72°C 50 sec.] followed by 35 cycles of
[95°C 30 sec., 60°C 30 sec., 72°C 2 min.] and a post treatment of 72°C for 7 min.
before cooling. Amplified fragments were separated on a 1% ethidium bromide
stained agarose gel, excised from the gel between 500 and 1000 bp, and purified
with GFX gel purification columns (GE Healthcare). Purified products were 1:1 diluted
and nested PCR reactions with the primers MADS4 and AP3-polydT (Kramer pers.
communication; table 17) were performed as followed: initial treatment of 95°C for 3
min., 33 cycles of [95°C 50 sec., 53°C 50 sec., 72°C 2 min.], followed by post
treatment of 72°C for 5 min. before cooling.
Based on these sequences primers for the MCS species were improved and used in
the second approach. Here, cDNA synthesis was done with the specific AP3-polydT
primer (table 17) following the instruction of the supplier and subsequently the primer
combinations MADS4-Aeo/AP3-724R, AP3-11F/AP3-724R, AP3-351F/AP3-polydT,
and AP3-351F/AP3-724R (table 17) were tested using 1 µl diluted cDNA as
indicated: initial treatment of 95°C for 3 min., 33 cycles of [95°C 50 sec., 53°C 50
sec., 72°C 2 min.], followed by post treatment of 72°C for 5 min. before cooling. In
addition, another first PCR was performed with the primer combination ATG3/AP3noT2 (table 17) as followed: initial treatment of 95°C for 3 min., 33 cycles of [95°C 50
sec., gradient of 50°C-60°C 50 sec., 72°C 2 min.], followed by post treatment of 72°C
for 5 min. before cooling. Afterwards probes were 1:20 diluted and used for the
second PCR with a successful amplification for the primer combination AP3351F/AP3-724R (table 17) as followed: initial treatment of 95°C for 3 min., 33 cycles
of [95°C 50 sec., 55°C 50 sec., 72°C 2 min.], followed by post treatment of 72°C for 5
min. before cooling.
Finally primers were optimized for genomic DNA and the combination PI-F-Aeo/AP3724R (table 17) was used with Pfu (Promega; table 2) and the following PCR
program: initial treatment 95°C 3 min., 30 cycles [95°C 1 min., 54°C 50 sec., 72°C 5
min.] and post treatment of 72°C for 15 min. before cooling.
29
Materials and Methods
For Sedum further primer optimizations were done based on exon regions of the
MCS species. GoTaq (table 2) was used because amplification and/or subsequent
cloning failed using proofreading enzymes. The primer combinations PI-F-Aeo/AP31766R (comprises the MCS specific forward primer) as well as AP3-82F/AP3-1766R
(both Sedum specific; all table 17) were used and PCR cycling was performed under
the following conditions: initial treatment 95°C 3 min., 30 cycles [95°C 50 sec., 55°C
45 sec., 72°C 2 min.] and post treatment of 72°C for 10 min. before cooling.
Table 2: Overview of the components used in PCR reactions for the different polymerases.
GoTaq
Pfu (Promega)
Pfu (Fermentas)
(Promega)
Phusion
(NEB)
buffer
1x
1x
1x
1x
MgSO4
-
-
2 mM
-
dNTPs
0.2 mM each 0.4 mM each
0.2 mM each
0.2 mM each
primer each
0.325 µM
1.0 µM
1.0 µM
0.5 µM
polymerase
0.05 U/µl
0.032-0.048 U/µl
0.025 U/µl
0.02 U/µl
DNA
30-50 ng
10-30 ng
10-50 ng
30-80 ng
Since no information concerning the copy number was available for MCS and Sedum
species, one main question was if there are gene duplications within the MCS and/or
Sedum species, and if orthologous and/or paralogous gene copies could be
detected. Deduced primers were therefore specific for the gene region and species of
interest but not for potential gene copies and allow a screening for copy number in
the studied species groups.
2.2.4. Cloning of the low-copy nuclear genes
PCR products of the low-copy nuclear genes showed only a single band. However,
used primers were just gene specific whereas species of the MCS are polyploid and
the selected genes represented in multigene families. Thus, to prove and verify the
number of amplicons, PCR products were cloned using the pGEM-T Easy vector
system (Promega; T-cloning technique) or the Jet cloning kit (Fermentas; blunt-end
cloning technique). Instructions of the supplier were followed but only half of the
reaction mix containing buffer, vector, and ligase was used. PCR products resulting
from amplification with GoTaq were cloned directly using the pGEM-T Easy vector.
30
Materials and Methods
GoTaq creates A-overhangs at PCR products which directly ligate to the T-overhang
of the provided cloning vector. In contrast, PCR products of proofreading enzymes
miss A-overhangs. Using the pGEM-T Easy vector system additional A-addition was
done using GoTaq and following the instructions of the supplier, or PCR products
were cloned directly into the Jet blunt-end cloning vector. Ligation was done for 1h at
room temperature or over night at 4°C. The ligation reaction mix was transformed to
50 µl of competent cells (XL1 blue-cells) via heat-shock reaction for 60-90 sec. in a
42°C warm water bath or heat block and immediately returned on ice for 2 min.
Transformed cells were grown in 700-900 µl SOC medium for 1.5 h at 37°C with
shaking at ~300 rpm. 150-200 µl of the transformation culture were plated on
LB/ampicillin/IPTG/X-Gal plates and grown over night at 37°C. Positive clones were
selected by blue-white screening for the pGEM-T Easy vector system. Using Jet
cloning only positive clones grow on ampicillin plates.
2.2.5. Colony PCR
Cloning was followed by colony PCR. In average, eight white clones per gene and
taxon were selected randomly and amplified using the vector specific primers M13
(Promega; table 17) or Jet (Fermentas; table 17).
Reactions were performed in 20 µl final volume with one randomly selected white
colony using Pfu (Promega) or GoTaq (Promega) with components as indicated in
table 2, and the following program: initial treatment at 95°C for 3 min., 33 cycles of
[95°C 50 sec., 55°C 50 sec., 72°C 12 min.], followed by a final elongation of 20 min.
at 72°C before cooling. Using GoTaq (e.g., for MCS_AP3 of Sedum) an elongation
time of 3 min. was used and an annealing temperature of 57°C for the vector specific
primer Jet.
Screening for successful amplification of the insert was done on 1% ethidium
bromide stained agarose gels. Five successful amplified PCR products (equivalent to
clones or colonies) were chosen randomly, purified using Nucleofast 96 PCR plates
(Macherey-Nagel), and re-suspended in 20 µl TE buffer. Quantity and quality of
purified PCR products were checked on a 1% ethidium bromide stained agarose gel
and the amount of template product for the sequencing reaction estimated.
31
Materials and Methods
2.2.6. Amplification of nrITS and cpDNA regions
For the studied species the rDNA nrITS region and several cpDNA regions (matK,
trnL-trnF, and psbA-trnH spacer regions) were amplified using primers and conditions
as described by Mort et al. (2002). Forward and reverse sequences for each taxon
were aligned. Obtained consensus sequences were used to prove species identity
and to improve the nrITS alignment provided by Dr. M.E. Mort (University of Kansas)
for subsequent phylogenetic species reconstructions.
2.2.7. Sequencing
Cycle sequencing was performed in a T-gradient cycler (Biometra) with the following
cycling program: initial denaturation at 96°C for 2 min., 25 cycles of [96°C 10 sec.,
50°C 5 sec., 60°C 4 min.] with a ramp of 1°C/sec. and immediately cooled to 4°C. For
clones that were amplified with the primers Jet, a higher annealing temperature of
57°C was used while cycle sequencing to assure sequence quality.
Reactions were performed using 5 µl final volume with maximal 2.75 µl of the purified
PCR product, 0.5 µM primer, 0.83 µl of 5x sequencing-buffer, and 0.33 µl BigDye
Terminator (version 3.1., Applied Biosystem).
In addition to the vector specific primers M13 or Jet, several internal sequencing
primers (table 17) were deduced for the low-copy nuclear genes to generate
overlapping sequence regions. For the nrITS and cpDNA regions amplification
primers were used.
Sequencing products were purified on 96-well Sephadex (GE Healthcare) plates and
dried products were sent for sequencing. Sequencing was done on a capillary
sequencer (ABI PRISM; Applied Biosystems) in the Institute of Systematic Botany,
University of Zurich.
2.2.8. Sequence analysis
Obtained chromatograms of all sequences were checked and edited manually.
Consensus sequences for each clone were obtained based on overlapping sequence
regions using Sequencher 4.2.1. (Gene Codes).
Consensus sequences were aligned with help of Sequencher, manually improved,
and nucleotide substitutions were carefully checked by comparison with the
chromatogram of the original sequence.
32
Materials and Methods
Final alignments comprising all consensus sequences of respective clones were
done using Sequencher and manually improved in Se-Al V2.0a11 (Rambaut 1996)
and MEGA version 4 (Tamura et al. 2007).
2.2.9. Definition of the gene regions
Since amplification based on genomic DNA, coding and noncoding regions (exons
and introns) could be defined following the GT/AG rule for excisions of introns.
Comparisons were done with respective cDNA sequences of MCS species and/or
with coding sequences of Kalanchoe species that were obtained from the NCBI
database. Kalanchoe gracilis (AJ252946) and K. pinnata (AJ252919) were used for
MCS_PEPC; K. blossfeldiana (DQ479358) for MCS_AP3.
Three datasets were obtained for each gene: a full-length dataset including all
information of amplified templates (total amplified genomic DNA). Exclusion of exon
regions will result in the intron dataset whereas via excluding the intron partitions the
exon dataset is obtained. The exon alignment was carefully checked for the open
reading frame (ORF) and adapted to it by translation into amino acid alignments in
MEGA.
Furthermore, unambiguous alignments of intron sequences between genera are not
trivial. Therefore, the Sedum sequences were removed, arising gaps deleted, and
phylogenetic reconstructions done based on exclusive MCS intron regions.
2.2.10. Improvements of the datasets
Full-length nucleotide alignments for each specific taxon were compared and
identical sequences (clones) excluded from datasets to decrease calculation time.
Additionally, datasets were checked for PCR mediated chimeric sequences.
Occurrence of chimeric sequences is likely when similar templates are amplified
within a single PCR reaction. Testing was first done by careful comparison of taxon
specific sequences by eye, searching for recombination and break points. Afterwards
checking was done with help of the Bellerophon web server that uses a partial tree
building approach (Huber et al. 2004). Phylogenetic reconstructions are done with
sequence fragments before and after an assumed break point and the topologies
compared to infer incongruence that may indicate chimeric sequences. Here, the
whole dataset was compared using four different correction methods (HuberHugenholtz, Kimura, Jukes-Cantor, and no correction). Aligned datasets and an open
33
Materials and Methods
window size of 300 were used and only adapted for the MCS_AP1 dataset to an
open window size of 200. Full-length and exon datasets were checked and the
detected potential chimeric sequences again carefully proved by eye. Sequences
resulting with every method and after final proof as chimers were excluded.
Next to redundant and chimeric sequences, sequences with single base pair
deletions that lead to frameshift mutations and premature stop codons in the exon
alignment were excluded based on comparisons of the exon and protein alignments.
Each final alignment was converted into nexus and phylip formatted files for
phylogenetic reconstructions. MEGA formatted files were used for further analyses as
indicated below.
2.2.11. Partition Homogeneity Tests
Conflicting signals in datasets could be estimated using the Partition Homogeneity
Test (PHT) as implemented in PAUP* 4b10 (Swofford 2002). The reduced dataset of
each of the three gene regions was used to estimate conflicting evolutionary signals
concerning exon and intron regions. Datasets were changed by combining all exon
and intron regions together and define them as data partitions. 1000 replicates were
performed using the heuristic search modus. Starting trees were obtained by
stepwise addition with random addition of sequences and 100 replicates. Number of
trees hold in each step during random sequence addition was 1. Using TBR branchswapping no more than 10 trees were saved. If the resulted p-value is significant (p <
0.05) there is a conflicting signal in the dataset between exon and intron sequence
evolution. If the p-value is not significant no conflicting signal could be detected.
2.2.12. Phylogenetic reconstructions
The character based methods Bayesian Interference (BI) and Maximum likelihood
(ML) were used to reconstruct the phylogenetic history of the studied genes and
species. The advantage of both methods is the integration of a model of sequence
evolution. The model of sequence evolution was first calculated using Modeltest 3.7
and the model defined by the Akaike information criterion (AIC) was selected
(Posada and Crandall 1998). The obtained model was used to define the parameters
for each BI analysis using MrBayes 3.1.2 (Huelsenbeck and Ronquist 2001, Ronquist
and Huelsenbeck 2003) and for ML analyses using PHYML 3.0 (Guindon et al.
2005). Default settings (temperature 0.2, four chains, random starting tree, and
34
Materials and Methods
sampling every 100th generation) were used for BI. Only in one case, where after a
reasonable high number of generations the standard deviation of the split
frequencies of the four chains was not below 0.01, the analysis was stopped and
restarted with the lower temperature of 0.1 (see table 9). If the standard deviation of
the split frequencies for the chains was < 0.01 the analysis was finished. Based on
the number of generations, posterior probabilities (pp) as well as trees were
summarized using a burn-in of 30%.
For ML the web-based PHYML 3.0 program was used (Guindon and Gascuel 2003,
Guindon et al. 2005). Based on the obtained model of sequence evolution all
parameters needed for the ML analysis were specified, especially the model of
sequence evolution and the discrete gamma-model, if possible. As tree topology
search strategies NNI and SPR were used. A random BIONJ tree was generated as
starting point and in total five replicates performed. Bootstrap support (bs) was
obtained and the number of bootstrapped datasets range between 300 and 1000
depending on the computer resources and calculation time.
All trees were opened in TreeView 1.6.6 (Page 1996), rooted with the respective
outgroup and ordered. Posterior probabilities and bootstrap support are indicated at
the nodes and final trees were graphically improved in Adobe Illustrator 11.0.0.
2.2.13. Blast analyses
Obtained sequences were blasted against known sequences in the NCBI database
to confirm results deduced from phylograms. The megablast search algorithm (highly
similar sequences) was used and only if no results could be obtained the blastn
search algorithm (somewhat similar sequences).
2.2.14. Neighbor-joining reconstructions for homologs of PEPC, AP1, and AP3
To infer the genealogy over a broader sampling and to confirm orthologous and
paralogous copies for the studied genes and species, Neighbor-joining (NJ) analyses
were done. Therefore, sequences that showed highest sequence similarity and
randomly picked sequences from the NCBI database were added to the respective
alignments. Alignments were done and improved based on translated amino acids
and reverted nucleotide alignments subsequently used to infer the NJ phylogram.
Alignments, translations, and analyses were done in MEGA with the following
settings: all sites included, complete deletion of gaps, default model of “Maximum
35
Materials and Methods
Composite Likelihood”, and 1,000 re-sampled pseudoreplicates to obtain bootstrap
support.
2.2.15. Species phylogeny
To compare gene trees with the species phylogeny an existing nrITS dataset (M.E.
Mort, University of Kansas) was used and own sequences added. Improvements of
the given alignment were made and the resulting dataset used for a BI analysis.
The BI tree with branch lengths was furthermore used to construct an ultrametric tree
using r8s v.1.70, a program for reconstructing divergence times and absolute rates of
substitutions (Sanderson 2003). For this purpose a rate smoothing approach using
penalized likelihood (PL) with the TN algorithm as well as nonparametric rate
smoothing (NPRS) with Powell algorithm were used (Sanderson 1997). The age of
the oldest island (Fuerteventura = 20.7 My) was used to calibrate the node of the
Macaronesian taxa. The cross-validation procedure for PL was performed to
determine the appropriate smoothing levels and was set to S = 1. Both methods
resulted in highly similar ages. Therefore the ultrametric tree was constructed based
on NPRS and Powell algorithm.
2.2.16. Nucleotide differences, replacements, and amino acid substitutions
Next to estimate phylogenetic relationships, exon and deduced amino acid
alignments were used for further analyses. Based on exon phylograms orthologous
and paralogous gene copies were distinguished and this information used for all
further analyses.
Numbers of nucleotide differences were counted based on exon alignments and
numbers of replacements based on deduced protein alignments. Total numbers of
substitutions
between
all
species-specific
sequences
were
counted
which
corresponds mainly to paralogous gene copies (sequences between subclades). In a
second approach it was distinguished in orthologous gene sequences (sequences
within a subclade). Based on protein alignments amino acids were characterized and
proved if the replaced amino acid was very different to the original one concerning
charge, size, and further characteristics and if so, afterwards named and counted as
quite different amino acid (Stryer 1995).
36
Materials and Methods
2.2.17. Relative Rate Tests
Differences in the evolutionary rates of duplicated gene copies were estimated with
Tajima's relative rate test (RRT) as implemented in MEGA. Equality of evolutionary
rates between two paralogous sequences in reference to an outgroup sequence is
tested. P-values less than 0.05 are used to reject the null hypothesis of equal rates
between lineages. Thus, if the null hypothesis is rejected, molecular clock can be
rejected as well and sequences evolve differently.
2.2.18. Selection pressure
The selection pressure acting on the respective genes was estimated with three
different methods. In advance, a further reduction of identical species-specific exon
sequences was done. Furthermore, the Kalanchoe sequences were excluded from
the MCS_PEPC and MCS_AP3 datasets to calculate the selection pressure within
the studied species of MCS and Sedum.
First, Ka/Ks-values were estimated that define the selection pressure acting on a
gene with Ka/Ks = 1 indicating neutral selection, Ka/Ks < 1 purifying selection, and
Ka/Ks > 1 positive selection. Ka- and Ks-values were obtained by calculating pdistances with the Nei-Gojobori method and complete deletion of gaps as
implemented in MEGA. The Nei-Gojobori method calculates the numbers of
synonymous and nonsynonymous substitutions and the numbers of potentially
synonymous and nonsynonymous sites to define Ks and Ka, respectively (Nei and
Gojobori 1986). Jukes-Cantor correction was used to correct the computed pdistances for multiple substitutions at the same site and standard errors for Ka and
Ks were estimated from 1,000 bootstrap pseudoreplicates. Obtained Ka- and Ksvalues were used with three decimal positions to calculate Ka/Ks-values (ω), mean
Ka/Ks-values, and standard deviations in Excel 2003 (Microsoft). In addition, Ka/Ksvalues were classified in specific ranges and presented in balk-diagrams as
proportion of comparisons.
Two-tailed t-tests with unequal variances as implemented in Excel 2003 were done to
prove for significant differences of Ka/Ks-values. Tests were done between 1) copy A
and copy B of the MCS species; 2) gene copies of the MCS to sister- as well as
outgroup species, and 3) regulatory and structural genes.
37
Materials and Methods
Next to the calculation of Ka/Ks-values, selection pressure was tested for each gene
with the codon-based Z-test for large samples as implemented in MEGA. The null
hypothesis of neutrality was compared with the alternative hypotheses of positive and
purifying selection. Analyzed datasets were checked twice for each region, first, the
whole dataset including MCS and Sedum species and second comprising only MCS
species. Analyses were done with the Nei-Gojobori method with p-distance or JukesCantor distance, and 1,000 bootstrap re-samples with complete deletion of gaps.
Verifying positive selection using estimation of Ka/Ks-values is difficult because an
average over the whole protein sequence is done (Zhang 2003). Other methods
allow estimation of positive selection acting on single amino acid sites like the one
implemented in the Selecton web server (Doron-Faigenboim et al. 2005, Stern et al.
2007). Selecton enables to test the hypothesis of positive selection by comparing
different evolutionary models with each other. One of the selected models (M8)
allows for positive selection and is compared with the null model (M8a) which
assumes that there is no positive selection by setting Ka/Ks = 1. Testing the null
hypothesis of positive selection with the M8 model, the first result is a graphic
indicating with different colors the selection pressure acting on each amino acid site.
If there is evidence for positive selection the alternative model (M8a) is tested.
Subsequent likelihood ratio tests (LRT) estimate which model explains best the
analyzed data and indicate significance for positive selection.
Selecton enables the user to supply an accurate phylogeny to improve the results.
Therefore, BI phylograms were implemented in the analyses. Testing positive
selection within the MCS_PEPC gene region was done with two different datasets.
The first excludes the Kalanchoe sequences and the four Sedum sequences
S_pubescence_1c43 and 243 as well as S_surculosum_34341 and 54343, whereas
these Sedum sequences were included in the second analysis. Two datasets of
MCS_AP1 were used to search for positive selection, one containing all analyzed
sequences
and
the
second
excluded
the
highly
unique
sequences
of
A_aureum_13637 and 1b3637 as well as A_saundersii_33637. From the MCS_AP3
dataset the Kalanchoe sequence was excluded and the test of positive selection
performed.
38
Results
3. Results
3.1. Datasets
Alignments of MCS_PEPC, MCS_AP1, and MCS_AP3 were carefully checked for
sequences of chimeric origin, with ambiguous sites, and single base pair (bp)
deletions that result in frameshifts and premature stops in exon regions and
translated proteins. Sequences showing these features were excluded from the
datasets in a second approach although observation based on these sequences will
be discussed.
For each studied species one individual plant was used to extract DNA and to amplify
the gene of interest. Nevertheless, each species can be represented by several
different cloned sequences (named as clones or sequences) which may represent
different alleles or gene copies present in the species.
Identical clones are represented by a single sequence in the dataset and reduced
datasets were used for the analyses.
MCS_PEPC
In total, 83 sequences of MCS_PEPC homologs were obtained (table 3).
Table 3: Number of obtained total and reduced cloned sequences for MCS_PEPC.
taxon
total number reduced number
A. aureum
8
4
A. canariense
5
3
A. cuneatum
5
3
A. goochiae
5
5
A. nobile
5
4
A. rubrolineatum
6
4
A. saundersii
5
5
A. smithii
3
3
Ai. laxum
4
2
Ai. pachycaulon
5
3
M. anagensis
5
5
M. icterica
5
3
S. caeruleum
5
2
S. jaccardianum
5
5
S. modestum
0
0
S. pubescens
6
3
S. surculosum
6
5
∑ 83
∑ 59
39
Results
For S. modestum amplification failed and five sequences were excluded:
Ai_pachycaulon_313T showed a chimeric sequence type and an indel of one
nucleotide that will result in a frameshift mutation including several premature stop
codons. Likewise of potential chimeric origin and excluded were A_cuneatum_402,
A_nobile_361, A_canariense_41321, and M_icterica_472. Reduction of identical
species-specific clones resulted in a dataset comprising 59 sequences (table 3).
In total, aligned sequences have a full-length of 1708 bp and three exons and two
introns were detected (fig. 5, table 4).
Fig. 5: Schematic exon and intron structure of the MCS_PEPC gene sequences. Exons are shown as
boxes and introns as lines. The relative length of the respective parts is given in table 4.
The exon alignment has a length of 1107 bp resulting in a translated protein of 369
amino acids. Within exon 3 up to 15 nucleotides were exclusive for the four unique
sequences of S. pubescens and S. surculosum (compare below).
Table 4: Exon and intron positions and lengths of the MCS_PEPC sequences.
position
length
exon 1
1 – 429
429
intron 1
430 – 799
370
exon 2
800 – 1186
387
intron 2 1187 – 1415
229
exon 3
293
1416 – 1708
The intron alignment comprises 599 bp but intron regions are quite divers and
unambiguously alignable rather for taxa within the same genus. Therefore, MCS or
Sedum species were removed from the alignments, respectively. Arising gaps were
deleted but no further improvements of the MCS or Sedum intron alignments were
done. The MCS intron alignment comprises 177 bp whereas the intron region of the
Sedum species alone would result in 578 bp.
Deletion
of
the first
intron
was
observed
S_surculosum_44318 and 44342.
40
for S.
jaccardianum and for
Results
MCS_AP1
The most incomplete dataset was obtained for MCS_AP1. No amplification products
were obtained for Ai. laxum, Ai. pachycaulon, and S. pubescens. Sequencing failed
partly for S. caeruleum, therefore, only four exon sequences are available.
Three sequences were excluded because of chimeric origin: A_aureum_32223,
A_goochiae_12223, and A_nobile_3b3637. The first one showed additionally a
premature stop codon in the translated protein alignment and was excluded together
with further A. aureum sequences (12223, 23637, 33039, 4a2539) which shared this
feature. Furthermore, the sequence of A_goochiae_23637 was excluded because of
a 1 bp deletion that will result in a frameshift mutation and premature stop codon in
the coding sequence. In addition, this sequence showed an indel of 56 amino acids
between position 32 and 89 in the hypothetical protein. A frameshift mutation
resulting in a premature stop codon was also detected for A_smithii_23637. In
addition, this sequence had had an exclusive intron with a contrasting splice site.
Instead of GT/AG for excision of the intron CA/AG was detected. The features of
these sequences were confirmed by independent amplification. In total, 88
sequences were obtained and further reduction of identical species-specific clones
resulted in an alignment comprising 42 sequences (table 5).
Table 5: Number of obtained total and reduced cloned sequences for MCS_AP1.
taxon
total number reduced number
A. aureum
11
4
A. canariense
11
4
A. cuneatum
15
5
A. goochiae
7
1
A. nobile
5
3
A. rubrolineatum
4
1
A. saundersii
5
3
A. smithii
5
3
Ai. laxum
0
0
Ai. pachycaulon
0
0
M. anagensis
6
1
M. icterica
3
1
S. caeruleum
(4)
(3)
S. jaccardianum
4
4
S. modestum
8
5
S. pubescens
0
0
S. surculosum
8
4
∑ 88
∑ 42
41
Results
The full-length alignment comprises 3075 bp and eight exons and seven introns of
highly variable length were observed (fig. 6, table 6).
Fig. 6: Schematic exon and intron structure of the MCS_AP1 gene sequences. Exons are shown as
boxes and introns as lines. The relative length of the respective parts is given in table 6.
The exon alignment has a length of 651 bp (including S. caeruleum) which can be
translated into 217 amino acids. Focusing on the translated protein alignment of this
MADS-box gene, approx. amino acids 1-28 describe a part of the MADS-box domain,
amino acids 29-69 the I-domain, amino acids 70-137 the K-domain, and amino acids
138-217 a part of the C-terminal domain.
Table 6: Exon and intron positions and lengths of the MCS_AP1 sequences.
position
length
exon 1
1 – 98
98
intron 1
99 – 1758
1660
exon 2
1759 – 1861
103
intron 2 1862 – 2020
159
exon 3
2021 – 2085
65
intron 3 2086 – 2216
131
exon 4
2217 – 2316
100
intron 4 2317 – 2424
108
exon 5
2425 – 2466
42
intron 5 2467 – 2603
137
exon 6
2604 – 2645
42
intron 6 2646 – 2764
119
exon 7
2765 – 2925
161
intron 7 2926 – 3037
112
exon 8
38
3038 – 3075
The intron alignment comprises 2426 bp but only 1758 bp if the dataset was reduced
by the Sedum sequences. The intron alignment of the Sedum species comprised
1608 bp. Especially the first intron was long and extremely different between MCS
42
Results
and Sedum sequences. It comprises 1660 aligned base pairs if both, MCS and
Sedum species, were taken into account. However, only 522 bp were to some extent
alignable and shared between the species. If the Sedum species were removed, the
first intron comprises 1072 bp and 995 bp in a Sedum alignment.
MCS_AP3
For MCS_AP3 105 sequences of the studied species could be aligned (table 7). Four
sequences showed a premature stop codon: A_aureum_263 and A_cuneatum_302b
at the beginning, and A_rubrolineatum_26234 and A_rubrolineatum_305b at the end.
In addition, A_aureum_263 seemed to be chimeric and was excluded; likewise also
A_cuneatum_305b, A_rubrolineatum_301b, and A_saundersii_305. Reduction of
identical species-specific clones resulted in an alignment of 75 sequences (table 7).
Table 7: Number of obtained total and reduced cloned sequences for MCS_AP3.
taxon
total number reduced number
A. aureum
4
2
A. canariense
10
7
A. cuneatum
7
5
A. goochiae
5
3
A. nobile
12
3
A. rubrolineatum
7
4
A. saundersii
10
7
A. smithii
6
4
Ai. laxum
4
3
Ai. pachycaulon
7
5
M. anagensis
4
3
M. icterica
3
3
S. caeruleum
5
5
S. jaccardianum
4
4
S. modestum
6
6
S. pubescens
6
6
S. surculosum
5
5
∑ 105
∑ 75
The full-length alignment resulted in a length of 2265 bp. For K. blossfeldiana
(DQ479358) 30 nucleotides were missing at the beginning, for several Sedum
sequences the first 42 bp, and for all these sequences the 3´-UTR region was not
amplified.
43
Results
Sequences comprising seven exons and six introns and with position 2180 the 3´UTR region started (fig. 7; only amplified for MCS species). Exon and intron lengths
are quite variable and summarized in fig. 7 and table 8.
Fig. 7: Schematic exon and intron structure of the MCS_AP3 gene sequences. Exons are shown as
boxes and introns as lines. After the last exon box the 3´-UTR is indicated as line. The relative length
of the respective parts is given in table 8.
The exon alignment comprises 669 nucleotides and resulted in a protein alignment of
223 amino acids. Translation in the hypothetical MADS-box protein would result in a
part of the MADS-box (approx. amino acids 1-42), I-domain (43-71), K-domain (72138), and C-terminal domain (139-223).
Table 8: Exon and intron positions and lengths for MCS_AP3 sequences.
position
length
exon 1
1 – 143
143
intron 1
144 – 424
281
exon 2
425 – 491
67
intron 2
492 – 699
208
exon 3
700 – 761
62
intron 3
762 – 886
125
exon 4
887 – 986
100
intron 4
987 – 1303
317
exon 5
1304 – 1345
42
intron 5
1346 – 1546
201
exon 6
1547 – 1627
81
intron 6
1628 – 2005
378
exon 7
2006 – 2179
174
3´-UTR region 2180 – 2265
86
The intron alignment consists of 1510 bp and was reduced to 1166 bp if only the
MCS species were included. Introns of the Sedum species would result in 1318 bp.
Intron 7 is unique because most parts correspond exclusively to several MCS
species ordered in subclade B (see, e.g., fig. 14).
44
Results
3.2. Phylogenetic reconstructions
In the present study full-length as well as exon and intron regions were used to study
phylogenetic relationships. Combining exon and intron regions, which may evolve
quite differently, in a full-length dataset reveals the problem that conserved (exon)
and highly diverse (intron) regions are combined. Partition Homogeneity Tests
indicate significant differences for MCS_AP1 (p < 0.05) but not for MCS_PEPC (p =
0.719) and MCS_AP3 (p = 0.856).
Phylogenies were reconstructed using BI and ML. The model of sequence evolution
was first obtained by Modeltest and the respective model used to define the settings
for the BI and ML analyses. In table 9 the model for each dataset as well as the
settings and results of the BI analyzes are summarized.
Table 9: Selected model (AIC criterion) and results of the BI analyzes for different datasets. Default
th
settings of MrBayes (four chains, sample every 100 generation, random starting tree) were used.
dataset
model
temperature
p-value
generations burnin
MCS_PEPC_full-length
TVM+I+Γ
T = 0.2
0.005320 1,000,000
3,000
MCS_PEPC_exon
GTR+I+ Γ
T = 0.2
0.006626 1,000,000
3,000
MCS_PEPC_intron
HKY+I
T = 0.1
0.009362 2,000,000
6,000
MCS_PEPC_intron_Aeo HKY+I
T = 0.2
0.008906 600,000
1,800
MCS_AP1_full-length
TVM+ Γ
T = 0.2
0.003558 1,000,000
3,000
MCS_ AP1_exon
TrN+ Γ
T = 0.2
0.009073 1,000,000
3,000
MCS_ AP1_intron
GTR+ Γ
T = 0.2
0.009856 100,000
300
MCS_ AP1_intron_Aeo
TVM+ Γ
T = 0.2
0.007175 100,000
300
MCS_AP3_full-length
HKY+ Γ
T = 0.2
0.009833 1,800,000
5,400
MCS_AP3_full-
HKY+ Γ
T = 0.2
0.009303 1,500,000
4,500
MCS_AP3_exon
TIM+I+ Γ
T = 0.2
0.008387 150,000
4,500
MCS_ AP3_intron
GTR+ Γ
T = 0.2
0.009091 1,000,000
3,000
MCS_ AP3_intron_Aeo
K81uf+ Γ
T = 0.2
0.004791 1,000,000
3,000
length_without_3´-UTR
45
Results
MCS_PEPC
Phylograms based on full-length alignments of MCS_PEPC were in agreement
comparing BI (fig. 8) and ML (fig. 29; appendix). The BI phylogram provides a slightly
better resolution. Kalanchoe gracilis (AJ252946) and K. pinnata (AJ252919) were
used as outgroup. Several clones of S. pubescens (243 and 1c43) and S.
surculosum (34341 and 54343) were with a very long branch clearly separated from
the remaining Sedum and MCS species. These sequences are sister to the
remaining analyzed sequences in the BI phylogram (pp = 1.00) but show an
unresolved basal position for ML. Focusing on the main clade, S_pubescens_343 is
basal to all other sequences (pp = 1.00, bs = 55%). Sequences of S. surculosum are
found in different positions: S_surculosum_2c4391 basal to all other sequences (pp =
0.96, bs = 55%) and S_surculosum_44318 and 44342 as sister to S. jaccardianum
(pp = 1.00, bs = 100%). Sedum caeruleum is sister to this latter group supported with
a pp-value of 0.90 but without bootstrap support (43%) and these Sedum sequences
are sister (pp = 1.00, bs = 53%) to the monophyletic MCS species. The sequences of
the MCS species form two subclades supported with pp = 0.63 but without bootstrap
support (28%). Subclade A contains sequences of A. aureum, A. nobile, A.
rubrolineatum, A. smithii, Ai. pachycaulon, and M. anagensis. The Aeonium species
are sister to Ai. pachycaulon and M. anagensis (pp = 1.00, bs = 75%) and the latter
both species are also sister to each other (pp = 1.00, bs = 99%). Aeonium
rubrolineatum and A. nobile are sister to each other (pp = 1.00, bs = 99%); A. smithii
(pp = 1.00, bs = 95%) and A. aureum (pp = 1.00, bs = 100%) are successive basal to
them. Subclade B contains duplicated sequences of A. aureum, A. nobile, A.
rubrolineatum, A. smithii, and Ai. pachycaulon. Only M. anagensis shows no
duplicates within subclade B. Next to the duplicated sequences the remaining MCS
species, namely A. canariense, A. cuneatum, A. goochiae, A. saundersii, Ai. laxum,
and M. icterica, could be found within subclade B. Basal and sister to the Aeonium
species (pp = 1.00, bs = 100%) is a group comprising Ai. laxum and Ai. pachycaulon
in a sister relationship to each other (pp = 1.00, bs = 99%) and to M. icterica (pp =
1.00, bs = 98%). Within subclade B A. goochiae is basal to all Aeonium species (pp =
1.00, bs = 95%). The remaining Aeonium sequences form two distinct groups (pp =
1.00, bs = 88%). The first comprises A. aureum, A. saundersii, and A. smithii. The
latter two species are sister to each other (pp = 0.96, bs = 81%) and to A. aureum (pp
= 0.95, bs = 86%). The second group comprises A. canariense, A. cuneatum, A.
46
Results
nobile, and A. rubrolineatum. Aeonium cuneatum and A. rubrolineatum are sister to
each other (pp = 1.00, bs = 80%) and A_canariense_411 is basal to them (pp = 1.00,
bs = 81%). The remaining sequences of A. canariense, 412 and 41322, are sister to
A. nobile (pp = 1.00, bs = 99%). Both species groups are sister to each other (pp =
1.00, bs = 98%).
If the unique sequences of S. pubescens and S. surculosum were removed, the
same relationships were obtained. Also removing the Kalanchoe species and the
unique Sedum sequences did not change the topologies but decrease resolution
(data not shown).
Fig. 8: BI phylogram based on the full-length MCS_PEPC data. Posterior probabilities are given at the
nodes.
47
Results
Also in the exon phylograms K. gracilis (AJ252946) and K. pinnata (AJ252919) were
used as outgroup and both phylograms (BI fig. 9; ML fig. 30; appendix) were in
agreement. Sedum_pubescens_243 and 1c43 as well as S_surculosum_34341 and
54343 are separated from the remaining sequences by a long branch in the BI
phylogram (pp = 1.00). This relationship is unresolved in the ML phylogram. Sister to
this group are all other species sequences. Sedum_pubescens_343 is basal of the
main clade (pp = 0.99, bs = 76%). Positions of S. surculosum are divergent with
S_surculosum_2c4391 basal to all remaining sequences (pp = 0.97, bs = 73%) and
the other two S. surculosum sequences included in the Sedum subclade. Sedum
caeruleum is sister to S. jaccardianum and S. surculosum (pp = 0.97, bs = 65%),
which are also sister to each other (pp = 1.00, bs = 100%). The species of the MCS
form a monophyletic clade, separated in two subclades. The position of subclade A is
unresolved, comprising sequences of A. aureum, A. nobile, A. rubrolineatum, A.
smithii, Ai. pachycaulon, and M. anagensis. Two subgroups could be described, one
with A. rubrolineatum and A. nobile as sister to each other (pp = 0.88, bs = 87%) and
A. smithii (pp = 1.00, bs = 100%) as well as A. aureum (pp = 1.00, bs = 96%)
successive sister to them. This species group is sister (pp-value = 1.00, bs = 59%) to
the sequences of M. anagensis and Ai. pachycaulon, which are also sister to each
other (pp = 1.00, bs = 100%). Within subclade B duplicated sequences of A. aureum,
A. nobile, A. rubrolineatum, A. smithii, and Ai. pachycaulon were found whereas
sequences of M. anagensis were only detected within subclade A. Next to the
duplicates, sequences of A. canariense, A. cuneatum, A. goochiae, A. saundersii, Ai.
laxum, and M. icterica were found. Basal of subclade B are the sequences of Ai.
laxum and Ai. pachycaulon in a sister relationship to each other (pp = 1.00, bs =
98%) and to M. icterica (pp = 1.00, bs = 99%). The split of this group is highly
supported (pp = 1.00, bs = 100%) and A. goochiae is basal to all remaining Aeonium
species of subclade B (pp = 1.00, bs = 96%). Within subclade B two groups are
formed (pp = 1.00, bs = 84%). The first comprises A. aureum, A. saundersii, and A.
smithii; the second A. canariense, A. cuneatum, A. nobile, and A. rubrolineatum.
Aeonium saundersii and A. smithii are sister to each other (pp = 0.96, bs = 72%) with
A. aureum basal to them (pp = 0.96, bs = 86%). In a sister relationship are also A.
canariense and A. nobile (pp = 0.93, bs = 76%). Unique is the position of
A_canariense_411 (pp = 1.00, bs = 77%) basal to the sister relationship of A.
48
Results
cuneatum and A. rubrolineatum (pp 1.00 = bs = 75%). Both species groups are also
sister to each other (pp = 1.00, bs = 94%).
Exclusion of the unique sequences of S. pubescens and S. surculosum did not
change the topology. Also removing the Kalanchoe sequences and using
S_pubescens_343 as outgroup resulted in the same relationships within the ingroup
(data not shown).
Fig. 9: BI phylogram based on the MCS_PEPC exon data. Posterior probabilities are given at the
nodes.
49
Results
The phylograms based on the intron data are highly unresolved. However, the ML
phylogram (fig. 10) shows a better resolution compared to BI (fig. 31; appendix) and
will be described below. The Kalanchoe sequences were
excluded and
S_pubescens_243 and 1c43 were used as outgroup. These sequences as well as
S_surculosum_34341 and 54343 were unresolved at the basis and separated from
the main clade by a long branch. Sedum caeruleum is sister (bs = 70%) to all
remaining species and S_pubescens_343 and S_surculosum_2c4391 are basal to
the main clade (bs = 40%). Further S. surculosum sequences are mixed with the
MCS species in subclade B and were found together with Ai. laxum, Ai. pachycaulon,
and M. icterica. Subclade B contains also the sequences of S. jaccardianum in
unresolved positions. The separation of the main clade into two subclades obtains no
bootstrap support and relationships are also highly unresolved in the BI phylogram.
Subclade A comprises A. aureum, A. nobile, A. rubrolineatum, A. smithii, Ai.
pachycaulon, and M. anagensis in two subgroups but the separation is not supported
(bs = 25%). Aeonium nobile and A. rubrolineatum are sister to each other (bs = 60%)
and A. aureum (bs = 66%) and A. smithii (bs = 78%) are successive sister to them. In
the second subgroup Ai_pachycaulon_315T is sister to M. anagensis but without
support. Basal of subclade B are Ai. laxum and Ai. pachycaulon in a sister
relationship to each other and to M. icterica but no support was obtained for these
relationships. Subclade B is furthermore separated into two species groups, one
comprising A. canariense and A. nobile in a sister relationship to each other (bs =
70%) and A. cuneatum as well as A. rubrolineatum cluster to them. The second
group is formed by A. aureum, A. goochiae, A. saundersii, and A. smithii sequences
in unresolved relationships.
Given that unambiguous alignments of intron sequences between genera are not
trivial Sedum sequences were removed, arising gaps deleted, and Ai. laxum used as
outgroup. The ML phylogram shows a better resolution (fig. 33; appendix) but no
conflicts were found in comparison to the BI phylogram (fig. 32; appendix).
Aichryson_pachycaulon_311T and 312T cluster together unresolved basal of the
respective phylograms and M. icterica is basal to all remaining species (pp = 0.97, bs
= 92%). The separation of the main clade into two species groups is supported with a
pp-value of 0.80 and a bootstrap of 66%. Within subclade A, Ai_pachycaulon_315T
and M. anagensis are sister to each other (pp = 0.99, bs = 99%) and separated from
the duplicated Aeonium species (pp = 1.00, bs = 100%). Here, A. nobile is sister to A.
50
Results
rubrolineatum (pp = 0.96, bs = 71%) and both are successive sister to A. aureum (pp
= 0.99, bs = 82%) and A. smithii (pp = 1.00, bs = 95%). Focusing on the ML
phylogram, the remaining Aeonium sequences in subclade B form two subgroups but
without support (40%). One subgroup comprises A. canariense, A. cuneatum, A.
nobile, and A. rubrolineatum where A. canariense and A. nobile are sister to each
other (bs = 95%) but excluding A_canariense_411. The second subgroup is formed
by A. aureum, A. goochiae, A. saundersii, and A. smithii in unresolved relationships
to each other.
Fig. 10: ML phylogram based on the MCS_PEPC intron data. Bootstrap support is given at the nodes.
51
Results
MCS_AP1
Full-length BI (fig. 34; appendix) and ML (fig. 11) phylogenies of MCS_AP1 were in
agreement with each other and relationships slightly better resolved in the ML
phylogram. The analyzed Sedum species form a distinct group supported with a ppvalue of 1.00 and bootstrap of 100%. Sedum jaccardianum and S. surculosum are
sister to each other and to S. modestum; both relationships with highest support.
Sedum_surculosum_3 seems to be paralogous to the remaining S. surculosum
clones (pp = 1.00, bs = 78%). Nearly all sequences of the MCS species form one
main clade. However, the species group is not monophyletic since A_aureum_13637
and 1b3637 as well as A_saundersii_33637 were detected in derived positions (pp =
1.00, bs = 100%). The two sequences of A. aureum cluster together but are
separated from the additional A. saundersii sequence. The main clade consists of the
MCS species that form two distinct subclades (A and B) in a sister relationship to
each other (pp = 1.00, bs = 85%). Subclade B contains all studied and analyzed
species whereas subclade A comprises only duplicated sequences of A. aureum, A.
canariense, A. cuneatum, A. nobile, and A. smithii. Relationships within subclade A
are rather unresolved. Only a sister relationship between A. canariense and A. nobile
could be detected (pp = 1.00, bs = 86%) to which A. cuneatum is basal and sister (pp
= 1.00, bs = 88%). In the ML phylogram A. smithii sequences were basal to this
species group but without support (bs = 42%) and A_aureum_3b3637 basal to all of
them (bs = 100%). Within subclade B relationships are better resolved. Both
Monanthes species are basal to the Aeonium species even if their respective position
is unresolved. Aeonium goochiae is basal to all other Aeonium species (pp = 1.00, bs
= 90%) which form two distinct species groups (pp = 1.00, bs = 100%). Group one
comprises A. aureum which is sister to A. saundersii but the support for this
relationship is low (pp = 0.68, bs = 62%). Aeonium smithii is basal to them (pp = 0.99,
bs = 82%) and A. rubrolineatum is basal of this whole subgroup (pp = 1.00, bs =
95%). The second species group consists of A. canariense and A. nobile as sister to
each other (pp = 1.00, bs = 100%) and A. cuneatum basal of them (pp = 1.00, bs =
79%).
If the dataset is reduced by the first large intron relationships in the obtained BI
phylogram are mainly the same. The only difference is the position of
S_surculosum_3, which clusters separated from the remaining S. surculosum
sequences as sister to S. jaccardianum (data not shown).
52
Results
Fig. 11: ML phylogram based on the MCS_AP1 full-length data. Bootstrap support is given at the
nodes.
53
Results
Phylograms of the exon sequences are highly unresolved but in agreement
comparing BI (fig. 35; appendix) and ML (fig. 12) although the ML phylogram
provides a slightly better resolution. Sedum caeruleum sequences could be obtained
and were used as outgroup. The MCS species are not monophyletic because of the
unique A. aureum and A. saundersii sequences. Aeonium_aureum_13637 and
1b3637 cluster unresolved at the basis of all other analyzed sequences. The
sequence of A_saundersii_33637 is sister to the analyzed Sedum ser. Monanthoidea
species but this relationship is only weakly supported by a pp-value of 0.63 (bs =
42%). Within the Sedum group S. modestum is sister (pp = 1.00, bs = 100%) to the
sister species of S. jaccardianum and S. surculosum. Unique is again the position of
S_surculosum_3 intermixed with S. jaccardianum (pp = 0.98, bs = 88%) whereas the
remaining S. surculosum clones are sister to them (pp = 1.00, bs = 100%). The clade
comprising the Sedum species and A_saundersii_33637 is sister to the main clade
(pp = 0.51, bs = 54%) formed by the MCS species, which are distinguished into two
subclades (pp = 0.81, bs = 61%). Within subclade A, relationships between A.
aureum, A. cuneatum, and A. smithii are unresolved (pp = 0.69, bs = 57%). Aeonium
canariense (pp = 0.81, bs = 58%) and A. nobile are basal to them. Subclade B
comprises next to duplicated sequences of A. aureum, A. canariense, A. cuneatum,
A. nobile, and A. smithii also the sequences of A. goochiae, A. rubrolineatum, A.
saundersii, M. anagensis, and M. icterica. The latter two are unresolved at the base
of the subclade focusing on the BI phylogram. In the ML phylogram, M. anagensis is
basal of subclade B (bs = 84%) and M. icterica basal of all Aeonium species (bs =
45%). Basal of all other Aeonium species A. goochiae could be found (pp = 0.83, bs
= 46%) and further relationships within subclade B are rather unresolved.
Nevertheless, the sister relationship of A. canariense and A. nobile is observed (pp =
1.00, bs = 100%) and A. cuneatum is basal to them (pp = 0.66, bs = 53%). Focusing
on the ML phylogram, also A. aureum and A. smithii (bs = 40%) are sister to each
other.
54
Results
Fig. 12: ML phylogram based on the MCS_AP1 exon data. Bootstrap support is given at the nodes.
Similar clades and relationships are obtained using exclusively the intron region of
MCS_AP1. BI (fig. 13) and ML (fig. 36; appendix) phylograms were in major
agreement and without conflicting signals. Excluding S. caeruleum, three groups of
MCS species and one exclusive Sedum ser. Monanthoidea clade were obtained. A
unique position was detected for A_aureum_13637 and 1b3637 as well as
A_saundersii_33637. The first two sequences were selected as outgroup and
55
Results
A_saundersii_33637 is basal to all remaining analyzed sequences in the BI
phylogram (pp = 1.00). Again, due to these three sequences, the MCS species are
not monophyletic. Within the Sedum clade S. modestum is sister to S. jaccardianum
and S. surculosum (pp = 1.00, bs = 100%), which are also sister to each other (pp =
1.00, bs = 100%). Sedum_surculosum_3 showed again a unique position as sister to
the other S. surculosum sequences (pp = 1.00, bs = 86%). The Sedum clade is basal
(pp = 1.00, bs = 100%) to the main clade comprising the analyzed MCS. This main
clade is divided into two distinct subclades (pp = 0.95, bs = 69%). Subclade A,
comprising A. aureum, A. canariense, A. cuneatum, A. nobile, and A. smithii,
separated into two species groups (pp = 1.00, bs = 100%). Aeonium canariense is
sister to A. nobile (pp = 0.98, bs = 82%) and A. cuneatum is sister to them (pp = 1.00,
bs = 97%). This species group is sister to the sister relationship of A. aureum and A.
smithii (pp = 0.74, bs = 84%). The second subclade B contains next to the
sequences of the above mentioned species the two Monanthes species and the
sequences of A. goochiae, A. rubrolineatum, and A. saundersii. Focusing on the ML
phylogram, both Monanthes species are sister to each other (bs = 54%) and basal to
the species of Aeonium (bs = 100%). The relationship between the two Monanthes
species is not resolved in the BI phylogram but the separation from the Aeonium
species supported with a pp-value of 1.00. Basal of the Aeonium clade is A. goochiae
(pp = 0.95, bs = 80%). Two species groups are formed (pp = 1.00, bs = 100%) where
the first comprises A. canariense and A. nobile in a sister relationship to each other
(pp = 1.00, bs = 100%) and to A. cuneatum (pp = 1.00, bs = 84%). The second
species group is formed by A. aureum which is sister to A. saundersii (pp = 0.74, bs =
78%); A. smithii (pp = 0.98, bs = 83%) and A. rubrolineatum (pp = 1.00, bs = 93%)
successive basal to them.
After removing the Sedum sequences, A_aureum_13637 and 1b3637 were used as
outgroup (BI fig. 37 and ML fig. 38; both appendix). Aeonium_saundersii_ 33637 has
a unique position basal to the remaining analyzed sequences and species (pp = 1.00,
bs = 100%). The main clade is again divided into two subclades (pp = 1.00, bs =
100%) where exactly the same relationships as described above for the full intron
dataset could be observed with high supports.
56
Results
Fig. 13: BI phylogram based on the MCS_AP1 intron data. Posterior probabilities are given at the
nodes.
MCS_AP3
For phylogenetic reconstructions of the MCS_AP3 sequences K. blossfeldiana
(DQ479358) was used as outgroup in the full-length dataset. BI (fig. 39; appendix)
and ML (fig. 14) phylograms are in major agreement and ML provides a slightly better
resolution. Exclusion of the 3´-UTR results in the same topology and will be
described below.
57
Results
The sequences of S. pubescens form two distinct groups. One is unresolved at the
base and the other sister and basal to all remaining species (pp = 0.99, bs = 69%).
The remaining species form two major cluster (pp = 0.88, bs = 55%), one containing
the Sedum species and the second monophyletic group is formed by the species of
the MCS. Sedum caeruleum is sister to the Sedum ser. Monanthoidea species
supported by a pp-value of 0.87 and bootstrap of 54%. Sedum modestum is sister to
S. jaccardianum and S. surculosum (pp = 1.00, bs = 99%). The sequences of S.
jaccardianum form a monotypic group and S. surculosum is sister to them. However,
S_surculosum_1 and 8 are sister to S. jaccardianum (pp = 1.00, bs = 96%) whereas
S_surculosum_2, 3, and 6 are sister and basal to this group of sequences (pp = 0.87,
bs = 59%). The species of the MCS form two distinct subclades (pp-value = 1.00, bs
= 86%). Subclade A contains exclusively Aeonium species (A. aureum, A.
canariense, A. cuneatum, A. rubrolineatum, A. smithii) and subclade B comprises
also the Aichryson and Monanthes species. Within subclade A two distinct species
groups are formed (pp = 1.00, bs = 96%). Highly supported is the sister relationship
between A. canariense and A. cuneatum (pp = 1.00, bs = 100%). These species are
sister to the species group formed by A. aureum, A. rubrolineatum, and A. smithii.
Aeonium aureum is basal (pp = 0.92, bs = 64%) to A. rubrolineatum and A. smithii
which are sister to each other (pp = 1.00, bs = 100%). Within subclade B Ai.
pachycaulon and M. icterica are intermixed (pp = 1.00, bs = 100%) with Ai. laxum as
sister to them. Aichryson_laxum_26634 is in a derived position compared to the other
two Ai. laxum sequences and shows differences between the BI and ML phylograms.
This sequence is basal to the remaining Ai. laxum, Ai. pachycaulon, and M. icterica
sequences in the ML phylogram with 100% bootstrap support. In contrast, in the BI
phylogram the Ai. laxum clones form a monotypic species group (pp = 0.71) basal to
Ai. pachycaulon and M. icterica (pp = 1.00) and with Ai_laxum_26634 as sister to the
other Ai. laxum sequences. The group formed by Ai. laxum, Ai. pachycaulon, and M.
icterica is basal to the remaining sequences of subclade B in the ML phylogram (bs =
99%). In the BI phylogram this relationship is unresolved. The position of M.
anagensis is also unresolved in the BI phylogram, whereas this species is basal to all
Aeonium species in the ML phylogram (bs = 53%). Aeonium goochiae is basal to the
remaining Aeonium species (pp = 1.00, bs = 98%). Two distinct Aeonium species
groups were formed (pp = 1.00, bs = 100%), one comprising A. aureum, A.
saundersii, and A. smithii (pp = 0.97, bs = 68%). The second group combines A.
58
Results
canariense, A. cuneatum, and A. nobile with A. rubrolineatum basal to them (pp =
0.65, bs = 48%). Aeonium canariense and A. cuneatum are sister to each other (pp =
0.57, bs = 55%) with A. nobile highly supported (pp = 1.00, bs = 100%) basal to
them.
Fig. 14: ML phylogram based on the MCS_AP3 full-length data. The 3´-UTR region is excluded.
Bootstrap support is given at the nodes.
59
Results
BI (fig. 15) and ML (fig. 40; appendix) phylograms of the exon region are in major
agreement but some differences were observed in subclade A. Kalanchoe
blossfeldiana (DQ479358) was used as outgroup and sequences of S. pubescens
cluster into two distinct groups. One is unresolved at the base and one basal to all
remaining species (pp = 0.96, bs = 71%). Sedum caeruleum is sister and basal to all
further analyzed species (pp = 0.79, bs = 57%). Sedum ser. Monanthoidea species
cluster together (pp = 1.00, bs = 97%) and S. modestum is sister to S. jaccardianum
and S. surculosum. Sedum jaccardianum and S. surculosum are also sister to each
other but with a unique pattern for S. surculosum. Sedum_surculosum_1 and 8 are
sister to S. jaccardianum with a support of pp = 1.00 and bootstrap of 93%. In
contrast, S_surculosum_2, 3, and 6 are basal and sister to this above mentioned
group (pp = 0.98, bs = 65%). The Sedum species are separated from the species of
the MCS with a pp-value of 0.88 but no bootstrap support is obtained (41%). Within
the main clade of the MCS species two subclades could be distinguished (pp = 0.98,
bs = 77%). Subclade A comprises only duplicated sequences of several Aeonium
species whereas subclade B also contains the Aichryson and Monanthes species.
Two species groups were detected within subclade A focusing on the BI phylogram
and differences were observed between BI and ML phylograms. In the BI phylogram
A. rubrolineatum and A. smithii are sister to each other (pp = 0.86) and A. aureum is
basal to them (pp = 0.53). They are sister (pp = 1.00) to the second species group
which is formed by A. canariense and A. cuneatum in a sister relationship to each
other (pp = 1.00). In contrast, ML positioned A_smithii_226 basal of subclade A (bs =
97%); A. canariense and A. cuneatum are sister to each other (bs = 98%) with A.
aureum basal to them (bs = 57%). The sequences of A. rubrolineatum are, without
support, basal to this relationship. Relationships within subclade B are identical
comparing BI and ML phylograms. Next to duplicated gene copies of several
Aeonium species, subclade B also comprises the species of Aichryson and
Monanthes, A. goochiae, A. nobile, and A. saundersii. Aichryson laxum is sister (pp =
1.00, bs = 100%) to the complete intermixed sequences of Ai. pachycaulon and M.
icterica (pp = 0.99, bs = 69%). Aichryson_laxum_26634 is thereby basal to the both
remaining Ai. laxum sequences (pp = 0.83, bs = 64%) and the whole species group is
basal of subclade B (pp = 1.00, bs = 96%). Monanthes anagensis is basal to all
Aeonium species (pp = 0.94, bs = 81%). Aeonium goochiae is basal and sister (pp =
0.88, bs = 68%) to the remaining Aeonium species which form two groups within
60
Results
subclade B (pp = 1.00, bs = 84%). One subgroup (pp = 0.96, bs = 63%) comprises A.
aureum, A. saundersii, and A. smithii in unresolved relationships. The second
subgroup (pp = 0.97, bs = 58%) is formed by A. canariense and A. cuneatum which
are sister to each other (pp = 0.99, bs = 60%) and cluster together with A. nobile and
A. rubrolineatum in unresolved relationships.
The same relationships are also obtained if Kalanchoe is excluded and S. pubescens
used as outgroup and also if the focus lies only on orthologous sequences (data not
shown).
Fig. 15: BI phylogram based on the MCS_AP3 exon data. Posterior probabilities are given at the
nodes.
61
Results
For the intron phylogram Kalanchoe was excluded and S. pubescens used as
outgroup. Relationships inferred by BI (fig. 16) and ML (fig. 41; appendix) phylograms
are in agreement focusing on the relationships of the MCS species but show
conflicting signals for Sedum. Using S_pubescens_6, 7, and 8 as outgroup, the
sequences S_pubescens_1, 2, 3 are unresolved at the base of all remaining species
in the BI phylogram. In contrast, this species group is found basal to the main clade
of the MCS species in the ML phylogram but lacks bootstrap support (38%). Within
the BI phylogram, S. caeruleum is sister to the species of Sedum ser. Monanthoidea
(pp = 0.53) and form together with them a unique Sedum clade in an unresolved
position. Sedum modestum is sister and basal (pp = 1.00) to the sister relationship of
S. jaccardianum and S. surculosum. The two sequences of S_surculosum_1 and 8
are sister to S. jaccardianum (pp = 0.99) whereas the three remaining S. surculosum
sequences (2, 3, 6) are sister to this group (pp = 0.60). Relationships within the ML
phylogram are differently resolved. The three species of Sedum ser. Monanthoidea
form a unique clade in an unresolved position and S. modestum is sister and basal to
S. jaccardianum and S. surculosum (bs = 100%). As in the BI phylogram,
S_surculosum_1 and 8 are sister to S. jaccardianum (bs = 78%) whereas
S_surculosum_2, 3, and 6 are sister to this group (bs = 69%). Sedum caeruleum is
basal to the main MCS clade and S_pubescens_1, 2, and 3 but without support
(37%). Relationships within the main clade comprising all studied Aeonium,
Aichryson, and Monanthes species are identical comparing both methods. The main
clade is separated into two subclades (pp = 1.00, bs = 95%). Subclade A comprising
duplicated sequences of A. aureum, A. canariense, A. cuneatum, A. rubrolineatum,
and A. smithii. Aeonium canariense and A. cuneatum are sister to each other (pp =
1.00, bs = 100%) and to the group formed by A. aureum, A. rubrolineatum, and A.
smithii (pp = 1.00, bs = 100%). Aeonium rubrolineatum is sister to A. smithii (pp =
1.00, bs = 100%) and A. aureum basal to them (pp = 0.91, bs = 74%). Subclade B
contains all remaining analyzed sequences of Aeonium, Aichryson, and Monanthes.
Monanthes anagensis is, with highest support, basal to all other species. The two
Aichryson species and M. icterica form a subclade basal to the Aeonium species (pp
= 0.61, bs = 64%). Sequences of Ai. pachycaulon and M. icterica are intermixed (pp
= 1.00, bs = 100%) and two sequences of Ai. laxum are sister to them (pp = 0.61, bs
= 53%). Basal to this group is Ai_laxum_26634 (pp = 1.00, bs = 100%). Aeonium
goochiae is basal to all other Aeonium species (pp = 1.00, bs = 99%) which could be
62
Results
separated into two subgroups (pp = 1.00, bs = 100%). One subgroup describes the
sister relationship of A. canariense and A. nobile to each other (pp = 0.97, bs = 88%)
and A. cuneatum basal to them (pp = 1.00, bs = 100%). Relationships within the
second subclade are unresolved containing A. aureum, A. rubrolineatum, A.
saundersii, and A. smithii.
Fig. 16: BI phylogram based on the MCS_AP3 intron data. Posterior probabilities are given at the
nodes.
63
Results
If only the MCS species were taken into account (BI fig. 42 and ML fig. 43; both
appendix) and Ai. laxum is considered as outgroup, identical relationships are
obtained in the BI and ML phylograms. Monanthes icterica is intermixed with Ai.
pachycaulon (pp = 1.00, bs = 100%) in a sister relationship to Ai. laxum (pp = 0.66,
bs = 55%) but the position of the whole subgroup is unresolved at the base of the
respective phylograms. For the remaining species two subclades (pp = 1.00, bs =
100%) could be detected reflecting mainly the above mentioned relationships for the
full intron dataset.
3.3. Blast and Neighbor-joining analyses
To confirm results deduced from the phylograms sequences were blasted. For further
confirmation NJ analyses were done that help to classify the sequences and to
distinguish more powerfully between orthologous and paralogous gene copies.
Interesting results were obtained for MCS_PEPC where the clear separation of the
four unique Sedum sequences (S_pubsecens_243, 1c43 and S_surculosum_34341,
54343) is supported by Blast and NJ analyses. Whereas all other species clones
show highest similarity to the same sequences – PEPC genes of several Kalanchoe
species (e.g., AJ252917, X87819, X87818, AJ231288, AJ252946, AJ344052), Cycas
revoluta (AJ312617) or Euphorbia tirucalli (AJ312660) – the separated Sedum show
other similarities. For S_pubescens_1c43 highest similarity was as well found to
PEPC genes of Kalanchoe species but all of them describe different isoforms as the
above mentioned ones. High similarity was found to, e.g., isogenes 5 (AJ344056)
and 6 (AJ344057) of K. pinnata and to PEPC genes of other species such as Vitis
vinifera (AF236126). Sedum_pubescens_243 shows similarities to, e.g., the K.
pinnata PEPC isogenes 5 (AJ344056), 6 (AJ344057), 7 (AJ244058), and to PEPC
genes of Crataegus (EU500593), Digitaria didactyla (AM690213), Lotus japonicus
(AB092820), Lupinus luteus (AM237200), and Vicia faba (AJ011303). Also
S_surculosum_34341 and 54343 show highest sequence similarity to PEPC genes
of different taxa like Lupinus albus (AY663388) or Oryza sativa (AY187619).
The clear separation of the unique sequences of S. pubescens and S. surculosum
was also confirmed in the NJ phylogram (fig. 45; appendix). All Aeonium, Aichryson,
Monanthes, and the remaining Sedum sequences are closely related to each other
and in a sister relationship to the mentioned PEPC isogenes 1, 2, and 3 of diverse
Kalanchoe species. In contrast, the unique S. pubescens and S. surculosum
64
Results
sequences are closely related to the isogenes 5, 6, and 7 of several Kalanchoe
PEPC genes.
For MCS_AP1 most sequences show highest similarity to AP1-like sequences of
Corylopsis sinensis (AY306146) and Heuchera americana (AY306148). Clade
specific sequences could be observed for subclade B. Namely A. goochiae, A.
rubrolineatum, A. saundersii (including the sequence 33637), and both Monanthes
species show high sequence similarity to AP1-like genes of Pyrus pyrifolia
(EF423915, EF423916), Malus x domestica (AB458503, AY071921, EU672877),
Prunus persica (EU079377), and Eriobotrya japonica (AY880261, AY880262).
Furthermore, all species of subclade B, with exception of the both Monanthes, show
a high similarity to Citrus sinensis AP1-like sequences (AY338974, AY338975). Also
the Sedum species have sequence similarity to the above mentioned species.
Therefore, in general no convincing clade or subclade specific pattern was observed.
The NJ analysis of the enlarged MCS_AP1 dataset revealed that all analyzed
sequences of the MCS and Sedum species are in close relationship to each other
and more distantly related to AP1-like sequences of other species. Even the three
unique sequences of A. aureum and A. saundersii are imbedded. They show the
closest relationship to S. caeruleum but belong clearly to the MCS/Sedum clade.
Sister to MCS_AP1 are AP1-like sequences of several genera such as Citrus, Malus,
Prunus, and Pyrus. Basal in a sister relationship are homologs of AGL8 and
FRUITFUL-like sequences (data not shown).
A clear pattern was observed for the blasted MCS_AP3 sequences of the studied
species. All sequences show highest sequence similarity to the AP3-like protein of K.
blossfeldiana (DQ479358). In addition, the sequences of subclade A show a high
sequence similarity to DEF-like genes of Pedicularis groenlandica (AY524010) or in
case of A. rubrolineatum to DEF-like genes of Mazus reptans (AY530538) and
Torenia fournieri (AB359951). Exceptions are the sequences A_cuneatum_26734
and 26634 that show only similarity to the above mentioned K. blossfeldiana AP3-like
gene (DQ479358). Nevertheless, a clear AP3-like subclade (B) and a DEF-like
subclade (A) could be distinguished (e.g., fig. 15).
In the NJ analysis MCS and Sedum sequences are closely related to each other and
cluster together in a sister relationship to AP3-like gene sequences of other species
(data not shown).
65
Results
3.4. Species phylogeny based on nrITS
One aim was to compare the genealogies of the low-copy nuclear coding genes with
the species phylogeny based on neutral markers. The nrITS alignment was improved
and the obtained BI phylogram (fig. 17) shows some differences to the nrITS
Maximum parsimony (MP) phylogram of Mort et al. (2002). The MCS are
monophyletic comprising all Aeonium, Aichryson, and Monanthes species. Using S.
caeruleum as outgroup it is together with S. pubescens basal to the remaining
species. Sedum jaccardianum and S. surculosum are sister to each other (pp = 1.00)
and to S. modestum (pp = 1.00). Sedum ser. Monanthoidea species are in a sister
relationship to the MCS species (pp = 1.00), which are distinguished in three main
clades. Monophyletic groups are formed by Aichryson (pp = 1.00), the perennial
Monanthes (pp = 1.00), and Aeonium (pp = 1.00). The position of the annual M.
icterica is unresolved but still, the species is closer related to Aichryson than to the
other Monanthes species. If the dataset of Mort et al. (2002) is used without
optimization, the BI phylogram resolves M. icterica as sister to the Aichryson clade
(data not shown). The species group comprising M. icterica and Aichryson is sister to
Aeonium and to the perennial Monanthes species (p = 1.00). Mort et al. (2002)
resolved this relationship differently with M. icterica basal and sister to Aeonium and
to the remaining Monanthes species but not closely related to Aichryson. In the BI
phylogram, the perennial Monanthes species are sister to Aeonium (pp = 1.00).
Within the Aeonium clade relationships are mainly unresolved, nevertheless several
subclades could be recognized. Basal of the Aeonium clade is the species group of
A. goochiae, one studied species, and A. lindleyi (pp = 0.99). The position of A.
cuneatum as further studied species is unresolved but seems to be rather basal to
the remaining Aeonium species which form four further subclades. The second
unresolved subclade B consists of five Aeonium species (pp = 0.94) including A.
canariense as studied species. Subclade C comprises seven different species
including A. aureum, A. saundersii, and A. smithii as studied species in sister
relationships. Aeonium smithii is basal to the other species (pp = 0.70) and A.
saundersii is sister (pp = 0.52) to the unresolved species group comprising A.
viscatum and the former Greenovia species (pp = 1.00). Subclade D consists of 13
species with A. nobile as a further studied species and A. glutinosum and A.
glandulosum basal (pp = 0.78) of this highly unresolved clade. Finally subclade E
comprises 14 species, including the studied species A. rubrolineatum but
66
Results
relationships are unresolved. The corresponding ultrametric tree is shown in fig. 46
(appendix).
Fig. 17: BI phylogram based on the improved nrITS dataset. Own sequences are marked with
numbers. Posterior probabilities are given at the nodes.
3.5. Gene duplications
Duplicated gene copies were detected based on exon phylograms for all three lowcopy nuclear genes; for different species and distinguished into orthologs and
paralogs. A general observation was the subclade specific duplication for the MCS
species, a phenomenon which was not observed for the Sedum species. For MCS
67
Results
species the main clade was separated into different subclades. The respective gene
copies within one subclade are orthologous to each other and paralogous to the
duplicated sequences (gene copies) of the other subclade. This pattern was mainly
obtained for the Aeonium species with one exception: the duplication of MCS_PEPC
for sequences of Ai. pachycaulon. Paralogous gene duplications were observed for
several Sedum species but also for MCS species within the respective subclades.
Summarizing the obtained pattern: species of the MCS, especially Aeonium, display
two or more copies of MCS_PEPC, MCS_AP1, and MCS_AP3 in their genomes
whereas most Sedum species only possess one copy.
For MCS_PEPC duplicates were found for A. aureum, A. nobile, A. rubrolineatum, A.
smithii, Ai. pachycaulon, S. pubescens, and S. surculosum (e.g., fig. 9). For the
Sedum species the observed pattern is unique. Four of the duplicated S. pubescens
and S. surculosum sequences are separated by a very long branch. Nevertheless, for
S. surculosum a duplication was also detected within the main clade. The MCS
species show a subclade specific duplication. The main clade represented by MCS
could be separated into two subclades, which each comprises duplicated gene
copies of the above mentioned Aeonium and Aichryson species. A unique pattern
was observed for A. canariense. Duplicated genes copies were detected within
subclade B where two clones are sister to A. nobile and A_canariense_411 is basal
to the species group formed by A. cuneatum and A. rubrolineatum.
For MCS_AP1 subclade specific duplications were detected for A. aureum, A.
canariense, A. cuneatum, A. nobile, A. saundersii, and A. smithii (e.g., fig. 12).
Sequences of A. aureum could be found in three different positions in the phylogram.
In each subclade one copy of A. aureum was detected. Additionally, A. aureum forms
a unique subclade unresolved at the basis of all analyzed sequences. Also the
sequences of A. saundersii show a unique duplication pattern. Whereas
A_saundersii_33637 is sister to the species of Sedum ser. Monanthoidea, further
sequences were found in the main clade comprising all studied MCS species.
Identical patterns of gene duplication were detected for A. canariense, A. cuneatum,
A. nobile, and A. smithii, which each have a copy in both subclades. Paralogous
duplications were detected for the Sedum species. Sedum_caeruleum_4 is in a
derived position compared to the other S. caeruleum sequences. For S. surculosum:
S_surculosum_3 clusters together with S. jaccardianum whereas further sequences
form a unique clade as sister to them.
68
Results
For MCS_AP3 again two subclades were obtained for the MCS species (e.g., fig.
15). Duplications were detected for A. aureum, A. canariense, A. cuneatum, A.
rubrolineatum, and A. smithii. These species are represented in each of the
orthologous subclades A and B. Single paralogous gene duplications were detected
within Monanthes and Aichryson. Further single duplication events resulting in
paralogs were detected for other species such as A. canariense in subclade A or A.
nobile and A. smithii in subclade B. Also for nearly all analyzed Sedum species single
gene duplications were detected. Sedum surculosum shows hereby the strongest
pattern. Sedum_surculosum_1 and 8 are sister to S. jaccardianum whereas
S_surculosum_2, 3, and 6 are basal and sister to this relationship. Paralogs were
also detected for S. modestum and S. pubescens.
In general, two or more subclade specific gene copies were detected for the species
of the MCS but only one copy for the Sedum species. For MCS_AP1 and MCS_AP3
each time a main copy (one subclade) is detected comprising all studied species. In
addition, a second (or more) subclades were observed comprising duplicated gene
copies of Aeonium species. Single duplications were detected for several species
within subclades and the deduced copies paralogous. Sedum pubescens and S.
surculosum show duplications for all three studied gene regions. Focusing on the
MCS species, A. aureum and A. smithii show for all three genes duplications; A.
canariense, A. cuneatum, A. nobile, and A. rubrolineatum have gene duplications at
least for two genes, and A. saundersii shows only a duplication for MCS_AP1.
Several species like A. goochiae, M. anagensis, and M. icterica show never a gene
duplication event.
3.6. Nucleotide differences, replacements, and amino acid
substitutions
Nucleotide differences were counted between all obtained species-specific
sequences and also distinguished into paralogous and orthologous species-specific
sequences of subclade A and B, respectively. The number and percentage of
differences is generally higher between paralogs. Even if synonymous and
nonsynonymous substitutions were observed, only nonsynonymous replacements
were counted because they may have an important impact on protein evolution.
Additionally, it was recognized if the replaced amino acid was quite different
compared to the original one.
69
Results
For MCS_PEPC a unique pattern was observed. Four unique Sedum sequences are
separated from the remaining sequences with a very long branch (e.g., fig. 9).
Therefore, for S. pubescens and S. surculosum highest numbers and percentage of
nucleotide differences and replacements were observed. They range between 24.9%
and 27.6% for the nucleotide differences and between 15.7% and 16.8% for the
replacements, respectively (table 10). For the remaining duplicates of Sedum 7.5%
(S. pubescens) and 11.6% (S. surculosum) nucleotide differences could be observed.
Those were clearly higher than that obtained for orthologs where values range
between 0.1% and 0.6%; for S. surculosum a higher value of 2.2% was detected. For
the MCS species comparisons of nucleotide differences for sequences from subclade
A with sequences from subclade B range between 5.5% (A. aureum) and 9.4% (Ai.
pachycaulon). Low values were observed for A. canariense (1.1%) and likewise for S.
jaccardianum (1.4%). However, the relationships of these respective sequences have
to be described as “paralogous alleles” since the compared species-specific
sequences were found in the same subclade but in derived positions to each other.
Nevertheless, their values are still different from those of true alleles or paralogs (for
comparison, e.g., Fortune et al. 2007 or Zhang et al. 2008).
The values for replacements range for paralogous gene copies of MCS species
between 4.1% (A. aureum) and 6.5% (Ai. pachycaulon). For orthologous sequences
the values range between 0.3% and 0.8% with 0% for Ai. pachycaulon. For Sedum
the values of replacements range between 0.8% and 5.7% for paralogs, and 0.3%
and 0.5% for orthologous sequences (0% for S. caeruleum).
The values for quite different substituted amino acids (aa) range for paralogs
between 54.2% (Ai. pachycaulon) and 75.8% (S. surculosum) with a mean of 66.9%.
As already indicated in table 10 the replaced amino acids in orthologous gene copies
are mostly quite similar to the original one.
70
Results
Table 10: Number and percentage of nucleotide differences, replacements, and quite different amino
acids (aa) between species-specific sequences of MCS_PEPC distinguished into orthologous and
paralogous sequences.
species
A. aureum
A. canariense
A. cuneatum
A. goochiae
A. nobile
A. rubrolineatum
A. saundersii
A. smithii
Ai. laxum
Ai. pachycaulon
M. anagensis
M. icterica
S. caeruleum
S. jaccardianum
S. modestum
S. pubescens
S. pubescens
S. surculosum
S. surculosum
nucleotide
differences
no.
%
61
5.5
12
1.1
paralogs
replacements
no.
%
15
4.1
4
1.1
quite
different aa
no.
%
10
66.7
3
75.0
76
78
6.9
7.0
17
16
4.6
4.3
11
11
64.7
68.8
82
7.4
18
4.9
12
66.7
104
9.4
24
6.5
13
54.2
15
276
83
305
128
1.4
3
58
5
62
21
0.8
2
41
3
47
14
66.7
24.9
7.5
27.6
11.6
15.7
1.4
16.8
5.7
70.7
60.0
75.8
66.7
nucleotide
differences
no.
%
2
0.2
1
0.1
3
0.3
3
0.3
2
0.2
2
0.2
5
0.5
3
0.3
5
0.5
1
0.1
5
0.5
3
0.3
1
0.1
7
0.6
2
24
0.2
2.2
orthologs
replacements
no.
%
1
0.3
1
0.3
1
0.3
1
0.3
1
0.3
1
0.3
2
0.5
2
0.5
1
0.3
0
0.0
3
0.8
1
0.3
0
0.0
2
0.5
1
2
quite
different aa
no.
%
0
0
0
0
1
50
0
0
0
0
0
0
1
50
1
50
0
0
0
0
2
66.7
0
0
0
0
1
50
-
0.3
0.5
0
1
0
50
Sequence divergence between paralogs of MCS_AP1 range between 1.2% for S.
surculosum and 8.3% for A. aureum (table 11). For the species of the genus
Aeonium a mean of 5.8% could be estimated and a range between 4.3% and 8.3%
(both A. aureum) observed. For S. caeruleum a value of 2.8% was found for the
paralogous sequences. For orthologous sequences of the MCS species the obtained
values are significantly lower than the one of the paralogs ranging between 0.2%
(several species) and 0.8% for A. smithii. However, most species-specific orthologs
showed no nucleotide differences at all. For the Sedum species the values range
between 0.2% (S. caeruleum and S. modestum) and 0.3% (S. jaccardianum).
The mean value for replacements is 7.6%. Focusing on paralogous sequences the
highest value of 12.9% was observed for A. aureum and the lowest value of 6.5% for
A. nobile for the MCS species. For the Sedum species the values 2.8% (S.
surculosum) and 4.6% (S. caeruleum) were observed.
Values for replaced amino acids with quite different characteristics were high (table
11) and ranged between 65.2% (A. saundersii) and 87.5% (A. aureum) for
paralogous sequences with a mean of 74.7%.
71
Results
Table 11: Number and percentage of nucleotide differences, replacements, and quite different amino
acids (aa) between species-specific sequences of MCS_AP1 distinguished into orthologous and
paralogous sequences.
species
A. aureum
A. aureum
A. canariense
A. canariense
A. cuneatum
A. cuneatum
A. goochiae
A. nobile
A. rubrolineatum
A. saundersii
A. smithii
Ai. laxum
Ai. pachycaulon
M. anagensis
M. icterica
S. caeruleum
S. jaccardianum
S. modestum
S. pubescens
S. surculosum
nucleotide
differences
no.
%
54
8.3
28
4.3
32
4.9
paralogs
replacements
no.
%
28
12.9
16
7.4
17
7.8
quite
different aa
no.
%
23
82.1
14
87.5
12
70.6
nucleotide
differences
no.
%
1
0.2
orthologs
replace
ments
no.
%
1
0.5
quite
different aa
no.
%
1
100
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
31
4.8
15
6.9
11
73.3
29
4.5
14
6.5
10
71.4
0
0
0
0
0
0
43
46
-
6.6
7.1
23
20
-
10.6
9.2
15
17
-
65.2
85.0
1
5
-
0.2
0.8
1
2
-
0.5
0.9
1
2
-
100
100
18
2.8
10
4.6
7
70.0
0.2
0.3
0.2
6
2.8
4
66.7
0
0
0
0
0
0
0
1.2
1
0
0
0
0.5
0
0
8
1
2
1
0
0
0
0
The mean value for nucleotide differences between paralogous sequences of
MCS_AP3 is 5.6%. The values range between 0.7% (A. canariense and S.
jaccardianum), 1.2% for Ai. laxum, and 9.3% for S. pubescens (table 12). For A.
canariense, Ai. laxum, and S. surculosum duplications were observed within one
subclade and should thus rather be defined as “paralogous alleles”. In contrast, the
sequences of S. pubescens cluster into two well defined clades. Paralogs of Aeonium
species have a mean value of 5.3% and the observed values were quite similar (table
12). For orthologs lower values were observed that range between 0.1% (A.
canariense and S. jaccardianum) and 1.3% (M. anagensis and S. caeruleum) with a
mean of 0.6%.
Focusing again only on paralogs, the mean value for replacements is 4.5% and
range between 0.9% for A. canariense and S. jaccardianum and 9% for S.
pubescens.
Calculations of replacements with quite different amino acids resulted in a mean
value of 53.6% and range between 30% (S. pubescens), 66.7% (A. canariense and
A. rubrolineatum) and 100% for A. canariense and S. jaccardianum.
72
Results
Table 12: Number and percentage of nucleotide differences, replacements, and quite different amino
acids (aa) between species-specific sequences of MCS_AP3 distinguished into orthologous and
paralogous sequences.
species
A. aureum
A. canariense
A. canariense
A. canariense
A. cuneatum
A. cuneatum
A. goochiae
A. nobile
A. rubrolineatum
A. rubrolineatum
A. saundersii
A. smithii
Ai. laxum
Ai. pachycaulon
M. anagensis
M. icterica
S. caeruleum
S. jaccardianum
S. jaccardianum
S. modestum
S. pubescens
S. pubescens
S. surculosum
S. surculosum
nucleotide
differences
no.
%
41
6.1
42
6.3
5
0.7
paralogs
replacements
no.
%
9
4.0
9
4.0
2
0.9
42
6.3
10
4.5
6
60.0
42
6.3
12
5.4
8
66.7
39
8
5.8
1.2
10
5
4.5
2.2
5
3
50.0
60.0
5
0.7
2
0.9
2
100
62
9.3
20
9.0
6
30
22
3.3
5
2.2
2
40
quite
different aa
no.
%
5
55.6
6
66.7
2
100
nucleotide
differences
no.
%
1
2
1
2
0
2
3
2
3
6
2
2
7
9
2
9
1
2
6
3
4
2
6
0.1
0.3
0.1
0.3
0.0
0.3
0.4
0.3
0.4
0.9
0.3
0.3
1.0
1.3
0.3
1.3
0.1
0.3
0.9
0.4
0.6
0.3
0.9
orthologs
replacements
no.
%
0
1
0
0
0
0
1
2
1
1
0
1
2
3
0
4
0
2
3
1
3
1
1
0.0
0.4
0.0
0.0
0.0
0.0
0.4
0.9
0.4
0.4
0.0
0.4
0.9
1.3
0.0
1.8
0.0
0.9
1.3
0.4
1.3
0.4
0.4
quite
different aa
no.
%
0
1
0
0
0
0
1
2
1
1
0
1
2
2
0
4
0
2
1
0
1
0
1
0
100
0
0
0
0
100
100
100
100
0
100
100
66.7
0
100
0
100
33.3
0
33.3
0
100
3.7. Relative Rate Tests
MCS_PEPC sequences revealed significant or highly significant rate differences for
all analyzed paralogs. In dependence of the outgroup paralogs of A. aureum, A.
nobile, A. rubrolineatum, A. smithii, and Ai. pachycaulon showed different rates. For
Ai. pachycaulon this pattern of sequence evolution is only weakly supported
compared to the Aeonium species.
For MCS_AP1 relative rate tests did never reveal any differences in the evolutionary
rate of the studied species and their paralogous gene copies.
For MCS_AP3 rate differences were found for some of the studied species and their
respective paralogous sequences. Several comparisons were made for A. cuneatum
and A. rubrolineatum and in some cases significant p-values were observed. Results
depend strongly on the selected outgroup species but were confirmed using different
73
Results
outgroup references. MCS_AP3 sequences of A. aureum, A. canariense, and A.
smithii did not indicate any differences in their evolutionary rate.
3.8. Ka/Ks-values
All estimated mean Ka/Ks-values were significantly lower than 1 and define purifying
selection. Mean values for the MCS species were highest for MCS_AP1 copy A and
lowest for MCS_PEPC copy B. Overall the lowest mean value was detected for the
sistergroup species of Sedum ser. Monanthoidea for MCS_PEPC and the highest for
the Sedum ser. Monanthoidea species (sistergroup) for MCS_AP1 (table 13).
Table 13: Mean Ka/Ks-values. Indication of the copies follows the phylograms in fig. 8-16.
copy A
copy B
Sedum_sister
MCS_PEPC
0.110 ± 0.101
0.106 ± 0.065
0.047 ± 0.034
MCS_AP1
0.412 ± 0.445
0.373 ± 0.185
0.417 ± 0.510
0.286 ± 0.020
MCS_AP3
0.195 ± 0.163
0.340 ± 0.479
0.092 ± 0.084
0.124 ± 0.115
Sedum_outgroup
Figures 18-28 summarize the Ka/Ks-values. Most individual Ka/Ks-values were
clearly below 1 but exception were detected for MCS_AP1 copy A, for MCS_AP1 of
the Sedum ser. Monanthoidea species (sistergroup), and MCS_AP3 copy B.
MCS_PEPC copy_B
1
0,75
0,5
0,25
0
0- 0.25- 0.5- 0.75- 1.0- 1.25- 1.5- 1.75- 2.0- 2.25- >2.5
0.25 0.5 0.75 1.0 1.25 1.5 1.75 2 2.25 2.5
1
0,75
0,5
0,25
0
0- 0.25- 0.5- 0.75- 1.0- 1.25- 1.5- 1.75- 2.0- 2.25- >2.5
0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
Ka/Ks
Ka/Ks
MCS_PEPC Sedum _sistergroup
proportion of comparisons
proportion of comparisons
proportion of comparisons
MCS_PEPC copy_A
1
0,75
0,5
0,25
0
0- 0.25- 0.5- 0.75- 1.0- 1.25- 1.5- 1.75- 2.0- 2.25- >2.5
0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
Ka/Ks
74
Results
MCS_AP1 copy_B
1
0,75
0,5
0,25
0
0- 0.25- 0.5- 0.75- 1.0- 1.25- 1.5- 1.75- 2.0- 2.25- >2.5
0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
proportion of comparisons
proportion of comparisons
MCS_AP1 copy_A
1
0,75
0,5
0,25
0
0- 0.25- 0.5- 0.75- 1.0- 1.25- 1.5- 1.75- 2.0- 2.25- >2.5
0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
Ka/Ks
Ka/Ks
MCS_AP1 Sedum _outgroup
1
0,75
0,5
0,25
0
0- 0.25- 0.5- 0.75- 1.0- 1.25- 1.5- 1.75- 2.0- 2.25- >2.5
0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
proportion of comparisons
proportion of comparisons
MCS_AP1 Sedum _sistergroup
1
0,75
0,5
0,25
0
0- 0.25- 0.5- 0.75- 1.0- 1.25- 1.5- 1.75- 2.0- 2.25- >2.5
0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
Ka/Ks
Ka/Ks
MCS_AP3 copy_B
1
0,75
0,5
0,25
0
0- 0.25- 0.5- 0.75- 1.0- 1.25- 1.5- 1.75- 2.0- 2.25- >2.5
0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
propotion of comparisons
proportion of comparisons
MCS_AP3 copy_A
1
0,75
0,5
0,25
0
0- 0.25- 0.5- 0.75- 1.0- 1.25- 1.5- 1.75- 2.0- 2.25- >2.5
0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
Ka/Ks
Ka/Ks
MCS_AP3 Sedum _outproup
1
0,75
0,5
0,25
0
0- 0.25- 0.5- 0.75- 1.0- 1.25- 1.5- 1.75- 2.0- 2.25- >2.5
0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
proportion of comparisons
proportion of comparisons
MCS_AP3 Sedum _sistergroup
1
0,75
0,5
0,25
0
0- 0.25- 0.5- 0.75- 1.0- 1.25- 1.5- 1.75- 2.0- 2.25- >2.5
0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
Ka/Ks
Ka/Ks
Fig. 18-28: Distribution of Ka/Ks-values for MCS_PEPC, MCS_AP1, and MCS_AP3 genes for the
MCS and the Sedum sister- as well as outgroup species. Indication of the copies follows the
phylograms in fig. 8-16.
T-tests compare for each gene the Ka/Ks-values of 1) gene copy A versus copy B of
the MCS species, 2) gene copies of the MCS vs. sistergroup and outgroup species,
respectively and 3) regulatory versus structural genes (table 14).
75
Results
Significant differences were observed for the two copies of MCS_AP3 where the
mean Ka/Ks-value of copy B is significantly higher than the one of copy A. For
MCS_PEPC and MCS_AP1 similar values were observed for each of the two copies
(table 13 and 14).
In general, the gene copies of the MCS species have higher mean Ka/Ks-values
compared to the Sedum specific copies. An exception is the copy of the sister
species (Sedum ser. Monanthoidea) for MCS_AP1 that has with 0.417 ± 0.510 a
higher mean value than MCS_AP1 copy A (0.412 ± 0.445) and copy B (0.373 ±
0.185); but the results are not significant (table 14).
The regulatory genes MCS_AP1 and MCS_AP3 have higher mean Ka/Ks-values as
the structural gene MCS_PEPC and all these relationships are significant (table 14).
Table 14: Comparison of Ka/Ks-values for the respective genes for 1) copy A and B of the MCS
species, 2) copies of the MCS vs. sister- and outgroup (OG) species, respectively, and 3) regulatory
vs. structural genes based on t-tests (n.s. = not significant, * significant (p < 0.05) and ** highly
significant (p < 0.01)).
MCS_
PEPC
_A
MCS_
PEPC
_B
MCS_
PEPC_A
-
n.s.
MCS_
PEPC_B
n.s.
-
MCS_
AP1_A
**
MCS_
AP1_B
MCS_
PEPC
_sister
MCS_
AP1
_A
MCS_
AP1
_B
**
**
**
MCS_
AP1_
sister
MCS_
AP1_
OG
MCS_
AP3
_A
MCS_
AP3
_B
MCS_
AP3_
sister
MCS_
AP3_
OG
**
**
**
**
**
**
**
**
-
n.s.
n.s.
n.s.
*
n.s.
**
**
n.s.
-
n.s.
**
**
n.s.
MCS_
AP3_A
**
**
*
**
-
**
**
*
MCS_
AP3_B
**
**
n.s.
n.s.
**
-
**
**
3.9. Selection pressure
No positive selection pressure was revealed for MCS_PEPC with the codon-based Ztest if the whole dataset is analyzed. Focusing only on the MCS species, positive
selection was detected between M_anagensis_303T and 305T. In general, purifying
selection is favored over strict neutrality and explains best the gene evolution.
For MCS_AP1, the null hypothesis of strict neutral evolution was rejected favoring
the alternative hypothesis of purifying selection. Positive selection was detected only
between S_jaccardianum_1 and S_surculosum_3 (only for p-distance setting).
76
Results
The analyzed sequences of MCS_AP3 showed in most cases purifying selection but
for several Aeonium sequences in subclade B the null hypothesis of strict neutrality
could not be rejected. In addition, positive selection was indicated between the
sequences of A. nobile and A_cuneatum_225, 26134, and 301b of the subclade B.
Focusing only on the MCS species positive selection was additionally detected
between M_anagensis_311T and 313T. Even if purifying selection is the main
selection pressure, positive and especially neutral selection cannot be ruled out.
Selection pressure acting on a particular codon can differ and thus, analyses were
done with Selecton to detect sites under positive selection. Sites under positive
selection were detected for MCS_PEPC, MCS_AP1, and MCS_AP3 but were only
significant for the two regulatory genes.
Only one site under positive selection could be found for MCS_PEPC, 37 amino
acids indicating positive selection for MCS_AP1, and three for MCS_AP3. For
MCS_AP3 all three amino acids were found within the MADS-box domain but must
be handle with caution because missing data at the beginning of several Sedum
sequences could influence the results. For MCS_AP1 17% of codons under positive
selection were observed. Three codons indicating positive selection were found
within the MADS-box (10.7%), six within the I-domain (14.6%), 10 within the Kdomain (14.7%), and 18 in the C-terminal domain (22.5%).
77
Discussion
4. Discussion
4.1. Phylogenetic reconstructions
Utilization of neutral markers for reconstructing phylogenetic relationships for species
evolving within fast radiations often results in poorly resolved topologies or
unresolved polytomies (Mes and ‘t Hart 1996, Baldwin et al. 1998). However,
accurate knowledge of these relationships is critical for studying processes of
speciation and evolution, and for the identification and proof of key innovations
(Hodges 1997, Jorgensen and Frydenberg 1999, Jorgensen and Olesen 2001).
Since low-copy genes can be involved in the determination of phenotypes and
adaptation, a nuclear gene phylogeny might answer questions about morphological
and physiological evolution influenced by the studied genes (Sang 2002). In addition,
low-copy nuclear genes may provide more information to resolve relationships and
thus, could overcome the problem of missing synapomorphic characters (reviewed in
Sang 2002 or Small et al. 2004). Therefore, numerous studies have attempted to
evaluate the use of low-copy nuclear genes to reconstruct phylogenies and to
estimate their evolutionary impact (e.g., Bailey and Doyle 1999, Barrier et al. 2001,
Malcomber 2002, Fan et al. 2004, Grob et al. 2004, Álvarez et al. 2005, Purugganan
and Robichaux 2005, Syring et al. 2005, Janssens et al. 2007, Tu et al. 2008).
In the present study homologs of PEPC, AP1, and AP3 genes were used to
reconstruct phylogenetic relationships within the island radiation of the
Macaronesian Crassulaceae Sempervivoideae (MCS). The results obtained from the
full-length genic region, and exons and introns only of the studied low-copy nuclear
genes, were very complex but some generalizations could be drawn especially by
focusing on the main copies and clades.
The MCS and Sedum species are well separated. The studied Sedum species could
be distinguished in the outgroup species S. caeruleum and S. pubescens and in the
sistergroup species of Sedum ser. Monanthoidea. The species of the MCS cluster
together in a main clade, and are separated into at least two subclades. The main
copy (subclade) comprises in all but one case all studied MCS species and a
separation between the three genera Aeonium, Aichryson, and Monanthes could be
observed. The obtained relationships both support and contradict relationships
inferred by other markers.
78
Discussion
Within Aeonium a well supported relationship was found between A. aureum, A.
saundersii, and A. smithii. This relationship is confirmed by morphological and
molecular data (Mes and ‘t Hart 1996, Jorgensen and Frydenberg 1999, Mort et al.
2002, reanalyzed nrDNA ITS) even if Liu (1989) and Mes (1995) separated the
species into different sections (see also table 15; appendix). These species share an
increased number of flower organs that is highest for A. aureum (28-32), followed by
A. saundersii (12-16), and A. smithii (8-12) compared to 6-11 for the remaining
Aeonium species. Further common features are their yellow flower color (Liu 1989)
and to some extent their CO2 fixation pathway. For A. aureum C3 fixation is reported;
A. saundersii and A. smithii have an intermediated pathway including C3 and CAM
activity (Tenhunen et al. 1982, Pilon-Smits et al. 1992, Mort et al. 2007).
Contradicting, Lösch (1990) reported strong CAM for A. smithii. Analysis of
MCS_AP1 places one additional species in this group: A. rubrolineatum. Like the
other species, A. rubrolineatum has a slightly increased number of flower organs (911), one of the main characters encoded by homologs of AP1. In addition, it shares
the yellow flower color with all species in the group and a subshrub habit with A.
saundersii and A. smithii (Liu 1989).
Another close relationship, confirmed by all studied genes, was detected between A.
canariense, A. cuneatum, and A. nobile. Analyses of both MCS_PEPC and
MCS_AP3 also include A. rubrolineatum in the group, contradicting the results of
MCS_AP1. These relationships are surprising and generally not supported by other
markers (Liu 1989, Mes 1995, Mes and ‘t Hart 1996, Jorgensen and Frydenberg
1999, Mort et al. 2002). Morphological data combine at least the herbaceous rosette
plants A. canariense and A. cuneatum. Both belong to sections Canariensia (Lems
1960, Mes 1995) and Patinaria (Liu 1989), respectively. A sister relationship was also
suggested by MCS_AP3 but is not confirmed using neutral molecular markers (Mes
and ‘t Hart 1996, Jorgensen and Frydenberg 1999, reanalyzed nrITS). That
MCS_AP3 combines both species could be due to the general similar habit and
especially because of similar petals (8 to 10), a character encoded by AP3 homologs.
In the MCS_PEPC phylograms a sister relationship between A. cuneatum and A.
rubrolineatum could be detected. For both species C3 fixation prevails and weak
CAM activity is possible. For MCS_AP1, A. cuneatum is basal to the sister
relationship of A. canariense and A. nobile. Mes and ‘t Hart (1996) showed that A.
cuneatum is closely related to A. nobile and A. rubrolineatum based on morphological
79
Discussion
data and restriction site mutations of cpDNA, respectively. However, the latter
relationship was not confirmed in the respective combined dataset (Mes and ‘t Hart
1996). No other markers support the relationship between A. nobile and A.
canariense as suggested by MCS_PEPC and MCS_AP1, or that between A. nobile
and A. rubrolineatum (MCS_AP3). Lems (1960) and Liu (1989) treated A. nobile as
monotype of sect. Megalonium, and Mes (1995) ordered it to sect. Leuconium not
closely related to any of the other above mentioned species. Mort et al. (2002) sorted
it together with species such as A. haworthii, A. pseudourbicum or A. urbicum in
clade 4. In contrast, A. rubrolineatum belongs to sect. Holochrysa (Lems 1960), Liu
(1989) and Mes (1995) ordered it to sect. Aeonium, and Mort et al. (2002) in clade 3
(see also table 15 and fig. 2).
The close relationship of these four species is also surprising in the present study.
Focusing on relationships inferred by MCS_PEPC data, the strong CAM species A.
nobile is closely related to species where C3 prevails (Lösch 1990). Concerning the
number of flower organs the species are not very different (between 7 and 11; Liu
1989) and a close relationship might be assumed. In general the species show no
strong morphological agreement. Species with pale yellow green flowers (A.
canariense) are combined with yellow ones (A. cuneatum), yellow with red stripes (A.
rubrolineatum), and dark red (A. nobile). Aeonium canariense and A. cuneatum are
herbaceous species occurring in the laurel forest on Tenerife whereas A. nobile and
A. rubrolineatum are subshrubs on La Palma and La Gomera (Liu 1989).
In all inferred phylograms and focusing on the main copy, A. goochiae is basal to all
other Aeonium species. Its position is supported by morphological data (Liu 1989)
and confirmed by nrITS but not when cpDNA and nrITS were combined (Mort et al.
2002). Jorgensen and Frydenberg (1999) reported a close relationship of A.
goochiae with A. lindleyi and A. viscatum unresolved in a larger clade. Mort et al.
(2002) confirmed the sister relationship to A. lindleyi in the combined cpDNA/nrITS
phylogram. They ordered both species together with A. aureum and A. saundersii to
clade 2. However, a relationship to both of the latter species is not detected in the
present study. Liu (1989) described A. goochiae as a distinct species within sect.
Petrothamnium. It can be distinguished by its pinkish flowers and very thin leaves
whereas other characters suggest a distant relationship to A. lindleyi. Aeonium
goochiae was suspected to connect the genus Aeonium with the genus Aichryson
80
Discussion
(Lems 1960, Liu 1989). The present data clearly support this hypothesis given the
basal position of A. goochiae in all phylograms.
Resolving relationships of Aichryson and Monanthes with the present data is
difficult since only a low number of species was included. However, one noteworthy
remark has to be made; as expected the Monanthes species show mostly diverging
phylogenetic patterns. Monanthes anagensis is mainly found in sister relationships to
the Aeonium species. A close relationship to Ai. pachycaulon is only resolved in the
MCS_PEPC phylograms, but missing representatives of M. anagensis for the main
copy may cover the pattern (e.g., fig. 9).
In contrast, M. icterica always shows a sister relationship to the Aichryson species
(MCS_PEPC) or is even imbedded and intermixed with them (MCS_AP3). For
MCS_AP1 missing amplification of Aichryson makes a deduction of the relationships
difficult. Most times the positions of both Monanthes are unresolved but they do not
cluster as sister species. The position of M. icterica is traditionally highly debated.
Nyffeler (1995) emphasis the Aichryson-like habit of M. icterica but the enlarged
nectariferous scales and bladder cell-idioblasts order it unambiguously to Monanthes.
Mes et al. (1997) discussed a sister relationship between M. icterica and Aichryson
based on morphological and molecular data. Mort et al. (2002) reported, based on
nrITS and combined cpDNA/nrITS data, that M. icterica is basal and sister to
Aeonium and to the perennial Monanthes species. Chloroplast markers resolve M.
icterica in a sister relationship to Aichryson. Reanalysis of the nrITS dataset also
suggests this sister relationship. Based on the present study, the genus Monanthes is
only monophyletic if M. icterica is separated from the perennial species.
As discussed above the uniqueness of M. icterica was shown before. The annual life
form and the deviating chromosome base number of x = 10 distinguish this species.
RAPD patterns and nrITS data differ significantly and support the assumption that M.
icterica is genetically only distantly related to the perennial Monanthes (Mes et al.
1997). No hybrids have been reported involving M. icterica. Even if this phenomenon
is assumed given the diverging chromosome base number, it is still interesting since
hybridization in the genus is considered to be frequent and easily possible (Nyffeler
1995). The inclusion of M. icterica in the genus Monanthes based mainly on flower
morphology but multiple origin of flowers with large nectariferous scales could be
considered like in S. surculosum or S. napiferum (for reference see Mes et al. 1997).
81
Discussion
Nevertheless, with respect to the studied low-copy nuclear genes M. icterica is clearly
closely related to the Aichryson species.
Relationships and positions of the Sedum species are mainly confirmed with the
analyzed low-copy nuclear genes as compared to existing morphological,
chromosomal, and neutral molecular markers. Sedum jaccardianum, S. modestum,
and S. surculosum represent the species of Sedum ser. Monanthoidea and cluster
together as sistergroup species of the MCS. As predicted by other markers, S.
jaccardianum and S. surculosum are thereby sister to each other and S. modestum is
sister and basal to them. However, the pattern of S. surculosum for all studied genes
is remarkable. For example, two S. surculosum sequences are sister to S.
jaccardianum
for
the
main
copy
of
MCS_PEPC
whereas
the
sequence
S_surculosum_2c4391 is basal to the remaining species in the exon phylogram (e.g.,
fig. 9). In the exon phylograms of MCS_AP1 S_surculosum_3 is mixed with S.
jaccardianum, whereas the remaining sequences are sister to S. jaccardianum (e.g.,
fig. 12). Focusing on the coding region of MCS_AP3 (e.g., fig. 15), S. jaccardianum
and S_surculosum_1 and 8 are sister to each other. Sister to them are the remaining
S. surculosum sequences (2, 3, 6) and these relationships were confirmed by the fulllength data. A possible explanation could be hybridization, which is generally easy
between Sedum species (for reference see van Ham and ‘t Hart 1998). Moreover, S.
surculosum is tetraploid which may hint at an ancient introgression event via
allopolyploidization. Given that in all cases some gene copies show a clear
relationship to S. jaccardianum, this species may be involved in the evolution of S.
surculosum.
Positions and relationships of both outgroup species, S. caeruleum and S.
pubescens, were difficult to resolve because of partly missing amplification. However,
they are basal to the remaining species and related to the analyzed species of
Sedum ser. Monanthoidea.
In general the three low-copy nuclear genes support similar species groups and
relationships; in several cases the species phylogeny based on morphological and
other molecular markers is contrasted. Jorgensen (2002) discussed that radiation
and morphological differentiation of Aeonium species is the result of adaptation to
ecological conditions and not a reflection of the species phylogeny. Traits such as
growth-form, plant height, leaf form, and inflorescence length (maybe partly
82
Discussion
influenced by MCS_AP1) vary and are adaptive in response to ecological conditions.
Thus, ecology may better predict the variance in a number of traits. If the characters
are adaptive, selection may act on the genes that are involved in these above
mentioned characters and would not evolve neutrally as phylogenetic markers like
nrITS and cpDNA do (see also Hodges and Arnold 1994b). Selection pressure may
fix mutations, which can be used as information for phylogenetic analyses (reviewed
in Sang 2002). Predicted by Jorgensen and Olesen (2001), application of markers
that may be related to these fitness factors could vary considerably and may better
resolve relationships between the species (see also Mes and ‘t Hart 1996).
On the other hand, pubescence of floral organs (maybe related to CAM and thus,
MCS_PEPC), flower size (petal length; MCS_AP3), number of ovaries, stamens per
flower, and number of petals (the latter both maybe partly influenced by MCS_AP3)
correlate with phylogeny (Jorgensen and Olesen 2001) suggesting that these genes
may evolve neutrally and, thus, reflect the species phylogeny.
The assumption that MCS_AP1 may have evolved in adaptation but MCS_PEPC and
MCS_AP3 under neutral selection could explain differences in the topologies inferred
by these different genes. At least for MCS_AP1 sites under positive selection could
be found, even if patterns for MCS_AP1 are quite diverse and difficult to explain (see
below). On the other hand, for MCS_AP3 evidence for neutral evolution was
obtained. However, in general, strong purifying selection was detected acting on all
analyzed genes and copies; substitution rates are restricted to avoid deleterious
mutations within the coding sequence (Sang 2002, Janssens et al. 2007). Thus,
conflicts between species and/or the respective gene topologies could maybe be
explained by conservation and selection. Whereas conflicts between the genealogies
may arise because different mutations are selectively fixed in the respective genes of
the species, conflicts to the species phylogeny may arise because of the random
accumulation of mutations in neutrally evolving nrITS and cpDNA regions.
Nevertheless, the obtained results must be handled with caution. The greatest
challenge of working with low-copy nuclear genes is the differentiation between
orthologous and paralogous gene copies since only the former resolve the species
phylogeny (Litt and Irish 2003). Preliminary results of the present study suggest that
the relationships inferred only from orthologs are similar to those of the whole dataset
including paralogs and orthologs. However, inclusion of more species, improvement
83
Discussion
of alignments comprising only orthologs, and subsequently further analyses are
necessary to confirm the results and to enable more reliable conclusions.
Phylogenies based on low-copy nuclear genes have confirmed in several studies the
relationships found for neutral evolving marker and/or other regulatory genes. Bailey
and Doyle (1999) confirmed the species phylogeny of several Brassicaceae species
using an intron region of the PISTILLATA gene. Koch et al. (2001) confirmed with
their phylogeny based on chalcone synthase (CHS) and AP3 gene promoter
sequences
relationships
deduced
by
a
combined
matK/CHS
analysis
for
Brassicaceae species. Analysis of Atmyb2 flanking sequence data resulted in highly
concordant phylogenies for Arabidopsis compared to a nrITS phylogeny (Beck et al.
2007). Durbin et al. (2003) found that relationships for Ipomoea based on CHS genes
were quite similar to results based on nrITS and waxy gene sequences. The
phylogeny of the tribe Andropogoneae (Poaceae) based on FLORICAULA/LEAFY
genes agreed largely with previously published phylogenies using other nuclear
genes (Bomblies and Doebley 2005) and the same was true for the genus
Amorphophallus (Araceae) using the FLORICAULA/LEAFY second intron (Grob et al.
2004). Zhang et al. (2008) confirmed the species relationships of Cornus with their
genealogy of a PISTILLATA-like gene copy and Fan et al. (2004) with the Myc-like
anthocyanin regulatory gene. Álvarez et al. (2005) resolved congruent clades
comparing previous phylogenies of Gossypium with one inferred by three low-copy
nuclear genes and Janssens et al. (2007) found highly congruent topologies for
Impatiens comparing the AP3/DEF K-domain with atpB-rbcL data.
In contrast to the multiplicity of the above mentioned studies is, e.g., the study of
Malcomber (2002). He found that all used markers and methods were insufficient in
constructing relationships of Rubiaceae and significant incongruent data partitions
between nrITS and the PEP-large intron were revealed by a PHT test. Syring et al.
(2005) found for four low-copy nuclear genes in Pinus that the individual loci do not
uniformly support either the nrITS or cpDNA hypotheses. In some cases the low-copy
nuclear genes produced even unique topologies. Fortune et al. (2007) detected
incongruences between species and gene trees in the waxy gene phylogeny of the
hexaploid Spartina species. Also phylogenetic analysis performed either on one or
the other duplicated waxy gene copy did not alter the topologies. Tu et al. (2008)
detected incongruences between cpDNA, waxy, and LEAFY gene phylogenies in the
84
Discussion
genus Nolana, which may have been due to reticulate evolution, lineage sorting, and
duplication of the waxy gene.
Just as the entire low-copy nuclear gene itself can alter the obtained topologies, the
different gene regions may reveal conflicting signals as well. Exons are quite
conserved and may not provide enough information to resolve relationships within
closely related species. In contrast, introns are attractive for evaluating relationships
among closely related taxa since they are diverse, fix mutations, and diverge at
relatively rapid rates (Syring et al. 2005). Resolution power of full-length datasets
may be simply explained by their length and amount of information that they provide
even if they combine conserved and quite diverse data partitions.
In the present study, exon and intron regions were combined and PHT tests suggest
significant differences between them for MCS_AP1 (p = 0.02). Nevertheless, both
datasets were combined due to the lack of well supported differences between the
topologies inferred by the two independent datasets (see Grob et al. 2004). The fulllength phylograms provide a better resolution and higher support values compared to
exon phylograms, maybe because more phylogenetic information was provided.
Conflicts were observed for the position of A_saundersii_33637. This sequence is
basal to the main clade and the Sedum species in the full-length phylograms (e.g.,
fig. 11) but sister to the analyzed Sedum species in the exon phylograms (e.g., fig.
12). Aeonium_saundersii_33637 is an additional sequence and maybe gene copy of
MCS_AP1. The observed conflicting position may be due to a diverging function.
Therefore, different selection pressure and fixation may have acted on this gene copy
and resulted in another position in the exon phylograms. Further conflicts were found
for S_surculosum_3 that is intermixed with S. jaccardianum inferred by the exons but
basal to the S. surculosum sequences in full-length phylograms. Hybridization and
gene duplication could be reasons for the observed pattern. Relationships inferred by
introns are similar to the full-length phylograms.
PHT tests revealed no differences between exon and intron regions for MCS_PEPC
and MCS_AP3 suggesting the same evolutionary rate for the data partitions.
Relationships inferred in full-length and exon phylograms of MCS_PEPC were
identical and comparisons to intron phylograms were difficult because of missing
resolution.
85
Discussion
For MCS_AP3 minor differences could be observed between the three gene regions.
Phylograms based on full-length data provided the best resolution. Conflicting
positions were detected for the Sedum species. In the exon phylograms S.
caeruleum is basal to the species of Sedum ser. Monanthoidea and to the main
clade. In contrast, S. caeruleum is basal and sister to the species of Sedum ser.
Monanthoidea in the full-length dataset and here, all Sedum species form a
separated Sedum specific subclade. Exon phylograms based on conserved regions
that were likely influenced by the gene functions may explain the classification of S.
caeruleum. In the full-length datasets more information was used and the diverse
intron data partitions may partly conflict the conserved exon partition. Intron
phylograms show similar topologies as full-length phylograms except within subclade
B where A. rubrolineatum and A. cuneatum showed conflicting positions (e.g., fig.
13).
In general, only few conflicts were found between intron, exon, and full-length
phylograms. There are several explanations why intron and exon regions could
evolve with similar rates. Low-copy nuclear genes could evolve in a neutral fashion
and similar selection pressure and mutation rates could be assumed for the data
partitions (see Beck et al. 2007). In accordance to Jorgensen and Olesen (2001) and
Jorgensen (2002) neutral selection pressure may be assumed for MCS_PEPC and
MCS_AP3 and is at least partly confirmed for MCS_AP3 in the present study.
Furthermore, conserved structures in introns, like regulatory elements, could
influence the divergence of this region (Bailey and Doyle 1999, Small et al. 2004) as
well as hitchhiking where the evolution of intron regions follows the evolution of exon
regions or concerted evolution may influence the diversification of intron and exon
regions (e.g., Small and Wendel 2002).
4.2. Gene duplications
Polyploidy is a common phenomenon in plant evolution and duplication of genes in a
genome is the most obvious molecular consequence (Kramer et al. 1998, Blanc and
Wolfe 2004, Moore and Purugganan 2005). Polyploid species appear to vary from
their diploid progenitors in a variety of ecologically important traits and genome
duplication may provide a molecular mechanism for ecological diversification
(Lawton-Rauh et al. 2003).
86
Discussion
Gene duplication can also arise independently and randomly within species. MADSbox and PEPC genes are represented in gene families and therefore the studied lowcopy nuclear genes are well-known members of multicopy gene families. Gene
duplication could lead to increased diversity and functional innovation; recent studies
described diverse potential outcomes: pseudogenization, functional redundancy,
subfunctionalization, neofunctionalization or simple gene loss (Lynch and Conery
2000, Zhang 2003, Moore and Purugganan 2005).
Summarizing the results of the present study, gene duplications could be observed
for all three analyzed genes and for different species. For A. aureum and A. smithii
gene duplications were detected for all three analyzed genes. Other species such as
A. nobile, A. rubrolineatum, and A. saundersii only show duplications for one or two
of the genes. In addition, A. goochiae, both Monanthes, and Ai. laxum did not show
any gene duplication. Aichryson pachycaulon showed duplication for MCS_PEPC but
being tetraploid, detection of duplicates is likely. In contrast, A. goochiae or M.
anagensis, tetra- or even hexaploid species (Mes et al. 1997, Jorgensen and Olesen
2001) never showed any duplicates. Subclade specific gene duplications for the
polyploid MCS – especially Aeonium – was a common feature, whereas the diploid
Sedum species showed duplications only within the Sedum specific clade. Thus,
whereas for the MCS species at least two subclades comprising each orthologous
gene copies could be obtained, single duplications that result in paralogous gene
copies were observed for the Sedum species and additionally for some MCS species.
Gene duplications within MCS species could be the result of two different events.
Island colonization was connected with polyploidization and the duplication of the
genome, and thus genes, is the most obvious consequence (Blanc and Wolfe 2004).
On the other hand duplicates may have arisen randomly within species by single,
independent duplication of a gene. In the present study both outcomes are
suggested by the obtained topologies. The MCS_PEPC exon phylograms suggest
that the duplication was the consequence of polyploidization. Two subclade specific
gene copies were detected for the MCS species whereas only one CAM-specific
gene copy was found for the Sedum species. Both MCS copies are under strong
purifying selection, seem to be functional, and show so far no evidence for
neofunctionalization. The main copy for the MCS species is copy B and five species
exhibit the duplicated gene copy A (e.g., fig. 9). All species analyzed contain the
87
Discussion
main copy B except M. anagensis which possesses only copy A. If redundant copies
and thus gene expression is available, random silencing and/or loss of one copy can
be tolerated in a species (Zhang 2003). Given that M. anagensis is tetraploid and of
potential hybrid origin (Mort et al. 2002) additional copies should be detectable.
However, if PCR failure is excluded, it must be assumed that one copy of PEPC is
enough to fulfill the gene function and other copies were lost.
For A. nobile and A. smithii both MCS_PEPC copies were detected and the species
are known for strong CAM activity (Lösch 1990). For these species, a duplication of
the gene or retaining both copies may be advantageous. Harder to explain is the
observed duplication for A. aureum and A. rubrolineatum because both are C3
species with weak CAM activity. On the other hand, this observation may support the
hypothesis that the duplicates are rather randomly retained or lost than gained. Since
both gene copies are also under strong purifying selection further support for
retention instead of gain is obtained. It is also noteworthy that A. aureum shows gene
duplications for all three analyzed genes (see also below). Aichryson pachycaulon is
another C3 species that possesses both copies. In addition to polyploidization the
tetraploid status (Uhl 1961) could be an explanation and that the duplicates may
have arisen by hybridization or allopolyploidization. Still, the ordering of the
duplicated copies in the two specific orthologous subclades suggests that
polyploidization may explain this pattern better.
For MCS_AP1 two main copies and two additional copies, for A. aureum and A.
saundersii, were detected for the MCS species. For Sedum two copies were found;
one copy being specific for the sistergroup species of Sedum ser. Monanthoidea and
one specific for the outgroup species S. caeruleum. Concerning the gene
duplications, the observed relationships seem to support both hypotheses: random
and independent duplication within species and duplication as consequence of
polyploidization. It could be imagined that the two copies in the main clade are the
consequence of polyploidization whereas the additional copies of A. aureum and A.
saundersii may result from single gene duplications. That MCS_AP1 does not resolve
the MCS species as monophyletic may support this suggestion. However, AP1 is part
of an extended gene family (Irish and Litt 2005) and AP1-like genes are not restricted
to flower organs since they play broad roles in both vegetative and reproductive
development (Shan et al. 2007). In the present study, amplification was based on
genomic DNA and all potential copies could be amplified without restriction to specific
88
Discussion
tissues. Additional copies could just represent different members of the gene family
or tissue specific copies. Nevertheless, Blast and NJ analyses cluster the additional
copies of A. aureum and A. saundersii with the other studied MCS_AP1 sequences.
Intensive estimation of sequence similarity, phylogenetic reconstruction, studies of
cDNA sequences, predictions of expression patterns (paralogs may differ in timing or
tissue), southern blotting, and genetic mapping would be powerful tools to
understand the pattern better and in more detail and should be included in future
studies. However, it is important to remember that MCS_AP1 amplification for
Aichryson and S. pubescens failed and may be an indication that the MCS_AP1
gene family could be diverse and exist in several copies in the analyzed species.
Also the specific gene copies of Sedum ser. Monanthoidea and S. caeruleum are in a
derived relationship to each other and may represent different gene copies of
MCS_AP1 (e.g., fig. 12).
Focusing on the additional copies, whereas A. aureum shows duplications for all
three analyzed genes, MCS_AP1 is the only gene duplication detected for A.
saundersii. This duplication is not subclade specific suggesting that maybe a derived
copy was amplified. The AP1 gene family is known for frequent major gene
duplications such as AP1 and CAL in Arabidopsis (Shan et al. 2007). In addition,
recent duplications within one or few closely related species are extremely common.
New genes are frequently recruited in the genome, however it is not clear whether
recent duplicates will be functionally fixed in genomes or will eventually become
pseudogenes (Shan et al. 2007). This phenomenon is also suggested in the present
study and will be discussed in detail below.
Altogether, all observations – also the results of MCS_AP3 with subclade specific
duplications and an accelerated number of gene copies for the MCS compared to the
Sedum species – rather suggest that the duplicated subclade specific gene copies
may be the result or consequence of the polyploidization event that was connected
with island colonization. This observation is in line with Barrier et al. (2001).
Consistent with the allopolyploid origin of the studied HSA species (Barrier et al.
1999), these contain mostly two copies of the studied genes ASAP1, ASAP3, and
ASCAB9 whereas their recent common ancestor, the diploid North American
Tarweeds, only possess one copy. Barrier et al. (2001) used copy specific primers to
amplify the different target gene copies. A disadvantage of the present study is that
no copy specific primers were used. Thus it is not possible to determine if missing
89
Discussion
gene copies are the result of gene loss or PCR failure. However, in the study of
Lawton-Rauh et al. (2003), a corresponding population based study to the one of
Barrier et al. (2001), extensive tests revealed missing gene copies for some
individuals that seemed to be deleted rather than not amplified despite evidence that
both copies were expressed in the species.
Still, in the present study gene duplication could also have occurred later in the
speciation process; after the split of Aichryson and Monanthes or after the separation
of A. goochiae and A. lindleyi (compare indicated nodes 1 and 2 in fig. 46; appendix).
However, chromosome base numbers do not support these hypotheses since they
do not diverge for the species of Aeonium and Monanthes and given that also the
Aichryson species have accelerated chromosome base numbers connected with
polyploidization. Thus, random loss of gene copies in the polyploid species is a more
likely explanation for the pattern of missing duplicated gene copies for several
species, if PCR failure is excluded. In addition, the observed pattern of Ai.
pachycaulon and M. anagensis for MCS_PEPC highly suggests that polyploidization
may explain the general pattern of the duplicated genes. Independent single gene
duplication or strong pseudogenization may play an additional role for MCS_AP1;
however, more species have to be included in further studies to answer the question
what triggers the gene duplication.
So far,
all gene copies
show purifying selection
and no
evidence for
neofunctionalization. Single and independent duplication of each gene in the
respective species is also rather unlikely since there was only a very weak trend that
duplicated genes were connected with function. For example A. nobile, the species
with the strongest CAM activity (Lösch 1990), had two MCS_PEPC copies that
showed allelic differences. For MCS_AP1 such evidence could be found for A.
saundersii which has a accelerated number of flower organs and exhibits duplication
only for this gene. Nevertheless, this gene duplication is not subclade specific and
until now non-, neo-, or subfunctionalization could not be ruled out. Further studies
have to be done to verify the results.
If genes were duplicated because of polyploidization one may assume that
phylogenetic relationships within orthologs and between paralogs would be the
same. Unfortunately, amplification of both copies was not successful to the same
degree for both copies. Independent loss of one gene copy and/or different or
preferential amplification may obscure patterns and make interpretation difficult (see
90
Discussion
also, e.g., Baum et al. 2005 or Fortune et al. 2007). If missing amplification was the
problem, gene copy specific primers or more extensive amplification, screening, and
sequencing of clones may help. Mort and Crawford (2004) suggest to screen at least
five clones to detect the potential gene copies of low-copy nuclear genes in a
species. In addition, copy specific restriction enzyme digests might help to detect if
different copies were amplified but by chance not cloned and sequenced.
Nevertheless, Kramer et al. (1998) stated that if orthologs of both duplicated products
could be found in more than one species, the duplication event must have occurred
before the last common ancestor of the species in question splits. In the present
study this would suggest duplications at the base of the MCS species, before the
three genera separated. The basal duplication implies a possible connection with
polyploidization and island colonization.
Still, random duplication may be an explanation for the observed duplicates. To prove
if gene duplications in the present study are the consequence of single independent
events or of polyploidization, more nuclear coding genes should be studied. Both,
PEPC and the MADS-box genes are encoded by multigene families and duplicated
within the angiosperms (Gehrig et al. 1995, Lawton-Rauh 2003). Other floral
regulatory genes appear to be more conservative in copy number. If whole genome
duplication was the driving force for gene duplication in the MCS, paralogs of those
single-copy nuclear genes should be found in a corresponding study. For example,
Aagaard et al. (2005) observed a gene duplication for FLORICAULA/LEAFY within
the Lamiales even though this MADS-box gene was thought single-copy in all diploid
angiosperms, and showing duplication only in tetraploid species. They discussed that
their resolved pattern was consistent with an ancient duplication. Classification of the
analyzed copies into gene clades and the same relative position of duplicated genes
in the Lamiales phylogeny support a whole genome duplication theory.
For MCS_PEPC single duplications could be found for Sedum species. A
duplication within the main clade was observed for the tetraploid S. surculosum that
may explain this observed duplication. In addition, a clear separation between four
sequences of S. pubescens and S. surculosum and all other Sedum and MCS
sequences was observed in the genealogy and supported by a Blast analysis.
Furthermore, the NJ analysis (fig. 45; appendix) reveals for nearly all sequences a
close relationship to the mRNA sequences PEPC Kb1 (X87818) or Kb2 (X87819) of
91
Discussion
K. blossfeldiana (Gehrig et al. 1995). In contrast, the four unique Sedum clones are
closely related to PEPC Kb3 (X87820) or Kb4 (X87821). Gehrig et al. (1995) defined
these isogenes. Kb1, Kb2, Kb3, and Kb4 form two gene pairs, Kb1/Kb2 and
Kb3/Kb4, with 95-98% sequence homology within the pairs and only 75% between
them. The orthologous sequence pair of Kb1/Kb2 is attributed to the CAM state of
Mesembryanthemum crystallinum whereas Kb3 and Kb4 are more closely related to
C3 specific PEPC isoforms. Thus, the unique pattern in the present study revealed a
separation between C3 and CAM specific PEPC gene copies. Most analyzed species
possess the expected CAM specific gene copy whereas copies specific for the C3
state were observed for two Sedum species. C3 fixation prevails in Sedum and it is
also the ancestral state for the evolution of the CAM biosynthetic pathway in the MCS
species (Pilon-Smits et al. 1992, Mort et al. 2007).
As mentioned, two main and two additional copies were detected for MCS_AP1. One
copy was specific for Aeonium whereas the other also comprises the studied
Monanthes species.
The most complex pattern could be detected for A. aureum. Four different gene
copies were observed even though several were subsequently excluded because of
premature stops. However, proteins that arise from sequences with premature stops
may have an influence in evolution and should be considered. Aeonium aureum is
the species with the most derived phenotype concerning the impact of AP1
homologs. It has the highest number of sepals and petals representing partly one of
the main functions encoded by AP1. Also A. saundersii, exhibiting 12-16merous
flowers, shows a gene duplication. But this duplication was not subclade specific and
there is no knowledge regarding the function of this additional copy. Within the MCS
subclade gene duplications were observed for A. canariense, A. cuneatum, and A.
nobile that have 7-10merous flowers. Aeonium rubrolineatum, with 9-11 flower
organs, exhibited no duplicated gene copies and contrast the pattern. Other species
like A. goochiae, M. anagensis, and M. icterica without duplicates have between 6
and 8 flower organs. If gene duplications within the main clade are indeed the result
of polyploidization, independent gene loss in species such as A. goochiae, A.
rubrolineatum or the Monanthes species with a low number of flower organs could be
assumed; species with derived numbers of flower organs, like A. aureum and A.
saundersii, simply retained duplicated copies. In general, Lynch and Conery (2000)
92
Discussion
discussed that 30-50% of duplicated genes were preserved over periods of 10-100
million years following polyploidization. This number is exceptionally high and indeed
the majority of duplicated genes, even without polyploidization, will be silenced rather
than preserved and therefore are lost. The percentage of retained or lost duplicates
varies between species depending on functions (Moore and Purugganan 2005).
For MCS_AP3 MCS species showed subclade specific duplications compared to
Sedum but a unique pattern could be found for S. pubescens. Two subgroups of
sequences were detected, one positioned unresolved at the base of the phylogram
and the other basal to all remaining analyzed sequences (e.g., fig. 15). It has to be
assumed that this gene duplication arose within the species resulting in paralogous
gene copies. This assumption is supported by the number of observed nucleotide
differences (compare, e.g., studies of Baum et al. 2005, Fortune et al. 2007, Shan et
al. 2007, Zhang et al. 2008) and the use of varying outgroups to root the obtained
phylograms. Both copies showed significantly different Ka/Ks-values (data not
shown). However the duplication does not correlate with the number of flower organs,
and thus partly function of AP3 homologs, since S. pubescens has 5-6merous
flowers in agreement with the other Sedum species. Nevertheless, one has to keep in
mind that two different primer combinations were used to amplify homologs of
MCS_AP3 in Sedum. For S. pubescens, sequences amplified with one or the other
primer combination were clearly separated. However, this pattern was not found for
the other Sedum species; their species-specific sequences cluster together.
Nucleotide differences between these sequences were not primer specific and in the
range of that for orthologous gene copies. Only S. surculosum showed a unique
pattern and it was also the only studied Sedum species that had duplications in all
three analyzed gene regions. Since S. surculosum is tetraploid additional gene
copies were expected.
Homologs of AP3 control petal and sepal organ identity and maybe size (Juenger et
al. 2000, Lohmann and Weigel 2002). The number of flower organs for species with
duplicated gene copies range between 7 and 32. Comparable to MCS_AP1, for
species with the lowest number of flower organs no duplications were observed like
for A. goochiae (7-8 petals), A. nobile (7-9), M. anagensis (6-8), and M. icterica (6-7).
Again this pattern is not consistent because, e.g., A. saundersii has no duplicated
gene copies but an accelerated number of flower organs (12-16). The high number of
93
Discussion
analyzed clones, ten, makes it unlikely that additional copies were simply not
detected. Nevertheless, the possibility of preferential amplification of a specific gene
copy over other copies could not be ruled out. There were no specific nucleotide
differences in the primer binding site between paralogs but primer affinity could also
be related to differences in primary or secondary structure of the DNA at the potential
target sites (Tu et al. 2008). This potential problem could be solved by designing
gene copy specific primers for amplification, where one primer is designed to the
conserved region and one to the variable region.
The subclade specific duplication of the MCS species with one main copy comprises
all analyzed MCS species and a gene duplication for several Aeonium species
(subclade A; e.g., fig. 15) was strongly support by Blast analyses. While the main
copy showed the highest similarity to AP3-like sequence of Kalanchoe, the second
copy showed the highest similarity to a DEF-like gene copy of P. groenlandica. Both
copies are under strong purifying selection, indicating that they seem to be functional
and no hint of neofunctionalization or pseudogenization could be detected. Kramer et
al. (1998) showed that DEF, as AP3 ortholog in A. majus, was able to largely replace
the endogenous AP3 function in A. thaliana. Thus it can be assumed that both gene
copies fulfill their task as B-function MADS-box genes. The observed pattern
suggests that divergence and subfunctionalization increase the chance of retention.
Even if these phenomena could not be unequivocally proved with the present data,
retention of duplicates is often connected with functional divergence. Different
expression and/or subfunctionalization may be assumed for the MCS_AP3 gene
copies. Given that significant differences in the calculated mean Ka/Ks-values and in
the evolutionary rates were observed (discussed below) the hypothesis of divergent
functionalization is supported. Also the retention of one copy in all analyzed species
and the loss of the other copy in a subgroup of the studied species suggest that
different evolutionary fates may have acted on the gene copies (see Zhang et al.
2008).
Functional diversification of duplicated genes is an important feature in the long-term
evolution of polyploids. Blanc and Wolfe (2004) showed that 57% of newly duplicated
and 73% of old duplicated gene pairs of A. thaliana had divergent expression.
Likewise Moore and Purugganan (2005) discussed that 57% of the duplicated genes
had divergent expression patterns whereas only 20% had asymmetric rates of protein
94
Discussion
evolution. Thus, a large majority of polyploid-derived duplicates in Arabidopsis, which
remain duplicated, acquired divergent functions and became specialized.
One of the outcomes of the classic model for gene duplication is gene loss
associated with pseudogenization (Ohno 1970). Since amplification in the present
study was based on genomic DNA, it is not possible to distinguish between functional
and nonfunctional gene copies but hints for pseudogenization were detected in some
cases. Frameshift mutations and subsequent premature stops in the coding region
may be signs of pseudogenization and were revealed in the present study for
MCS_PEPC and MCS_AP3 with weak patterns and a strong pattern for MCS_AP1.
Several MCS_AP1 sequences showed specific features and studying them could
help to distinguish between potential outcomes of the gene duplication even if these
sequences were finally excluded. Therefore, sequences of A. aureum with premature
stops and two unique sequences, A_goochiae_23637 and A_smithii_23637, were
included in a phylogenetic analysis that revealed the following pattern (fig. 44;
appendix). Within the main clade, A. aureum sequences with premature stops
clustered with two sequences of A. aureum without stops. The additional sequences
A_aureum_13637 and 1b3637 clustered together with the unique sequence of
A_goochiae_23637 that had an indel of 1 bp in the exon position 528 resulting in a
frameshift mutation and two premature stop codons. In addition, A_goochiae_23637
had a deletion between base pair 96 and 265 resulting in 56 missing amino acids
within the K-domain. Since the K-domain is quite conserved and is involved in
protein-protein interactions it is unlikely that the hypothetical protein of this particular
sequence would be functional. This assumption is in line with Durbin et al. (2003).
They observed early stops for cDNA sequences of Ipomoea and for one transcript a
33 bp deletion that made it highly unlikely that the hypothetical protein was functional
and pointed toward pseudogenization.
The other two A. aureum copies with stops clustered with A_saundersii_33637 and
were not substituted by copies without premature stop codons. The unique sequence
A_smithii_23637
also
clustered
in
this
above
mentioned
species
group.
Aeonium_smithii_23637 had an indel at position 635 and the resulting frameshift
mutation yields in a premature stop. In addition, A_smithii_23637 had a unique intron
with a divergent recognition site for the excision of the intron of CA/AG instead of
GT/AG. Beside pseudogenization, this could also indicate alternative splicing since
95
Discussion
premature stop codons were also observed for cDNA sequences of MCS_AP1.
Pseudogenization
or
alternative
splicing
was
also
assumed
for
alcohol
dehydrogenase (Adh) genes in cotton. For all sequences of the A-subgenome of
Gossypium barbadense 67 bp were deleted from exon 3 and intron 4. Additionally,
the first nucleotide of intron 6 in the D-subgenome of G. hirsutum was found to be
polymorphic; 12 alleles having a guanine (G) and 32 alleles an adenine (A). Small
and Wendel (2002) discussed that this could be an alternate splice recognition site
that may also change the expression. Since diverging expression could be
associated with subfunctionalization the phenomenon may also be assumed for the
additional copies of A. aureum and A. saundersii, if they should be functional. This is
especially interesting since A_saundersii_33637 showed a sister relationship to the
MCS_AP1 gene copy of the Sedum ser. Monanthoidea and likewise A. aureum to S.
caeruleum. As already discussed above, both Sedum specific copies showed a
derived relationship to each other which may represent different gene copies and
functions. Both Sedum specific copies had Ka/Ks-values which indicate purifying
selection, and no hints for pseudogenization or loss of activity could be detected.
This relationship to potential functional gene copies weakens the assumption that the
additional copies of A. aureum and A. saundersii are in the progress to become
pseudogenes. Alternative splicing has been reported in several MADS-box genes
and in most cases was believed to have negative consequences. On the other hand,
occasionally it may have enhanced the function of a gene because novel transcripts
may had the potential to function as different proteins (Shan et al. 2007).
Nevertheless Álvarez et al. (2005), who detected an unusual 3´ intron splice site for
the AdhC sequences of several studied cotton accessions, discussed that this feature
in addition to premature stops may indicate pseudogenization in some species. They
discussed also an unusually high number of replacements for some CesA1b
sequences that also suggest pseudogenization. Baum et al. (2005) described a
frameshift mutation alternating 16 amino acids of the exon 2 for amplified LEAFY
sequences of some Brassicaceae species. If the same splice sites as in Arabidopsis
are utilized, this frameshift mutation would result in a premature stop codon at the
start of exon 3. However, the calculated pairwise Ka/Ks-value was 0.21 in contrast to
1 which would be predicted if neutral selection was acting (see, e.g., Zhang et al.
2008). This implies that purifying selection has continued to act after gene duplication
and suggests that both loci are functional. In the present study another observation
96
Discussion
was made. Calculated mean Ka/Ks-values of the above mentioned additional A.
aureum and A. saundersii sequence groups vary between 0.819 ± 0.059 and 1.143 ±
0.436 (0.885 excluding the two A. aureum sequences; fig. 44; appendix),
respectively. These values are significantly higher than those calculated for the main
copies (discussed below) and values close to 1 indicate neutral selection associated
with pseudogenization. Positive selection acting on a gene, and thus probably the
development of a new function, could not be detected since calculated Ka/Ks-values
were not significantly higher than 1. Even if neofunctionalization could not be ruled
out, this result would rather support nonfunctionalization and formation of
pseudogenes in the additional copies of A. aureum and A. saundersii. However, that
contradicts the above discussed phenomenon of alternative splicing and the potential
outcome of subfunctionalization. A reliable statement will only be possible after
further analyses, e.g., after studying expression patterns. If diverging or maybe even
tissue specific expression should be observed, neo- or subfunctionalization would
have to taken into stronger account. Janssens et al. (2008) observed in DEF-like
sequences of Impatiens a nucleotide deletion that caused a frameshift and resulted
in a premature stop. However, this truncation could be detected for almost all
Impatiens species suggesting that the derived protein may have a specific function
throughout the genus. In general, frameshift mutations which mostly result in
premature stop codons and truncated proteins were often considered as detrimental
for protein function and therefore of little evolutionary significance. However, more
recently, combinations of gene duplication and frameshift mutations were assumed to
be an evolutionary important mechanism for the emergence of new biological
functions
and
maybe
important
mechanisms
driving
speciation
processes
(Vandenbussche et al. 2003, Kramer et al. 2006, Janssens et al. 2008).
4.3. Selection pressure
Various evolutionary patterns have been found for different genes in plants. Selective
diversification, balancing selection, positive selection, relaxation of selective
constraints, and purifying selection could be estimated (e.g., Bishop et al. 2000,
Barrier et al. 2001, Olsen et al. 2002, Chen et al. 2004, Purugganan and Robichaux
2005). In the present study all analyzed genes and copies were under strong
purifying selection. Ka/Ks-values range between 0.106 and 0.110 for the structural
97
Discussion
gene (MCS_PEPC) and between 0.195 and 0.412 for the regulatory genes
(MCS_AP1 and MCS_AP3).
In general, Purugganan et al. (1995) detected that plant MADS-box genes evolved
more rapidly than typical eukaryotic loci but changed slower than other plant
regulatory genes. Olsen et al. (2002) found diverging nucleotide diversity and
contrasting selection pressure studying floral regulatory MADS-box genes of A.
thaliana in detail. For example, coding regions of TERMINAL FLOWER 1 (TFL1) and
LEAFY displayed significant reduction in nucleotide variation suggesting a recent
adaptive sweep. In contrast, coding regions of AP3, PI, AP1, and CAL showed similar
levels of nucleotide diversity and no evidence of either positive or balancing
selection, possibly because they control evolutionary conserved floral organ traits
(Olsen et al. 2002). However, Kramer et al. (1998) discussed that genes encoding for
A- and C-function such as AP1 or CAL showed high levels of constraint due to their
pleiotropic roles in floral meristem development and organ identity. B-class MADSbox genes such as AP3 or PI evolved 20-40% faster than all other plant MADS-box
genes since they are only involved in organ identity, with enormous morphological
plasticity of petals and stamens (for reference see Kramer et al. 1998).
This general observation cannot be confirmed in the present study. The highest
Ka/Ks-values were calculated for MCS_AP1 with Ka/Ks = 0.412 (copy A) and 0.373
(copy B). Both copies of MCS_AP3 had lower Ka/Ks-values. Also for the codon sites
that indicate positive selection, MCS_AP1 showed with 37 the highest number. The
results of the present study are also in line with the one of Barrier et al. (2001) who
detected the highest Ka/Ks-values for ASAP1, followed by ASAP3 for species of the
HSA. Barrier et al. (2001) as well as Lawton-Rauh et al. (2003; corresponding
population study) discussed that even if no positive selection could be detected, both
regulatory genes seemed to have a huge potential to influence the speciation
process. The greater number of nonsynonymous over synonymous sites in most loci,
and the potential therein to change the hypothetical protein in a drastic way, is
consistent with the possibility of positive selection. However, homologs of AP1 and
AP3 for HSA species showed clearly accelerated Ka/Ks-values (0.98 and 0.79,
respectively) compared to the values inferred in the present study.
Still, the calculated mean Ka/Ks-values of MCS_AP1 and MCS_AP3 were higher
than other estimated Ka/Ks-values for floral regulatory genes. Purugganan et al.
(1995) estimated for such genes of A. thaliana similar levels of sequence constraints
98
Discussion
and Ka/Ks-values ranging from 0.119 to 0.185. For the MCS species, only copy A of
MCS_AP3 showed a Ka/Ks-value (0.195) close to the above mentioned ones. This
value is also similar to the one estimated by Hernández-Hernández et al. (2007) for
all B-class MADS-box genes over a broad taxonomic range (Ka/Ks = 0.151). All other
Ka/Ks-values of the present study were significantly higher, ranging between 0.340
and 0.412. This suggest a partial influence in the speciation process since these
values are also much higher than mean Ka/Ks-values. Tiffin and Hahn (2002)
estimated a mean Ka/Ks-value of 0.14 for 218 randomly analyzed sequences of
Brassicaceae species. Lawton-Rauh et al. (1999) calculated a Ka/Ks-value of 0.18
for four genes compared between A. thaliana and A. lyrata. Barrier et al. (2003)
calculated a mean Ka/Ks-value of 0.213 for Arabidopsis species. The values for the
MCS species are also much higher than Ka/Ks-values estimated in other studies with
comparable experimental designs to that of the present study. Baum et al. (2005)
estimated for the transcription factor LEAFY in Brassicaceae a Ka/Ks-value of
approximately 0.1. However different Ka/Ks-values were estimated for rosetteflowering lineages compared to inflorescence flowering lineages suggesting an
impact on speciation. Guillet-Claude et al. (2004) estimated a Ka/Ks-value of 0.143
for conifer knox-I paralogs but further analyses suggested an influence of these
genes in the evolution of conifers. Within the genus Impatiens Janssens et al. (2007)
found a low Ka/Ks-value of 0.1 for DEF/AP3-like genes. This value is lower than the
Ka/Ks-values for both copies of MCS_AP3 in the present study indicating that
homologs of AP3 may be involved in the evolution of the MCS species. That
homologs of AP1 may also play a role in the speciation of MCS is suggested by
higher Ka/Ks-values calculated for both MCS_AP1 copies. The values were similar to
the one estimated for the Myc-like anthocyanin regulatory gene in Cornus. Fan et al.
(2004) estimated a mean value of 0.407 that exceeded the values reported for most
structural and partly for regulatory genes.
Nevertheless, the present values are still below the levels seen for some other
regulatory genes. For example, variable domains of the plant R-genes show Ka/Ksvalues close to 1 (see Purugganan et al. 1995, Bishop et al. 2000). Even if the genes
in the present study showed purifying selection, other studies concluded the
influence of several genes in the speciation process based on the detection of
positive selection. Positive selection, as fixation of advantageous mutations, has
been an exciting topic to evolutionary biologist since adaptive changes in genes and
99
Discussion
genomes are responsible for evolutionary innovations and species differences (Yang
2005). Nevertheless, beneficial mutations and positive selection are expected to be
extremely rare and patterns of sequence variation generally support the view that
many differences between and within species are nonadaptive (Ford 2002). In the
present study, sites under positive selection could be detected using codon-based
model tests but estimation over whole proteins suggested strong purifying selection
acting on all analyzed genes and copies. So far, positive selection was detected for
genes involved in sexual reproduction, for host-parasite interaction genes, for genes
which encode for enzymes involved in energy metabolism, and for regulatory genes
involved in plant morphology (Bishop et al. 2000, Ford 2002, Barrier et al. 2003,
Zhang 2003, Purugganan and Robichaux 2005). Barrier et al. (2003) found for 304
orthologous loci compared between A. thaliana and A. lyrata 14 (5%) with an
estimated Ka/Ks > 1, indicating positive selection and likely adaptive divergence
between both species. 5% of genes under positive selection is an exceptionally high
estimation, likely due to the fact that closely related species were studied. In contrast,
Endo et al. (1996) found for 3595 analyzed genes only 17 (0.5%) with evidence for
positive selection. However, their criterion, the estimation of Ka/Ks-values, averaged
the selection pressure over entire genes and is not powerful for detection of positive
selection. Also Tiffin and Hahn (2002) found no genes that indicated positive
selection, maybe because only a small fraction of the approx. 26,000 genes of A.
thaliana were analyzed, and the method of calculating Ka/Ks-values used had been
insensitive. Positive selection acts in most cases only on a small region of a gene,
may occur in an episodic fashion, and only in a narrow window of evolutionary time
(Ford 2002, Zhang 2003). Thus, the signal of positive selection may be overwhelmed
by purifying selection (Guillet-Claude et al. 2004, Zhang 2003). For example, the 14
genes under positive selection detected by Barrier et al. (2003) could not be
confirmed in the corresponding population study since all Ka/Ks-values were lower
than 1. An explanation may be heterogeneities in selective constraints across loci
and the use of longer sequences.
Also several other authors discuss that substitution rates between protein coding
regions are not equally distributed (e.g., Vergara-Silva et al. 2000, Ford 2002, Fan et
al. 2004, Yang 2005). Most proteins have highly conserved regions where
replacement mutations are not tolerated. For example, active sites are usually the
subject of intense evolutionary constraint and thus are highly conserved to preserve
100
Discussion
function (Bishop et al. 2000, Hughes et al. 2000, Koch et al. 2001). In the present
study only three of the 37 amino acids indicating positive selection for MCS_AP1
were found within the MADS-box but 18 amino acids indicated positive selection in
the more variable C-terminal domain. In general, the MADS-box domain, encoding
the putative DNA-binding region, has the lowest nonsynonymous substitution rate
whereas the K-domain and C-terminal domain display greater amounts of protein
sequence variation (Purugganan et al. 1995). Hernández-Hernández et al. (2007)
observed positive selection mostly within the K-domain of some B-class MADS-box
genes. These observations were confirmed in the present study where 14.7% (Kdomain) and 22.5% (C-domain) of positive selected amino acids were found within
these two domains respectively but only 11% within the partly amplified MADS-box
domain.
Besides differential selection pressure acting on specific genes, duplicated gene
copies may provide the raw material for evolution and species could evolve by
divergence of paralogs (Ohno 1970, Zhang et al. 2002, Zhang 2003, Purugganan
and Robichaux 2005). Many studies showed that selection pressure acting on
orthologous and paralogous gene copies could be different. Gene duplications may
result in redundancy of a gene function and allow the duplicated copies to evolve in
different ways (Ohno 1970, Lynch and Conery 2000, Zhang 2003, Moore and
Purugganan 2005). Conant and Wagner (2003) discussed that at least 20% of gene
duplicates diverged asymmetrically in an average genome. In contrast, there is
expanding literature demonstrating that gene duplications frequently showed
evidence of purifying selection, that both copies evolved at similar rates, and are
maintained long-term (see Bomblies and Doebley 2005). In the present study, results
for MCS_AP3 and partly for MCS_PEPC would support the first hypothesis whereas
the results of MCS_AP1 are in line with the second mentioned phenomenon. Both
copies of MCS_AP3 are under purifying selection but there are significant differences
for selection pressure and evolutionary rates. MCS_PEPC paralogs showed
significant differences in their evolutionary rates but purifying selection acts on both
gene copies and was not significantly different. No evidence for differential selection
pressure and evolutionary rates at all were estimated for MCS_AP1 sequences.
No rate differences were observed in the study of Barrier et al. (2001) either, where
paralogs of ASAP1, ASAP3, and ASCAB9 evolved with similar rates. However,
101
Discussion
Lawton-Rauh et al. (2003) detected different levels and patterns of variation for
duplicated ASAP3 genes, whereas in contrast similar levels of nucleotide diversity
were detected for the duplicated AP1 homologs in HSA species. Similar observations
for the present MCS_AP1 and MCS_AP3 sequences may suggest an influence in the
speciation of MCS as discussed by Barrier et al. (2001) and Lawton-Rauh et al.
(2003) for their species.
Baum et al. (2005) discussed that purifying selection acts on the duplicated gene loci
of the flower regulatory gene LEAFY in the Brassicaceae. However, duplicates
showed a significant tendency to have elevated Ka/Ks-ratios (Baum et al. 2005). The
paralogous loci AtHVA22d and AtHVA22e of A. thaliana studied by Chen et al.
(2004) evolved under purifying selection. Nevertheless, replacement changes in the
AtHVA22d locus were accelerated indicating relaxation of purifying selection after
gene duplication in one copy. Partially non-overlapping modes of expression
between the two functional paralogs suggest that subfunctionalization explain the
maintenance of the duplicated loci (Chen et al. 2004). Similar patterns may be
assumed for the homologs of MCS_AP3 and should therefore be studied in detail in
further analyses. Different substitution rates were found for duplicated genes in
Gossypium. Small and Wendel (2002) detected varying evolutionary pressures acting
on the two subgenomes (A and D). Nucleotide diversity was consistently higher for
genes of the D-subgenome that evolved at a significantly faster rate than the Asubgenome sequences. The genetic redundancy caused by polyploidy or large gene
families may have allowed relaxed selection in the D-subgenome whereas purifying
selection was maintained in the A-subgenome sequences (Small and Wendel 2002).
Zhang et al. (2008) detected two copies of a PISTILLATA-like gene in Cornus.
Purifying selection dominated the evolution of this gene and the estimated mean
Ka/Ks-values, 0.51 and 0.22, differed significantly and were larger than the one
inferred for the outgroup taxa. The increased Ka/Ks-ratio was the result of
accelerated nonsynonymous substitutions. Also Aagaard et al. (2005) showed that
paralogs of FLORICAULA and DEF in Lamiales appeared to avoid silencing as
typical fate of most gene duplicates. They found no evidence for adaptive divergence
acting on duplicated copies but relaxed purifying selection followed the duplication in
one or both copies. Purifying selection, despite known functional redundancy, was
also detected for the genes CYCLOIDEA (CYC) and DICHOTOMA (DICH) in A.
majus (Hileman and Baum 2003). CYC and DICH arise via a single duplication and
102
Discussion
purifying selection acting on both loci is consistent with subfunctionalization. Both
paralogs showed some differences in their expression pattern, a well-known
mechanism for subfunctionalization and long-term maintenance of duplicated genes.
In the present study, this phenomenon could be assumed for the redundant AP3- and
DEF-like genes (compare, e.g., fig. 15). Huttley et al. (1997) found heterogeneity in
the rate of substitution between members of the CHS gene family. The higher
nonsynonymous substitution rate for one gene copy may indicate that it evolved a
new function. Also Durbin et al. (2000) pointed out that new CHS genes were
recruited in flowering plants and the nucleotide substitution rate was frequently
accelerated for them. Different evolutionary rates were detected for the five CHS
genes in Ipomoea. CHS-A, B, and C genes evolved about 2.7 times faster than CHSD and E genes with a higher rate of replacement mutations in CHS-A, B, and C
(Durbin et al. 2000). Wang et al. (2007) studied selective modes among CHS
orthologs and paralogs in A. thaliana and A. halleri subsp. gemmifera. Purifying
selection is the main source influencing the evolution of the CHS loci but diverging
ratios of Ka/Ks-values imply different levels of functional constraints.
Conflicting to the above mentioned studies, and in line with the results observed for
MCS_AP1, is the study of Fortune et al. (2007). The three divergent homologous
waxy genes in the hexaploid Spartina species seem to have evolved under selective
constraints and relative rate tests indicated no significant rate heterogeneity between
the sequences. Zhang et al. (2002) found no evidence for positive selection and only
little evidence that paralogs evolved at different rates among duplicated gene pairs of
A. thaliana. Yang et al. (2002) studied duplicated CHS genes of the Asteraceae and
in contrast to other studies found no significant rate differences comparing different
copies. Altogether, the huge differences in selection pressure, in the evolutionary
rates and the diverse patterns reflect the complexity of evolution for duplicated genes
(Chen et al. 2004).
4.4. Regulatory versus structural genes
In adaptive radiations extensive studies were done to define major forces acting in
the speciation process. Accelerated evolution of regulatory genes compared to
structural genes has been proposed as main explanation for the decoupled rates of
molecular and morphological evolution (King and Wilson 1975, Remington and
Purugganan 2002, Durbin et al. 2003, Purugganan and Robichaux 2005). This rise
103
Discussion
the question after major and candidate genes or gene classes responsible for rapid
morphological evolution (e.g., Remington and Purugganan 2002). The present study
contributes to this research work and the two regulatory genes studied showed
significantly higher Ka/Ks-values when compared to the structural gene (table 13 and
14) even though all studied gene copies evolved under strong purifying selection.
The species of the MCS are extremely different concerning growth-form and flower
morphology. This is also true for the species of the HSA, the classic example of an
adaptive radiation in plants. Several genes such as the growth regulatory gene GAI,
R-genes, and floral regulatory genes ASAP1 and ASAP3 were analyzed within the
HSA to estimate their impact on the speciation process (Barrier et al. 2001,
Remington and Purugganan 2002, Lawton-Rauh et al. 2003, Purugganan and
Robichaux 2005). Selection pressures acting on these genes, gene copies, coding,
and promoter regions, and therefore evolutionary consequences, were extremely
different (Purugganan and Robichaux 2005). Barrier et al. (2001) found gene
duplication and accelerated mutation rates in the HSA homologs of AP1 and AP3
compared to the North American Tarweed species. The estimated Ka/Ks-values for
these homologs in the MCS species were significantly lower. To explain this
observation it is important to take into account that the species of the MCS are not
considered as an adaptive radiation (Mes 1995, Jorgensen and Frydenberg 1999,
Mort et al. 2002). Strong selection by ecological factors is not the major force driving
speciation in the MCS species. Instead, inter-island dispersal and subsequent
speciation is the main impact confirmed, e.g., by the fact that the five major growthforms arose only once and spread subsequently over different islands. Even if
adaptation cannot be ruled out it likely plays only a minor role in the MCS (Mes and ‘t
Hart 1996) whereas it is the driving force for speciation in the HSA (for reference see,
e.g., Barrier et al. 1999).
Confirming the theory of King and Wilson (1975) both regulatory genes, MCS_AP1
and MCS_AP3, evolved faster than the structural gene MCS_PEPC. Even if the
Ka/Ks-values for the regulatory genes in the present study show no evidence for
positive selection, they are higher than estimated mean Ka/Ks-values in other studies
(see discussion above). This may indicate an impact of the studied regulatory genes
in the speciation process even if maybe not as main driving forces. A pollination
study, conducted on Tenerife, partly support this observation (Esfeld et al. 2009).
104
Discussion
With help of a fluorescent dye powder (pollen analog) pollen transfer between
sympatric Aeonium species with overlapping flowering times was estimated. Despite
an overlapping spectrum of flower visitors, infraspecific pollen transfer clearly
exceeded the interspecific pollen transfer. The degree of pollen transfer seems
thereby be linked to flower morphology and biology. However, other characters than
those encoded by AP1 and AP3 seem to influence reproductive isolation since
highest pollen transfer rates were observed between species that had similarly
colored flowers and the same reward of pollen and nectar. Therefore, it could be
assumed that other genes such as genes encoding for flower color (anthocyanin2,
e.g., Quattrocchio et al. 1999 or CHS, e.g., Durbin et al. 2003) and genes that have
an impact on nectar or pollen may have had a higher influence in reproductive
isolation and thus speciation of the MCS. In addition different flowering times leading
to temporal isolation is discussed as the major reproductive barrier in Aeonium. Thus,
flowering time genes should be studied to estimate their influence in the speciation
process (Liu 1989, Jorgensen and Olesen 2001, Esfeld et al. 2009).
However, it is important to consider that significant differences in mean Ka/Ks-values
were found between the MCS species and the respective sister- and outgroup
Sedum species, suggesting an influence of the studied genes in the speciation
process. For MCS_PEPC the species of Sedum ser. Monanthoidea had a
significantly lower Ka/Ks-values than the MCS species. The same observation could
be made for MCS_AP3, where both outgroup (S. caeruleum and S. pubescens) and
sistergroup species (S. jaccardianum, S. modestum, and S. surculosum) showed a
significant lower mean Ka/Ks-value compared to copy A and copy B of the MCS
species. The difference between the mean Ka/Ks-values of the outgroup and
sistergroup species was not significant (data not shown). In contrast the obtained
pattern for MCS_AP1 was different. Both copies of MCS_AP1 for the MCS species
showed no significant difference when compared to the mean Ka/Ks-values of the
sister- and outgroup species. However, relationships between the amplified gene
copies are more complicated (compare, e.g., fig. 12). The exon phylogram suggests
a sister relationship between one additional A. saundersii sequence (33637) and the
Sedum ser. Monanthoidea species. The outgroup species (S. caeruleum) could be
found basal but the relationship is unresolved. Thus, it is possible that derived and
different gene copies were amplified that do not reflect the true relationships and
105
Discussion
should not be compared as sister- and outgroup relationships (see also discussion
above).
In general not all regulatory genes are involved in morphological evolution or with the
same strength in the speciation process (Olsen et al. 2002, Remington and
Purugganan 2002). Despite high growth-form divergence selective constraints remain
strong for duplicated GAI genes in the HSA. The constraint was somewhat relaxed in
one of the two Hawaiian copies compared to the North American lineage but relative
rate tests revealed no significant differences in the evolutionary rates of both gene
copies. Remington and Purugganan (2002) failed to detect any evidence of positive
selection and the calculated mean Ka/Ks-value was 0.37. Similar Ka/Ks-values were
obtained in the present study for both copies of MCS_AP1 and for copy B of
MCS_AP3. This mild degree of relaxed selective constraint for the regulatory genes
is in sharp contrast to the results of Barrier et al. (2001) who detected Ka/Ks-values
of 0.98 (ASAP1) and 0.79 (ASAP3), respectively. However, the present Ka/Ks-values
are slightly accelerated compared to that of the structural gene ASCAB9 (Ka/Ks =
0.21; Barrier et al. 2001) pointing toward an impact of MCS_AP1 and MCS_AP3 in
the speciation of MSC, although maybe a limited one.
Accelerated sequence evolution and positive selection is not a general phenomenon
or only true for a small subset of the coding region of regulatory loci. For instance,
Lukens and Doebley (2001) found no evidence for positive selection on transcription
factors involved in growth regulation. Since many genes are involved in this process,
overlaying effects may obscuring an existing pattern. This suggests that candidate
genes or single speciation genes may not exist (see also Hodges 1997, Remington
and Purugganan 2002, Dias et al. 2003).
In contrast, gene regulation due to changes in the coding sequence of regulatory loci
or in cis acting promoter sequences can play a key role in morphological evolution
(Baum 1998). Remington and Purugganan (2002) confirmed the importance of gene
regulation in the speciation process. They found no evidence for positive selection
acting on the coding region of the GAI gene in the HSA. In contrast, the ~900 bp
upstream flanking region showed variable rates and patterns of evolution that may
reflect positive selection. Further studies, e.g., the one of Doebley and Lukens
(1998), have found higher rates of sequence polymorphism and divergence in
functionally important promoter segments compared to the corresponding gene
regions. Subsequently, changes in the level or expression pattern of genes were
106
Discussion
observed that lead to novel phenotypes in maize. The influence of gene regulation
was also extensively studied concerning the impact of flower color in the speciation
process. Striking differences in flower color and morphology is associated with
speciation, e.g., in Aquilegia, Ipomoea, and Mimulus (Hodges and Arnold 1994a,
Schemske and Bradshaw 1999, Durbin et al. 2003). Plant pigments influencing flower
color and hence pollinator attraction. To figure out their impact on the speciation
process, many studies focused on the structural gene CHS, a starting point in the
flavonoid synthesis (Clegg and Durbin 2000). However, most species differences
found so far were associated with changes in regulation of gene expression. Durbin
et al. (2003) found only one case where phenotypic differences in pigmentation
patterns could be explained by mutation of the structural gene. There are only a
limited number of ways that a mutation in the structural gene can lead to a change in
flower color or this change may rather result in the total loss of pigmentation. Thus,
changes in a structural gene will not have the same impact in terms of adaptive
evolution. Changes in regulatory genes on the other hand often result in novel
patterns of pigmentation that can influence pollinator behavior or even attract new
pollinators highly influencing the speciation process (Durbin et al. 2003).
Thus, there are many more ways in which mutations in gene regulation can occur
considering the complexity and many different types of genes involved in regulation
(Purugganan and Robichaux 2005). Mutations in regulatory genes can provide a
more rapid response to the environment and a larger adaptive success whereas in
contrast, structural gene evolution may not be the key to rapid and extensive
diversification in organismal form (Barrier et al. 2001, Durbin et al. 2003, Purugganan
and Robichaux 2005).
107
Summary
5. Summary
Evolution and speciation belong to the most exciting topics in biology. The diversity of
species evident in radiations provides a valuable tool to study these processes.
Numerous studies have shown that the morphological variation of radiated species
does not correlate with their molecular diversity and that regulatory genes evolve
faster than structural genes.
However, the determination of candidate genes involved in speciation is difficult.
Based on results of Barrier et al. (2001) the present study contributes to the ongoing
debate about what triggers speciation and evolution. The two regulatory genes,
homologs of the homeotic floral genes APETALA1 and APETALA3, as well as the
structural gene encoding for phosphoenolpyruvate carboxylase (PEPC) were studied
in the radiation of the Macaronesian Crassulaceae Sempervivoideae (MCS).
Robust phylogenies are needed to determine the influence of traits in speciation, and
their lack is a well known problem for radiations and fast species evolution. However,
comparisons of species phylogenies and genealogies may enable the observation of
major impact factors since low-copy genes can be involved in the determination of
phenotypes and adaptation. The studied low-copy nuclear genes both support and
contradict relationships observed for the MCS species based on morphological and
molecular markers. Several relationships like the one between A. aureum, A.
saundersii, and A. smithii or between the Sedum species are highly supported. On
the other hand, conflicting relationships were observed for A. canariense, A.
cuneatum, A. nobile, and A. rubrolineatum. In addition, the low-copy nuclear genes
shed light on highly debated relationships like the one of A. goochiae at the base of
all Aeonium species and M. icterica closely related to the Aichryson clade. MCS_AP3
is the most promising phylogenetic marker and may also have been involved in the
speciation as it showed significant differences in the estimated Ka/Ks-values and in
the evolutionary rates between the amplified gene copies.
An important observation concerns the copy number of the studied genes. Subclade
specific gene duplications were observed for the polyploid species of the MCS island
radiation but not for the diploid sister- and outgroup species of Sedum. The obtained
pattern strongly suggests that the gene duplication was
connected with
polyploidization and island colonization. Non-species specific gene loss or retention
could be observed as well as potential random gain and pseudogenization.
108
Summary
Gene duplications may provide the raw material for evolution and in general, the
influence of genes and gene copies in speciation could be defined by the selection
pressure. Positive selection connected with the development of a new function would
provide a strong evidence. However, purifying selection was observed for all studied
genes and copies. Nevertheless, the estimated Ka/Ks-values, defining the selection
pressure, for the regulatory genes were higher than mean Ka/Ks-values and also
higher than values observed in other studies with a comparable study design.
MCS_AP1 showed the highest Ka/Ks-values followed by the second regulatory gene
MCS_AP3. As predicted by the hypothesis of King and Wilson (1975), the structural
gene MCS_PEPC had the lowest Ka/Ks-values.
However, overall a limited influence of the analyzed genes in the speciation process
was suggested. Therefore, ongoing studies should focus on additional genes as well
as on gene regulation (promoter sequences) and should include more species of the
MCS for reliable conclusions. Following the results and observations of the present
study, e.g., genes influencing growth-form, flowering time, flower color, position of
inflorescence, number of floral organs, and genes influencing woodiness should be
studied to gain insights into the speciation process of the Macaronesian
Crassulaceae Sempervivoideae.
109
Literature
Literature
Aagaard, J.E., Olmstead, R.G., Willis, J.H. and Phillips, P.C. (2005): Duplication of floral regulatory
genes in the Lamiales. American Journal of Botany: 92, pp. 1284-1293.
Álvarez, I., Cronn, R. and Wendel, J.F. (2005): Phylogeny of the New World diploid cottons
(Gossypium L., Malvaceae) based on sequences of three low-copy nuclear genes. Plant Systematics
and Evolution: 252, pp. 199-214.
Andrus, N., Trusty, J., Santos-Guerra, A., Jansen, R.K. and Francisco-Ortega, J. (2004): Using
molecular phylogenies to test phytogeographical links between East/South Africa-Southern Arabia and
the Macaronesian islands – a review, and the case of Vierea and Pulicaria section Vieraeopsis
(Asteraceae). Taxon: 53, pp. 333-346.
Arnedo, M.A., Oromí, P. and Ribera, C. (1996): Radiation of the genus Dysdera (Araneae,
Haplogynae, Dysderidae) in the Canary Islands: the western islands. Zoologica Scripta: 25, pp. 241274.
Bailey, C.D. and Doyle, J.J. (1999): Potential phylogenetic utility of the low-copy nuclear gene pistillata
in dicotyledonous plants: comparison to nrDNA ITS and trnL intron in Sphaerocardamum and other
Brassicaceae. Molecular Phylogenetics and Evolution: 13, pp. 20-30.
Baker, H.G. and Baker, I. (1983): A brief historical review of the chemistry of floral nectar. In: Bentley,
B. and Elias, T. (eds): The biology of nectaries. Columbia University Press, New York.
Baldwin, B.G. and Robichaux, R.H. (1995): Historical biogeography and ecology of the Hawaiian
silversword alliance: new molecular phylogenetic perspectives. In: Wagner, W.L. and Funk, V.A. (eds):
Hawaiian biogeography: evolution on a hot spot archipelago. Smithsonian Institution Press,
Washington, D.C.
Baldwin, B.G., Crawford, D.J., Francisco-Ortega, J., Kim, S.-C., Sang, T. and Stuessy, T.F. (1998):
Molecular phylogenetic insights on the origin and evolution of oceanic island plants. In: Soltis, D.E.,
Soltis, P.S. and Doyle, J.J. (eds): Molecular Evolution and Plant Systematics.
Barber, J.C., Francisco-Ortega, J., Santos-Guerra, A., Marrero, A. and Jansen, R.K. (2000): Evolution
of endemic Sideritis (Lamiaceae) in Macaronesia: insights from a chloroplast DNA restriction site
analysis. Systematic Botany: 25, pp. 633-647.
Barrier, M., Baldwin, B.G., Robichaux, R.H. and Purugganan, M.D. (1999): Interspecific hybrid
ancestry of a plant adaptive radiation: allopolyploidy of the Hawaiian silversword alliance (Asteraceae)
inferred from floral homeotic gene duplications. Molecular Biology and Evolution: 16, pp. 1105-1113.
110
Literature
Barrier, M., Robichaux, R.H. and Purugganan, M.D. (2001): Accelerated regulatory gene evolution in
an adaptive radiation. Proceedings of the National Academic Society of the USA: 98, pp. 1020810213.
Barrier, M., Bustamante, C.D., Yu, J. and Purugganan, M.D. (2003): Selection on rapidly evolving
proteins in the Arabidopsis genome. Genetics: 163, pp. 723-733.
Baum, D.A. (1998): The evolution of plant development. Current Opinion in Plant Biology: 1, pp. 79-86.
Baum, D.A., Yoon, H.-S. and Oldham, R.L. (2005): Molecular evolution of the transcription factor
LEAFY in Brassicaceae. Molecular Phylogenetics and Evolution: 37, pp. 1-14.
Beck, J.B., Al-Shehbaz, I.A., O´Kane Jr., S.L. and Schaal, B.A. (2007): Further insights into the
phylogeny of Arabidopsis (Brassicaceae) from nuclear Atmyb2 flanking sequence. Molecular
Phylogenetics and Evolution: 42, pp. 122-130.
Becker, A., Winter, K.-U., Meyer, B., Saedler, H. and Theissen, G. (2000): MADS-box gene diversity in
seed plants 300 million years ago. Molecular Biology and Evolution: 17, pp. 1425-1434.
Berbel, A., Navarro, C., Ferrándiz, C., Cañas, L.A., Madueño, F. and Beltrán, J.P. (2001): Analysis of
PEAM4, the pea AP1 functional homologue, supports a model for AP1-like genes controlling both
floral meristem and floral organ identity in different plant species. The Plant Journal: 25, pp. 441-451.
Berger, A. (1930): Crassulaceae. In: Engler, A. and Prantl, K. (eds): Die natürlichen Pflanzenfamilien.
2nd ed. 18a. W. Engelmann, Leipzig.
Bishop, J.G., Dean, A.M. and Mitchell-Olds, T. (2000): Rapid evolution in plant chitinases: Molecular
targets of selection in plant-pathogen coevolution. Proceedings of the National Academic Society of
the USA: 97, pp. 5322-5327.
Blanc, G. and Wolfe, K.H. (2004): Functional divergence of duplicated genes formed by polyploidy
during Arabidopsis evolution. The Plant Cell: 16, pp. 1679-1691.
Blattner, F.R. (1999): Direct amplification of the entire ITS region from poorly preserved plant material
using recombinant PCR. BioTechniques: 27, pp. 173-178.
Böhle, U.-R., Hilger, H.H. and Martin, W.F. (1996): Island colonization and evolution of the insular
woody habit in Echium L. (Boraginaceae). Proceedings of the National Academic Society of the USA:
93, pp. 11740-11745.
111
Literature
Bomblies, K. and Doebley, J.F. (2005): Molecular evolution of FLORICAULA/LEAFY orthologs in the
Andropogoneae (Poaceae). Molecular Biology and Evolution: 22, pp. 1082-1094.
Bramwell, D. (1976): The endemic flora of the Canary Islands. In: Kunkel, G. (ed): Biogeography and
ecology in the Canary Islands. Junk Publishers, The Hague.
Carlquist, S. (1962): A theory of paedomorphosis in dicotyledonous woods. Phytomorphology: 12, pp.
30-45.
Carlquist, S. (1974): Island Biology. Columbia University Press, New York & London.
Carracedo, J.C. (1994): The Canary Islands: an example of structural control on the growth of large
oceanic-island volcanoes. Journal of Volcanology and Geothermal Research: 60, pp. 225-241.
Chen, C.-N., Chiang, Y.-C., Ho, T.-H.D., Schaal, B.A. and Chiang, T.-Y. (2004): Coalescent processes
and relaxation of selective constraints leading to contrasting genetic diversity at paralogs AtHVA22d
and AtHVA22e in Arabidopsis thaliana. Molecular Phylogenetics and Evolution: 32, pp. 616-626.
Chollet, R. (1996): Phosphoenolpyruvate carboxylase: a ubiquitous, highly regulated enzyme in plants.
Annual Review of Plant Physiology and Plant Molecular Biology: 47, pp. 273-298.
Clegg, M.T. and Durbin, M.L. (2000): Flower color variation: a model for the experimental study of
evolution. Proceedings of the National Academic Society of the USA: 97, pp. 7016-7023.
Coen, E.S. and Meyerowitz, E.M. (1991): The war of the whorls: genetic interactions controlling flower
development. Nature: 353, pp. 31-37.
Conant, G.C. and Wagner, A. (2003): Asymmetric sequence divergence of duplicate genes. Genome
Research: 13, pp. 2052-2058.
Cushman, J.C. and Bohnert, H.J. (1999): Crassulacean acid metabolism: Molecular genetics. Annual
Review of Plant Physiology and Plant Molecular Biology: 50, pp. 305-332.
Dias, A.P., Braun, E.L., McMullen, M.D. and Grotewold, E. (2003): Recently duplicated maize R2R3
Myb genes provide evidence for distinct mechanisms of evolutionary divergence after duplication.
Plant Physiology: 131, pp. 610-620.
Doebley, J. and Lukens, L. (1998): Transcriptional regulators and the evolution of plant form. The
Plant Cell: 10, pp. 1075-1082.
112
Literature
Doron-Faigenboim, A., Stern, A., Mayrose, I., Bacharach, E. and Pupko, T. (2005): Selecton: a server
for detecting evolutionary forces at a single amino-acid site. Bioinformatics Applications Note: 21, pp.
2101-2103.
Duarte, J.M., Cui, L., Wall, P.K., Zhang, Q., Zhang, X., Leebens-Mack, J., Ma, H., Altman, N. and
dePamphilis,
C.W.
(2006):
Expression
pattern
shifts
following
duplication
indicative
of
subfunctionalization and neofunctionalization in regulatory genes of Arabidopsis. Molecular Biology
and Evolution: 23, pp. 469-478.
Durbin, M.L., McCaig, B. and Clegg, M.T. (2000): Molecular evolution of the chalcone synthase
multigene family in the morning glory genome. Plant Molecular Biology: 42, pp. 79-92.
Durbin, M.L, Lundy, K.E., Morrell, P.L., Torres-Martinez, C.L. and Clegg, M.T. (2003): Genes that
determine flower color: the role of regulatory changes in the evolution of phenotypic adaptations.
Molecular Phylogenetics and Evolution: 29, pp. 507-518.
Emerson, B.C. (2002): Evolution on oceanic islands: molecular phylogenetic approaches to
understanding pattern and process. Molecular Ecology: 11, pp. 951-966.
Endo, T., Ikeo, K. and Gojobori, T. (1996): Large-scale search for genes on which positive selection
may operate. Molecular Biology and Evolution: 13, pp. 685-690.
Erbar, C. (2007): Current opinions in flower development and the evo-devo approach in plant
phylogeny. Plant Systematics and Evolution: 269, pp. 107-132.
Esfeld, K., Koch, M.A., van der Niet, T., Seifan, M. and Thiv, M. (2009): Little interspecific pollen
transfer despite overlap in pollinators between sympatric Aeonium (Crassulaceae) species pairs.
Flora. doi:10.1016/j.flora.2008.10.002.
Eyre-Walker, A. (2006): The genomic rate of adaptive evolution. Trends in Ecology and Evolution: 21,
pp. 569-575.
Fairfield, K.N., Mort, M.E. and Santos-Guerra, A. (2004): Phylogenetics and evolution of the
Macaronesian members of the genus Aichryson (Crassulaceae) inferred from nuclear and chloroplast
sequence data. Plant Systematics and Evolution: 248, pp. 71-83.
Fan, C., Purugganan, M.D., Thomas, D.T., Wiegmann, B.M. and Xiang, Q.-Y. (2004): Heterogeneous
evolution of the Myc-like anthocyanin regulatory gene and its phylogenetic utility in Cornus L.
(Cornaceae). Molecular Phylogenetics and Evolution: 33, pp. 580-594.
113
Literature
Fishbein, M., Hibsch-Jetter, C., Soltis, D.E. and Hufford, L. (2001): Phylogeny of Saxifragales
(Angiosperms, Eudicots): analysis of a rapid, ancient radiation. Systematic Biology: 50, pp. 817-847.
Ford, M.J. (2002): Applications of selective neutrality tests to molecular ecology. Molecular Ecology:
11, pp. 1245-1262.
Fortune, P.M., Schierenbeck, K.A., Ainouche, A.K., Jacquemin, J., Wendel, J.F. and Ainouche, M.L.
(2007): Evolutionary dynamics of Waxy and the origin of hexaploid Spartina species (Poaceae).
Molecular Phylogenetics and Evolution: 43, pp. 1040-1055.
Francisco-Ortega, J., Jansen, R.K. and Santos-Guerra, A. (1996): Chloroplast DNA evidence of
colonization, adaptive radiation, and hybridization in the evolution of the Macaronesian flora.
Proceedings of the National Academic Society of the USA: 93, pp. 4085-4090.
Francisco-Ortega, J., Santos-Guerra, A., Kim, S.-C. and Crawford, D.J. (2000): Plant genetic diversity
in the Canary Islands: a conservation perspective. American Journal of Botany: 87, pp. 909-919.
Francisco-Ortega, J., Barber, J.C., Santos-Guerra, A., Febles-Hernández, R. and Jansen, R.K. (2001):
Origin and evolution of the endemic genera of Gonosperminae (Asteraceae: Anthemideae) from the
Canary Islands: evidence from nucleotide sequences of the internal transcribed spacers of the nuclear
ribosomal DNA. American Journal of Botany: 88, pp. 161-169.
Francisco-Ortega, J., Fuertes-Aguilar, J., Kim, S.-C., Santos-Guerra, A., Crawford, D.J. and Jansen,
R.K. (2002): Phylogeny of the Macaronesian endemic Crambe section Dendrocrambe (Brassicaceae)
based on internal transcribed spacer sequences of nuclear ribosomal DNA. American Journal of
Botany: 89, pp. 1984-1990.
Gehrig, H., Taybi, T., Kluge, M. and Brulfert, J. (1995): Identification of multiple PEPC isogenes in
leaves of the facultative Crassulacean Acid Metabolism (CAM) plant Kalanchoe blossfeldiana Poelln.
cv. Tom Thumb. FEBS Letters: 377, pp. 399-402.
Grob, G.B.J., Gravendeel, B. and Eurlings, M.C.M. (2004): Potential phylogenetic utility of the nuclear
FLORICAULA/LEAFY
second
intron:
comparison
with
three
chloroplast
DNA
regions
in
Amorphophallus (Araceae). Molecular Phylogenetics and Evolution: 30, pp. 13-23.
Guillet-Claude, C., Isabel, N., Pelgas, B. and Bousquet, J. (2004): The evolutionary implications of
knox-I gene duplications in conifers: correlated evidence from phylogeny, gene mapping, and analysis
of functional divergence. Molecular Biology and Evolution: 21, pp. 2232-2245.
Guindon, S. and Gascuel, O. (2003): A simple, fast, and accurate algorithm to estimate large
phylogenies by maximum likelihood. Systematic Biology: 52, pp. 696-704.
114
Literature
Guindon, S., Lethiec, F., Duroux, P. and Gascuel, O. (2005): PHYML Online – a web server for fast
maximum likelihood-based phylogenetic inference. Nucleic Acids Research: 33, pp. W557-W559.
Hernández-Hernández, T., Martínez-Castilla, L.P. and Alvarez-Buylla, E.R. (2007): Functional
diversification of B MADS-box homeotic regulators of flower development: adaptive evolution in
protein-protein interaction domains after major gene duplication events. Molecular Biology and
Evolution: 24, pp. 465-481.
Hileman, L.C. and Baum, D.A. (2003): Why do paralogs persist? Molecular evolution of CYCLOIDAE
and related floral symmetry genes in Antirrhineae (Veronicaceae). Molecular Biology and Evolution:
20, pp. 591-600.
Hoballah, M.E., Gübitz, T., Stuurman, J., Broger, L., Barone, M., Mandel, T., Dell’Olivo, A., Arnold, M.
and Kuhlemeier, C. (2007): Single gene-mediated shift in pollinator attraction in Petunia. The Plant
Cell: 19, pp. 779-790.
Hodges, S.A. and Arnold, M.L. (1994a): Columbines: a geographically widespread species flock.
Proceedings of the National Academic Society of the USA: 91, pp. 5129-5132.
Hodges, S.A. and Arnold, M.L. (1994b): Floral and ecological isolation between Aquilegia formosa and
Aquilegia pubescens. Proceedings of the National Academic Society of the USA: 91, pp. 2493-2496.
Hodges, S.A. (1997): Floral nectar spurs and diversification. International Journal of Plant Science:
158, pp. S81-S88.
Hohenester, A. and Welss, W. (1993): Exkursionsflora für die Kanarischen Inseln. Ulmer,
Stuttgart.
Huber, T., Faulkner, G. and Hugenholtz, P. (2004): Bellerophon: a program to detect chimeric
sequences in multiple sequence alignments. Bioinformatics: 20, pp. 2317-2319.
Huelsenbeck, J.P. and Ronquist, F. (2001): MRBAYES: Bayesian inference of phylogenetic trees.
Bioinformatics: 17, pp. 754-755.
Hughes, A.L., Green, J.A., Garbayo, J.M. and Roberts, R.M. (2000): Adaptive diversification within a
large family of recently duplicated, placentally expressed genes. Proceedings of the National
Academic Society of the USA: 97, pp. 3319-3323.
Hurst, L.D. (2002): The Ka/Ks ratio: diagnosing the form of sequence evolution. TRENDS in Genetics:
18, pp. 486-487.
115
Literature
Huttley, G.A., Durbin, M.L., Glover, D.E. and Clegg, M.T. (1997): Nucleotide polymorphism in the
chalcone synthase-A locus and evolution of the chalcone synthase multigene family of common
morning glory Ipomoea purpurea. Molecular Ecology: 6, pp. 549-558.
Irish, V.F. and Litt, A. (2005): Flower development and evolution: gene duplication, and diversification
and redeployment. Current Opinion in Genetics and Development: 15, pp. 454-460.
Janssens, B.S., Geuten, K., Viaene, T., Yuan, Y.-M., Song, Y. and Smets, E. (2007): Phylogenetic
utility of the AP3/DEF K-domain and its molecular evolution in Impatiens (Balsaminiacea). Molecular
Phylogenetics and Evolution: 43, pp. 225-239.
Janssens, B.S., Viaene, T., Huysmans, S., Smets, E. and Geuten, K.P. (2008): Selection on length
mutations after frameshift can explain the origin and retention of the AP3/DEF-like paralogues in
Impatiens. Journal of Molecular Evolution: 66, pp. 424-435.
Jorgensen, T.H. and Frydenberg, J. (1999): Diversification in insular plants: inferring the phylogenetic
relationship in Aeonium (Crassulaceae) using ITS sequences of nuclear ribosomal DNA. Nordic
Journal of Botany: 19, pp. 613-621.
Jorgensen, T.H. and Olesen, J.M. (2001): Adaptive radiation of island plants: evidence from Aeonium
(Crassulaceae) of the Canary Islands. Perspectives in Plant Ecology, Evolution and Systematics: 4,
pp. 29-42.
Jorgensen, T.H. (2002): The importance of phylogeny and ecology in microgeographical variation in
the morphology of four Canarian species of Aeonium (Crassulaceae). Biological Journal of the
Linnean Society: 76, pp. 521-533.
Juan, C., Emerson, B.C., Oromí, P. and Hewitt, G.M. (2000): Colonization and diversification: towards
a phylogeographic synthesis for the Canary Islands. TRENDS in Ecology and Evolution: 15, pp. 104109.
Juenger, T., Purugganan, M.D. and Mackay, T.F.C. (2000): Quantitative trait loci for floral morphology
in Arabidopsis thaliana. Genetics: 156, pp. 1379-1392.
Kim, S.-C., Crawford, D.J., Francisco-Ortega, J. and Santos-Guerra, A. (1996): A common origin for
woody Sonchus and five related genera in the Macaronesian islands: molecular evidence for extensive
radiation. Proceedings of the National Academic Society of the USA: 93, pp. 7743-7748.
King, M.C. and Wilson, A.C. (1975): Evolution at two levels in humans and chimpanzees. Science:
188, pp. 107-116.
116
Literature
Koch, M.A., Weisshaar, B., Kroymann, J., Haubold, B. and Mitchell-Olds, T. (2001): Comparative
genomics and regulatory evolution: conservation and function of the Chs and Apetala3 promoters.
Molecular Biology and Evolution: 18, pp. 1882-1891.
Kramer, E.M., Dorit, R.L. and Irish, V.F. (1998): Molecular evolution of genes controlling petal and
stamen development: duplication and divergence within the APETALA3 and PISTILLATA MADS-box
gene lineages. Genetics: 149, pp. 765-783.
Kramer, E.M. and Hall, J.C. (2005): Evolutionary dynamics of genes controlling floral development.
Current Opinion in Plant Biology: 8, pp. 13-18.
Kramer, E.M., Su, H.-J., Wu, C.-C. and Hu, J.-H. (2006): A simplified explanation for the frameshift
mutation that created a novel C-terminal motif in the APETALA3 gene lineage. BMC Evolution and
Biology: 6, pp. 30-47.
Kull, U. (1982): Artbildung durch geographische Isolation bei Pflanzen – die Gattung Aeonium auf
Teneriffa. Natur und Museum: 112, pp. 33-40.
Lawton-Rauh, A., Buckler IV, E.S. and Purugganan, M.D. (1999): Patterns of molecular evolution
among paralogous floral homeotic genes. Molecular Biology and Evolution: 16, pp. 1037-1045.
Lawton-Rauh, A., Alvarez-Buylla, E.R. and Purugganan, M.D. (2000): Molecular evolution of flower
development. TRENDS in Ecology and Evolution: 15, pp. 144-149.
Lawton-Rauh, A. (2003): Evolutionary dynamics of duplicated genes in plants. Molecular
Phylogenetics and Evolution: 29, pp. 396-409.
Lawton-Rauh, A., Robichaux, R.H. and Purugganan, M.D. (2003): Patterns of nucleotide variation in
homoeologous regulatory genes in the allotetraploid Hawaiian silversword alliance (Asteraceae).
Molecular Ecology: 12, pp. 1301-1313.
Lems, K. (1960): Botanical notes on the Canary Islands. II. The evolution of plant forms in the islands:
Aeonium. Ecology: 41, pp. 1-17.
Litt, A. and Irish, V.F. (2003): Duplication and diversification in the APETALA1/FRUITFUL floral
homeotic gene lineage: implications for the evolution of floral development. Genetics: 165, pp. 821833.
Liu, H.-Y. (1989): Systematics of Aeonium (Crassulaceae). National Museum of Natural Science,
Taichung.
117
Literature
Lohmann, J.U. and Weigel, D. (2002): Building Beauty: the genetic control of floral patterning.
Developmental Cell: 2, pp. 135-142.
Lösch, R. (1990): Funktionelle Voraussetzungen der adaptiven Nischenbesetzung in der Evolution der
makaronesischen Semperviven. Dissertationes Botanicae: 146. Cramer, Berlin.
Lukens, L. and Doebley, J. (2001): Molecular evolution of the teosinte branched gene among maize
and related grasses. Molecular Biology and Evolution: 18, pp. 627-638.
Lynch, M. and Conery, J.S. (2000): The evolutionary fate and consequences of duplicate genes.
Science: 290, pp. 1151-1155.
Lynch, M., O’Hely, M., Walsh, B. and Force, A. (2001): The probability of preservation of a newly
arisen gene duplicate. Genetics: 159, pp. 1789-1804.
MacArthur, R.H. and Wilson, E.O. (1967): The theory of island biography. Princeton University Press,
Princeton.
Maire, R. (1976): Flore de l’Afrique du Nord. Vol. 14. Lechevallier, Paris.
Malcomber, S.T. (2002): Phylogeny of Gaertnera Lam. (Rubiaceae) based on multiple DNA markers:
evidence of a rapid radiation in a widespread, morphologically diverse genus. Evolution: 56, pp. 42-57.
Mandel, M.A. and Yanofsky, M.F. (1995): A gene triggering flower formation in Arabidopsis. Nature:
377, pp. 522-524.
Meimberg, H., Abele, T., Bräuchler, C., McKay, J.K., Pérez de Paz, P.L. and Heubl, G. (2006):
Molecular evidence for adaptive radiation of Micromeria Benth. (Lamiaceae) on the Canary Islands as
inferred from chloroplast and nuclear DNA sequences and ISSR fingerprint data. Molecular
Phylogenetics and Evolution: 41, pp. 566-578.
Mes, T.H.M. and ‘t Hart, H. (1994): Sedum surculosum and S. jaccardianum (Crassulaceae) share a
unique 70 bp deletion in the chloroplast DNA trnL (UUA) - trnF (GAA) intergenic spacer. Plant
Systematics and Evolution: 193, pp. 213-221.
Mes, T.H.M. (1995): Phylogenetic and systematic implications of chloroplast and nuclear spacer
sequence variation in the Macaronesian Sempervivoideae and related Sedoideae. In: ‘t Hart, H. and
Eggli, U. (eds): Evolution and Systematics of Crassulaceae. Backhuis Publishers, Leiden.
Mes, T.H.M. and ‘t Hart, H. (1996): The evolution of growth-forms in the Macaronesian genus Aeonium
(Crassulaceae) inferred from chloroplast DNA RFLPs and morphology. Molecular Ecology: 5, pp. 351363.
118
Literature
Mes, T.H.M., van Brederode, J. and ‘t Hart, H. (1996): Origin of the woody Macaronesian
Sempervivoideae and the phylogenetic position of the East African species of Aeonium. Botanica
Acta: 109, pp. 477-491.
Mes, T.H.M., Wijers, G.-J. and ‘t Hart, H. (1997): Phylogenetic relationships in Monanthes
(Crassulaceae) based on morphological, chloroplast and nuclear DNA variation. Journal of
Evolutionary Biology: 10, pp. 193-216.
Moore, R.C. and Purugganan, M.D. (2003): The early stages of duplicate gene evolution. Proceedings
of the National Academic Society of the USA: 100, pp. 15682-15687.
Moore, R.C. and Purugganan, M.D. (2005): The evolutionary dynamics of plant duplicate genes.
Current Opinion in Plant Biology: 8, pp. 122-128.
Mort, M.E., Soltis, D.E., Soltis, P.S., Francisco-Ortega, J. and Santos-Guerra, A. (2001): Phylogenetic
relationships and evolution of Crassulaceae inferred from matK sequence data. American Journal of
Botany: 88, pp. 76-91.
Mort, M.E., Soltis, D.E., Soltis, P.S., Francisco-Ortega, J. and Santos-Guerra, A. (2002):
Phylogenetics and evolution of the Macaronesian clade of Crassulaceae inferred from nuclear and
chloroplast sequence data. Systematic Botany: 27, pp. 271-288.
Mort, M.E. and Crawford, D.J. (2004): The continuing search: low-copy nuclear sequences for lowerlevel plant molecular phylogenetic studies. Taxon: 53, pp. 257-261.
Mort, M.E., Soltis, D.E., Soltis, P.S., Santos-Guerra, A. and Francisco-Ortega, J. (2007): Physiological
evolution and association between physiology and growth form in Aeonium (Crassulaceae). Taxon:
56, pp. 453-464.
Nei, M. and Gojobori, T. (1986): Simple methods for estimating the numbers of synonymous and
nonsynonymous nucleotide substitutions. Molecular Biology and Evolution: 3, pp. 418-426.
Nei, M. and Kumar, S. (2000): Molecular Evolution and Phylogenetics. Oxford University Press, New
York.
Nekrutenko, A., Makova, K.D. and Li, W.-H. (2001): The Ka/Ks ratio test for assessing the proteincoding potential of genomic regions: an empirical and simulation study. Genome Research: 12, pp.
198-202.
Nyffeler, R. (1992): A taxonomic revision of the genus Monanthes Haworth (Crassulaceae). Bradleya:
10, pp. 49-82.
119
Literature
Nyffeler, R. (1995): Die Gattung Monanthes Haworth. Kakteen und andere Sukkulenten: 46, pp. 156165.
Ohno, S. (1970): Evolution by gene duplication. Springer-Verlag, Berlin.
Olsen, K.M., Womack, A., Garrett, A.R., Suddith, J.I. and Purugganan, M.D. (2002): Contrasting
evolutionary forces in the Arabidopsis thaliana floral developmental pathway. Genetics: 160, pp. 16411650.
Page, R.D. (1996): TreeView: an application to display phylogenetic trees on personal computers.
Computer Applications in the Biosciences: 12, pp. 357-358.
Pilon-Smits, E.A.H., ‘t Hart, H., Maas, J.W., Meesterburrie, J.A.N., Kreuler, R. and van Brederode, J.
(1992): The evolution of the crassulacean acid metabolism in Aeonium inferred from carbon isotope
composition and enzyme activities. Oecologia: 91, p. 548-553.
Posada, D. and Crandall, K.A. (1998): Modeltest: testing the model of DNA substitution.
Bioinformatics: 14, pp. 817-818.
Praeger, R.L. (1932): An account of the Sempervivum group. Royal Horticulturae Society London.
Purugganan, M.D., Rounsley, S.D., Schmidt, R.J. and Yanofsky, M.F. (1995): Molecular evolution of
flower develpoment: diversification of the plant MADS-box regulatory gene family. Genetics: 140, pp.
345-356.
Purugganan, M.D. and Robichaux, R.H. (2005): Adaptive radiation and regulatory gene evolution in
the Hawaiian silversword alliance (Asteraceae). Missouri Botanical Garden: 92, pp. 28-35.
Quattrocchio, F., Wing, J., van der Woude, K., Souer, E., de Vetten, N., Mol, J. and Koes, R. (1999):
Molecular analysis of the anthocyanin2 gene of Petunia and its role in the evolution of flower color.
The Plant Cell: 11, pp. 1433-1444.
Rambaut,
A.
(1996):
Se-Al:
Sequence
Alignment
Editor.
Available
from:
http://tree.bio.ed.ac.uk/software/seal/
Remington, D.L. and Purugganan, M.D. (2002): GAI homologues in the Hawaiian silversword alliance
(Asteraceae-Madiinae): molecular evolution of growth regulators in a rapidly diversifying plant lineage.
Molecular Biology and Evolution: 19, pp. 1563-1574.
Ronquist, F. and Huelsenbeck, J.P. (2003): MrBayes 3: Bayesian phylogenetic inference under mixed
models. Bioinformatics: 19, pp. 1572-1574.
120
Literature
Saedler, H., Becker, A., Winter, K.-U., Kirchner, C. and Theissen, G. (2001): MADS-box genes are
involved in floral development and evolution. Acta Biochimica Polonica: 48, pp. 351-358.
Sanderson, M.J. (1997): A nonparametric approach to estimating divergence times in the absence of
rate constancy. Molecular Biology and Evolution: 14, pp. 1218-1231.
Sanderson, M.J. (2003): r8s: inferring absolute rates of molecular evolution and divergence times in
the absence of a molecular clock. Bioinformatics: 19, pp. 301-302.
Sang, T. (2002): Utility of low-copy nuclear gene sequences in plant phylogenetics. Critical Reviews in
Biochemistry and Molecular Biology: 37, pp. 121-147.
Sanmartín, I., van der Mark, P. and Ronquist, F. (2008): Inferring dispersal: a Bayesian approach to
phylogeny-based island biogeography, with special reference to the Canary Islands. Journal of
Biogeography: 35, pp. 428-449.
Schemske, D.W. and Bradshaw Jr., H.D. (1999): Pollinator preference and the evolution of floral traits
in monkeyflowers (Mimulus). Proceedings of the National Academic Society of the USA: 96, pp.
11910-11915.
Schönfelder, P. and Schönfelder, I. (1997): Die Kosmos-Kanarenflora. Franckh-Kosmos, Stuttgart.
Shan, H., Zhang, N., Liu, C., Xu, G., Zhang, J., Chen, Z. and Kong, H. (2007): Patterns of gene
duplication and functional diversification during the evolution of the AP1/SQUA subfamily of plant
MADS-box genes. Molecular Phylogenetics and Evolution: 44, pp. 26-41.
Small, R.L. and Wendel, J.F. (2002): Differential evolutionary dynamics of duplicated paralogous Adh
loci in allotetraploid cotton (Gossypium). Molecular Biology and Evolution: 19, pp. 597-607.
Small, R.L., Cronn, R.C. and Wendel, J.F. (2004): Use of nuclear genes for phylogeny reconstruction
in plants. Australian Systematic Botany: 17, pp. 145-170.
Stern, A., Doron-Faigenboim, A., Erez, E., Martz, E., Bacharach, E. and Pupko, T. (2007): Selecton
2007: advanced models for detecting positive and purifying selection using a Bayesian inference
approach. Nucleic Acids Research: 35, W506-W511.
Stevens, J.F. (1995): Chemotaxonomy of the Eurasian Sedoideae and Sempervivoideae. In: ‘t Hart, H.
and Eggli, U. (eds): Evolution and Systematics of Crassulaceae. Backhuis Publishers, Leiden.
Stryer, L. (1995): Biochemistry. 4.ed. Freeman, New York.
121
Literature
Swofford, D.L. (2002): PAUP*. Phylogenetic analysis using parsimony (*and other methods), version
4. Sinauer, Sunderland.
Syring, J., Willyard, A., Cronn, R. and Liston, A. (2005): Evolutionary relationships among Pinus
(Pinaceae) subsections inferred from multiple low-copy nuclear loci. American Journal of Botany: 92,
pp. 2086-2100.
Taberlet, P., Gielly, L., Pautou, G. and Bouvet, J. (1991): Universal primers for amplification of three
non-coding regions of chloroplast DNA. Plant Molecular Biology: 17, pp. 1105-1109.
Tamura, K., Dudley, J., Nei, M. and Kumar, S. (2007): MEGA4: Molecular Evolutionary Genetics
Analysis (MEGA) software version 4.0. Molecular Biology and Evolution: 24, pp. 1596-1599.
Tenhunen, J.D., Tenhunen, L.C., Ziegler, H., Stichler, W. and Lange, O.L. (1982): Variation in carbon
isotope ratios of Sempervivoideae species from different habitats of Tenerife in the spring. Oecologia:
55, pp. 217-224.
Theissen, G., Becker, A., di Rosa, A., Kanno, A., Kim, J.T., Münster, T., Winter, K.-U. and Saedler, H.
(2000): A short history of MADS-box genes in plants. Plant Molecular Biology: 42, pp. 115-149.
Theissen, G. (2001): Development of floral organ identity: stories from the MADS house. Current
Opinion in Plant Biology: 4, pp. 75-85.
Theissen, G. (2005): On the origin of flowers: molecular views. Nova Acta Leopoldonina: 342, pp. 209219.
Tiffin, P. and Hahn, M.W. (2002): Coding sequence divergence between two closely related plant
species: Arabidopsis thaliana and Brassica rapa ssp. pekinensis. Journal of Molecular Evolution: 54,
pp. 746-753.
Trusty, J.L., Olmstead, R.G., Santos-Guerra, A., Sá-Fontinha, S. and Francisco-Ortega, J. (2005):
Molecular phylogenetics of the Macaronesian-endemic genus Bystropogon (Lamiaceae): palaeoislands, ecological shifts and interisland colonizations. Molecular Ecology: 14, pp. 1177-1189.
Tu, T., Dillon, M.O., Sun, H. and Wen, J. (2008): Phylogeny of Nolana (Solanaceae) of the Atacama
and Peruvian deserts inferred from sequences of four plastid markers and the nuclear LEAFY second
intron. Molecular Phylogenetics and Evolution: 49, pp. 561-573.
Uhl, C.H. (1961): The chromosomes of the Sempervivoideae (Crassulaceae). American Journal of
Botany: 48, pp. 114-123.
122
Literature
Vandenbussche, M., Theissen, G., van der Peer, Y. and Gerats, T. (2003): Structural diversification
and neo-functionalization during floral MADS-box gene evolution by C-terminal frameshift mutations.
Nucleic Acids Research: 31, pp. 4401-4409.
van Ham, R.C.H.J. and ‘t Hart, H. (1998): Phylogenetic relationships in the Crassulaceae inferred from
chloroplast DNA restriction-site variation. American Journal of Botany: 85, pp. 123-134.
Vergara-Silva, F., Martínez-Castilla, L. and Alvarez-Buylla, E.R. (2000): MADS-box genes:
development and evolution of plant body plans. Journal of Phycology: 36, pp. 803-812.
Wang, W.-K., Schaal, B.A., Chiou, Y.-M., Murakami, N., Ge, X.-J., Huang, C.-C. and Chiang, T.-Y.
(2007): Diverse selective modes among orthologs/paralogs of the chalcone synthase (Chs) gene
family of Arabidopsis thaliana and its relative A. halleri ssp. gemmifera. Molecular Phylogenetics and
Evolution: 44, pp. 503-520.
White, T.J., Bruns, T., Lee, S. and Taylor, J. (1990): Amplification and direct sequencing of fungal
ribosomal RNA genes for phylogenetics. In: Innis, M.A., Gelfand, D.H., Sninsky, J.J. and White, T.J.
(eds): PCR Protocols. A Guide to Methods and Applications. Academic Press, San Diego.
Whitkus, R. (1998): Genetics of adaptive radiation in Hawaiian and Cook Islands species of
Tetramolopium (Asteraceae). II. Genetic linkage map and its implications for interspecific breeding
barriers. Genetics: 150, pp. 1209–1216.
Whittall, J.B., Medina-Marino, A., Zimmer, E.A. and Hodges, S.A. (2006): Generating single-copy
nuclear gene data for a recent adaptive radiation. Molecular Phylogenetics and Evolution: 39, pp. 124134.
Whittall, J.B. and Hodges, S.A. (2007): Pollinator shifts drive increasingly long nectar spurs in
columbine flowers. Nature: 447, pp. 706-710.
Yang, J., Huang, J., Gu., H., Zhong, Y. and Yang, Z. (2002): Duplication and adaptive evolution of the
chalcone synthase genes of Dendranthema (Asteraceae). Molecular Biology and Evolution: 19, pp.
1752-1759.
Yang, Z. (2005): The power of phylogenetic comparison in revealing protein function. Proceedings of
the National Academy of Sciences of the USA: 102, pp. 3179-3180.
Zhang, L., Vision, T.J. and Gaut, B.S. (2002): Patterns of nucleotide substitution among
simultaneously duplicated gene pairs in Arabidopsis thaliana. Molecular Phylogenetics and Evolution:
19, pp. 1464-1473.
123
Literature
Zhang, J. (2003): Evolution by gene duplication: an update. TRENDS in Ecology and Evolution: 18,
pp. 292-298.
Zhang, W., Xiang, Q.-Y., Thomas, D.T., Wiegmann, B.M., Frohlich, M.W. and Soltis, D.E. (2008):
Molecular evolution of PISTILLATA-like genes in the dogwood genus Cornus (Cornaceae). Molecular
Phylogenetics and Evolution: 47, pp. 175-195.
Websites:
1)
http://www.ncbi.nlm.nih.gov/ (last access: 09/2009)
2)
http://selecton.tau.ac.il/ (last access: 09/2009)
3)
http://foo.maths.uq.edu.au/~huber/bellerophon.pl (last access: 09/2009)
4)
http://www.atgc-montpellier.fr/phyml/(last access: 09/2009)
124
Abbreviations
Abbreviations
aa
amino acid
Adh
alcohol dehydrogenase
AIC
Akaike information criterion
AP1
APETALA1
AP3
APETALA3
AtpB
membrane-bound ATP synthase
At
Arabidopsis thaliana
BI
Bayesian Interference
bp
base pair
bs
bootstrap support
CAB9
CHLOROPHYLL A/B BINDING PROTEIN9
CAL
CAULIFLOWER
CAM
Crassulacean Acid Metabolism
cDNA
complementary DNA
Ces
carboxylesterase
CHS
chalcone synthase
cpDNA
chloroplast DNA
CYC
CYCLOIDEA
DEF
DEFICIENS
DICH
DICHOTOMA
DNA
deoxyribonucleic acid
F
Fuerteventura
G
La Gomera
GAI
gibberellic acid insensitive
GC
Gran Canaria
GLO
GLOBOSA
GTR
General Time Reversible Model
H
El Hierro
HKY
Hasegawa-Kishino-Yano-85
HSA
Hawaiian silversword alliance
I
invariate
IPTG
isopropyl-beta-D-thiogalactopyranoside
ITS
internal transcribed spacer
K81uf
Kimura-3-Parameter with unequal frequencies
knox
knotted related homeobox
LB
Luria-Bertani
LRT
likelihood ratio test
matK
maturase K
MCS
Macaronesian Crassulaceae Sempervivoideae
ML
Maximum likelihood
MO
Morocco
MP
Maximum parsimony
mRNA
messenger RNA
My
million year
125
Abbreviations
NCBI
National Center for Biotechnology Information
NEB
New England Biolab
NJ
Neighbor-joining
NNI
Nearest Neighbor Interchange
NPRS
Nonparametric rate smoothing
nr
nuclear
OG
outgroup
ORF
open reading frame
P
La Palma
PCR
polymerase chain reaction
PEP
phosphoenolpyruvate
PEPC
phosphoenolpyruvate carboxylase
PHT
Partition Homogeneity Test
PI
PISTILLATA
PL
Penalized likelihood
pp
posterior probabilities
QTL
quantitative trait loci
RbcL
RUBISCO large (subunit)
rDNA
ribosomal DNA
R-gene
resistence gene
RNA
ribonucleic acid
RRT
Relative Rate Test
RUBISCO
ribulose-1,5-bisphosphate
sect
section
ser
series
SMNS
State Museum of Natural History Stuttgart
SPR
Subtree Pruning and Regrafting
SQUA
SQUAMOSA
T
Tenerife
TBR
tree bisection reconnection
TE
Tris-EDTA (ethylenediaminetetraacetic acid)
TFL
TERMINAL FLOWER
TIM
Transitional Model
trnF
transfer RNA for Phenylalanine
trnL
transfer RNA for Leucine
TrN
Tamura-Nei
TVM
Transversional Model
UTR
untranslated region
ω
omega = Ka/Ks
X-Gal
bromo-chloro-indolyl-galactopyranoside
126
Overview of scientific contributions
Overview of scientific contributions
Talks:
Esfeld, K. and Thiv, M. (2004): Studying adaptive radiation at the molecular level: a case study in the
Macaronesian Crassulaceae-Sempervivoideae. 7. Annual Meeting of the GfBS. Stuttgart.
Esfeld, K. (2006): The process of speciation: evidence from an adaptive radiation. Vavilov-Seminar,
IPK Gatersleben.
Esfeld, K., Thiv, M. and Koch, M. (2008): The use of nuclear coding genes for phylogenetic
reconstructions in an adaptive radiation. Systematics. Göttingen.
Thiv, M., Esfeld, K. and Koch, M. (2009): The impact of floral genes in the evolution of an adaptive
radiation. Systematics 2009, Leiden.
Poster:
Esfeld, K. and Thiv, M. (2005): A comparison of floral morphology and physiology genes in an
adaptive radiation. XVII International Botanical Congress. Vienna.
Esfeld, K. and Thiv, M. (2006): The evolution of the Macaronesian Crassulaceae-Sempervivoideae:
th
Studying adaptive radiation at the molecular level. 17 International Symposium Biodiversity and
Evolutionary Biology, Bonn.
Publications:
Esfeld, K., Koch, M.A., van der Niet, T., Seifan, M. and Thiv, M. (2009): Little interspecific pollen
transfer despite overlap in pollinators between sympatric Aeonium (Crassulaceae) species pairs.
Flora. doi:10.1016/j.flora.2008.10.002.
Thiv, M., Esfeld, K. and Koch, M. (2009, in press): Studying “adaptive radiation” at the molecular level:
a case study in the Macaronesian Crassulaceae-Sempervivoideae. In Glaubrecht, M. and Schneider,
H. (eds.): Evolution in Action - Adaptive Radiations and the Origins of Biodiversity. Springer, Hamburg.
Other contributions:
Esfeld, K., Hensen, I., Wesche, K., Jakob, S.S., Tischew, S. and Blattner, F.R. (2008): Molecular data
indicate multiple independent colonizations of former lignite mining areas in Eastern Germany by
Epipactis palustris (Orchidaceae). Biodiversity and Conservation: 17, pp. 2441-2453.
Tadele, Z. and Esfeld, K. (2008): Applications of TILLING to the understudied crops from Africa: the
case of tef. FAO/IAEA - International Symposium on Induced Mutations in Plants, Vienna. Poster.
Esfeld, K., Plaza, S. and Tadele, Z. (2009): Bringing high-throughput techniques to orphan crop of
Africa: Highlights from the Tef TILLING Project. Gene Conserve: 8, pp: 783-788.
Tadele, Z., Esfeld, K. and Plaza, S. (2009): Applications of high-throughput techniques to the
understudied crops of Africa. Conference on Agriculture: Africa's "Engine for Growth" - Plant science &
biotechnology hold the key, Rothamsted Research, Harpenden, Herts.
Tadele, Z., Esfeld, K. and Plaza, S. (2009): Employing green revolution genes to improve orphan crop
tef. Tadele, Z. (ed.): New approaches to plant breeding of orphan crops in Africa: Proceedings of an
International Conference, 19-21 September 2007, Bern, Switzerland.
127
Appendices
Appendices
1.) Datasets and alignments on CD
A)
PEPC_full_length_Diss.fas
B)
PEPC_full_length_Diss.nex
C)
PEPC_exon_Diss.fas
D)
PEPC_exon_Diss.nex
E)
AP1_full_length_with_cDNA_Diss.fas
F)
AP1_full_length_with_cDNA_Diss.nex
G)
AP1_full_length_without_cDNA_Diss.fas
H)
AP1_full_length_without_cDNA_Diss.nex
I)
AP1_exon_Diss.fas
J)
AP1_exon_Diss.nex
K)
AP3_full_length_Diss.fas
L)
AP3_full_length_Diss.nex
M)
AP3_exon_Diss.fas
N)
AP3_exon_Diss.nex
O)
Thesis PDF
2.) Comparison of infrageneric classification (table 15).
3.) Location of the studied species (table 16).
4.) Overview of the used primers (table 17).
5.) Fig. 29-43: BI and ML phylograms based on various datasets.
6.) Fig. 44: BI phylogram based on the exon dataset of MCS_AP1 including all
amplified unique sequences.
7.) Fig. 45: NJ phylogram based on the enlarged exon MCS_PEPC dataset.
8.) Fig. 46: Ultrametric tree.
128
Appendices
Table 15: Comparison of the infrageneric classification of Aeonium: Lems (1960), Liu (1989), Mes
(1995), and Mort et al. (2002). Studied species are indicated in bold.
Lems (1960)
A. aizoon
A. aureum
A. arboreum
A. balsamiferum
A. canariense
A. castello-paivae
A. ciliatum
A. cuneatum
A. davidbramwellii
A. decorum
A. diplocyclum
A. dodrantale
A. glandulosum
A. glutinosum
A. gomerense
A. goochiae
A. gorgoneum
A. haworthii
A. hierrense
A. holochrysum
A. korneliuslemsii
A. lancerottense
A. leucoblepharum
A. lindleyi
A. mascaense
A. nobile
A. palmense
A. percarneum
A. pseudourbicum
A. rubrolineatum
A. saundersii
A. sedifolium
A. simsii
A. smithii
A. spathulatum
A. stuessyi
A. subplanum
A. tabuliforme
A. urbicum
A. undulatum
A. valverdense
A. vestitum
A. virgineum
A. viscatum
A. volkeri
Liu (1989)
Mes (1995)
Sect. Holochrysa
Sect. Holochrysa
Sect. Canariensia
Sect. Urbica
Sect. Urbica
Sect. Canariensia
Sect. Urbica
Sect. Urbica
Sect. Aeonium
Sect. Aeonium
Sect. Patinaria
Sect. Leuconium
Sect. Leuconium
Sect. Patinaria
Sect. Leuconium
Sect. Leuconium
Sect. Canariensia
Sect. Canariensia
Sect. Urbica
Sect. Goochiae
Sect. Holochrysa
Sect. Urbica
Sect. Urbica
Sect. Holochrysa
Sect. Holochrysa
Sect. Urbica
Sect. Holochrysa
Sect. Goochiae
Sect. Patinaria
Sect. Pittonium
Sect. Leuconium
Sect. Petrothamnium
Sect. Pittonium
Sect. Leuconium
Sect. Leuconium
Sect. Aeonium
Sect. Aeonium
Sect. Leuconium
Sect. Pittonium
Sect. Petrothamnium
Sect. Greenovia
Sect. Greenovia
Sect. Aeonium
Sect. Aeonium
Sect. Canariensia
Sect. Leuconium
Sect. Leuconium
Sect. Canariensia
Sect. Leuconium
Sect. Leuconium
Sect. Greenovia
Sect. Greenovia
Sect. Patinaria
Sect. Pittonium
Sect. Leuconium
Sect. Goochiae
Sect. Aeonium
Sect. Leuconium
Sect. Leuconium
Sect. Aeonium
Sect. Aeonium
Sect. Leuconium
Sect. Aeonium
Sect. Goochiae
Sect. Megalonium
Sect. Canariensia
Sect. Urbica
Sect. Megalonium
Sect. Patinaria
Sect. Leuconium
Sect. Leuconium
Sect. Canariensia
Sect. Leuconium
Sect. Holochrysa
Sect. Goochiae
Sect. Goochiae
Sect. Goochiae
Sect. Goochiae
Sect. Goochiae
Sect. Holochrysa
Sect. Canariensia
Sect. Canariensia
Sect. Urbica
Sect. Holochrysa
Sect. Urbica
Sect. Aeonium
Sect. Petrothamnium
Sect. Petrothamnium
Sect. Chrysocome
Sect. Chrysocome
Sect. Chrysocome
Sect. Pittonium
Sect. Patinaria
Sect. Patinaria
Sect. Leuconium
Sect. Aeonium
Sect. Leuconium
Sect. Aeonium
Sect. Petrothamnium
Sect. Petrothamnium
Sect. Aeonium
Sect. Chrysocome
Sect. Chrysocome
Sect. Aeonium
Sect. Canariensia
Sect. Canariensia
Sect. Leuconium
Sect. Aeonium
Sect. Leuconium
Sect. Canariensia
Sect. Goochiae
Sect. Patinaria
Sect. Petrothamnium
Sect. Canariensia
Sect. Goochiae
129
Mort et al.
(2002;
cpDNA/nrITS)
Clade 2
Clade 2
Clade 3
Clade 1
Clade 4
Clade 4
Clade 4
Clade 4
Clade 2
Clade 3
Clade 4
Clade 4
Clade 2
Clade 3
Clade 4
Clade 4
Clade 3
Clade 3
Clade 4
Clade 3
Clade 2
Clade 4
Clade 4
Clade 1
Clade 4
Clade 4
Clade 3
Clade 2
Clade 3
Clade 4
Clade 1
Clade 1
Clade 4
Clade 3
Clade 3
Clade 1
Clade 2
Clade 4
Appendices
Table 16: Location of the collection sites of the studied species. Indicated are area and name of the
collection site as well as coordinates in Degrees Minutes and Seconds.
species
area
location
N
W
remarks
A. aureum
T
Los Carrizales
28° 11' 27.47"N
16° 30' 40.82"W
A. canariense
T
San Andres
28° 18' 48.18"N
16° 06' 51.23"W
A. cuneatum
T
El Bailadero
28° 19' 28.63"N
16° 07' 23.83"W
A. goochiae
P
Los Tilos
provided by R. Lösch
A. nobile
P
provided by R. Lösch
A. rubrolineatum
G
Barranco de las
Angustias
Vallehermoso
28° 05' 41.63"N
17° 10' 26.96"W
A. saundersii
G
Lomito
28° 03' 57.06"N
17° 05' 39.25"W
A. smithii
T
Vila Flor
28° 05' 41.65"N
16° 22' 53.63"W
Ai. laxum
T
Genoves by Icod
28° 12' 57.22"N
16° 26' 39.54"W
Ai. pachycaulon
T
Taborno
28° 19' 18.77"N
16° 09' 38.57"W
M. anagensis
T
Taganana
28° 19' 49.02"N
16° 07' 19.43"W
M. icterica
T
Tabaiba
28° 11' 34.95"N
16° 30' 36.48"W
Sardinia
Arzana
39°55' N
9° 31' O
S. jaccardianum
MO
Bekrite
33° 00' 55.25"N
5° 08' 30.73"W
S. modestum
MO
Taddert
31° 24' 20.47"N
4° 06' 18.29"W
S. pubescens
Tunesia
S. surculosum
MO
City Succulent
Collection Zurich
Oukaimeden
two extractions,
cultivated from
collected seeds
RS-TAV 996096/0
31° 06' 15.37"W
7° 30' 40.97"W
two extractions
S. caeruleum
130
provided by R.
Lübenau-Nestle
two extractions
Appendices
Table 17: Primers used to amplify the respective gene regions. Indicated are name, gene region,
sequence, and application.
primer name
M13_F (Promega)
M13_R (Promega)
Jet_F (Fermentas)
Jet_R (Fermentas)
PEPC-F
PEPC-R
PEPC_Sed_for_2
PEPC_Sed_rev
PEPC_446F
PEPC_968R
PEPC_1331R
PEPC_342F
PEPC_293F_Sed
poly(T)-AP1
AP1MDS2
AP1MDS3
SQUAR
AP1-noT
AP1-11F
AP1-704R
AP1_Sed_for
AP1_Sed_rev
AP1_238F
AP1_286R
AP1_1631R
AP1_379R
AP1_482R
AP1_515
AP1_5567R
AP1_251F
AP1_559F
AP1_375SeqF
Seq_Sed_AP1_R
AP3-polydT
ATG3
MADS4
MADS4-Aeo
AP3-11F
AP3-noT2
AP3-351F
PI-F-Aeo
AP3-724R
AP3-82F
AP3-1766R
AP3_399R
AP3_570F
AP3_586SeqF
AP3_1109SeqR
region
vector
vector
vector
vector
PEPC
PEPC
PEPC
PEPC
PEPC
PEPC
PEPC
PEPC
PEPC
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP1
AP3
AP3
AP3
AP3
AP3
AP3
AP3
AP3
AP3
AP3
AP3
AP3
AP3
AP3
AP3
sequence
GTAAAACGACGGCCAG
CAGGAAACAGCTATGAC
GCCTGAACACCATATCCATCC
GCAGCTGAGAATATTGTAGGAGATC
TCWGATTCAGGAAAAGATGC
GCAGCGATRCCCYTCATTGT
TCWGATTCAGGAAAAGATGCWGG
GCAGCGATRCCCYTCATTGTCAA
TGTCGCCACAGAGCAMTATC
GTATGCATCACGCAGACGMA
GCTTGGCATACATTTAGTGT
CCTCATGGATGAAATGGC
AGCATGGYATGAATCCWCCT
GACTCGAGTCGACATCGATTTTTTTTTTTTTTTTT
TGGNYTKNTSAAGAARGCTCATGA
GTNCARYTNARRMGNATNGARAAYAAGAT
GCAAAGCATCCMAKATGGCATG
GACTCGAGTCGACATCGA
AAGAAGGCTCATGAGATCTC
TGCTCATTCTCCCTTCATCTTC
GAGATCTCYGTCTTGTGTGATGC
CTCCCTTCATCTTCTCCCTGGTAAG
GGCCAAGCTTGATCTCTTGCAGA
TGTAACTCTCTCATGCTCAATGC
AAGCTCAGAAATTGACTCATGC
GCTCAGAAATTGACTCATGCATA
CACTGTGCTTCGTCTTGCTGCAC
GCACTTAAGGCTCAATGCTT
TTIATSCAGCAAAGCATCCAAG
AAAATCACAGGCATTACTTGGG
CCGCTTAGCTGTAGAAACTGAG
TGCACTTTGTTTTTGGACTTG
AACATGGCTGTGATMTTTACACA
CCGGATCCTCTAGAGCGGCCGCTTTTTTTTTTTTTTTT
ATGGSIMGIMMIAARATISARAT
AAYMGRCARGTIACITWYAARMGRMG
AATAGGCAGGTGACGTTYAAGAG
TGACGTTTAAGAGGCGGAAC
ATCCTCTAGAGCGGCCGC
AGCTATTCGTGCTCGCAAGT
AATAGGCAGGTGACGTTYTCGAAGCGGAG
TTGGCAAAAACAACGAAACA
AARAAGGCAGAGGAGCTYAC
TCAGTTACCACTCATGAGRGTGTAA
TCAGTTCTGCTTGAAATTGCTT
GCAAGAAACTTTGAGGAAAGTGA
AATGCAAGAAACTTTGAGGAA
TCKGCTTGAAATTGCTTTG
131
applied for
vector
vector
vector
vector
amplification
amplification
amplification
amplification
sequencing
sequencing
sequencing
sequencing
sequencing
cDNA
cDNA
cDNA
cDNA
cDNA
amplification
amplification
amplification
amplification
sequencing
sequencing
sequencing
sequencing
sequencing
sequencing
sequencing
sequencing
sequencing
sequencing
sequencing
cDNA
cDNA
cDNA
cDNA
cDNA
cDNA
cDNA
amplification
amplification
amplification
amplification
sequencing
sequencing
sequencing
sequencing
Appendices
Fig. 29: ML phylogram based on the MCS_PEPC full-length data. Bootstrap support is given at the
nodes.
132
Appendices
Fig. 30: ML phylogram based on the MCS_PEPC exon data. Bootstrap support is given at the nodes.
133
Appendices
Fig. 31: BI phylogram based on the MCS_PEPC intron data. Posterior probabilities are given at the
nodes.
134
Appendices
Fig. 32: BI phylogram based on the MCS_PEPC MCS intron data (Sedum sequences were excluded).
Posterior probabilities are given at the nodes.
135
Appendices
Fig. 33: ML phylogram based on the MCS_PEPC MCS intron data (Sedum sequences were
excluded). Bootstrap support is given at the nodes.
136
Appendices
Fig. 34: BI phylogram based on the MCS_AP1 full-length data. Posterior probabilities are given at the
nodes.
137
Appendices
Fig. 35: BI phylogram based on the MCS_AP1 exon data. Posterior probabilities are given at the
nodes.
138
Appendices
Fig. 36: ML phylogram based on the MCS_AP1 intron data. Bootstrap support is given at the nodes.
139
Appendices
Fig. 37: BI phylogram based on the MCS_AP1 MCS intron data (Sedum sequences were excluded).
Posterior probabilities are given at the nodes.
140
Appendices
Fig. 38: ML phylogram based on the MCS_AP1 MCS intron data (Sedum sequences were excluded).
Bootstrap support is given at the nodes.
141
Appendices
Fig. 39: BI phylogram based on the MCS_AP3 full-length data. The 3´-UTR-region is excluded.
Posterior probabilities are given at the nodes.
142
Appendices
Fig. 40: ML phylogram based on the MCS_AP3 exon data. Bootstrap support is given at the nodes.
143
Appendices
Fig. 41: ML phylogram based on the MCS_AP3 intron data. Bootstrap support is given at the nodes.
144
Appendices
Fig. 42: BI phylogram based on the MCS_AP3 MCS intron data (Sedum sequences were excluded).
Posterior probabilities are given at the nodes.
145
Appendices
Fig. 43: ML phylogram based on the MCS_AP3 MCS intron data (Sedum sequences were excluded).
Bootstrap support is given at the nodes.
146
Appendices
Fig. 44: BI phylogram based on exon data of MCS_AP1 including all amplified sequences (chimers,
sequences with unique splice sites or introns, and with premature stops caused by frameshift
mutations). Posterior probabilities are given at the nodes.
147
Appendices
Fig. 45: NJ phylogram of the enlarged MCS_PEPC dataset. Bootstrap support is given at the nodes.
148
Appendices
Fig. 46: Ultrametric tree. Posterior probabilities are given at the nodes. Asterisks mark the positions of
potential alternative duplication events.
149
Acknowledgment
Acknowledgment
Mein erster Dank gilt Herrn Prof. Dr. Marcus Koch, der die Betreuung der Arbeit
übernommen und damit die Promotion möglich gemacht hat. Ganz besonders
herzlich danke ich ihm für seine Geduld während der Fertigstellung der Arbeit sowie
für die Diskussionen und Denkanstösse.
Auch bei Frau Prof. Dr. Claudia Erbar bedanke ich mich sehr herzlich für die
unkomplizierte Zusammenarbeit und für die freundliche Übernahme der
Zweitkorrektur.
Bei Dr. Mike Thiv bedanke ich mich für die Überlassung des Themas, für die
Bereitstellung des Labor- und des Arbeitsplatzes am Naturkundemuseum Stuttgart.
Ich danke ihm für zahlreiche Diskussionen und das Korrekturlesen der Arbeit sowie
die eingebrachten Verbesserungsvorschläge.
Mein ganz besonders herzlicher Dank gilt Dr. Frank Blattner – ohne ihn wäre die
Arbeit nie beendet worden. Ich danke ihm sehr für die vielen hilfreichen und
intensiven Diskussionen und Denkanstösse, die er mir unermüdlich während der
Entstehung der Arbeit gab. Er war für meine Sorgen und Probleme ansprechbar und
hat damit wesentlich zum Gelingen der Arbeit beigetragen. Auch danke ich ihm von
ganzem Herzen für das Asyl, dass er mir immer mal wieder in seiner Arbeitsgruppe
gegeben hat. Auch das ausdauernde Korrekturlesen war von Nöten wie Willkommen.
Ganz besonders bedanke ich mich auch bei Dr. Zerihun Tadele, meinem „neuen“
Chef, der immer ein Auge zugedrückt hat wenn auf dem Rechner mal wieder eine
Analyse lief. Damit und auf vielfältige andere Art und Weise motivierte und
unterstützte er mich.
Mein herzlichster Dank gilt meinen Kollegen aus dem Labor in Hohenheim, meinen
„Sequenzierknechten“ Timo, Aelys, Philip, Valentin und Chloe in Zürich und der ETX
in Gatersleben: Petra und Birgit für eure Hilfe und Unterstützung im Labor während
meiner Zeit in Gatersleben; Maia, Nico, Enoch habt Dank für die vielen Diskussionen
und das ihr mir MrBayes und PHYML näher gebracht habt, Maia auch lieben Dank
für die Unterstützung bei r8s. Inga, Martin, Elke, Dominik und Markus für die Zeit am
Museum. Und all meinen Kollegen an der Uni Bern, die mir im letzten Jahr soviel
beigebracht haben. Ohne euch alle wäre die vorliegende Arbeit nicht möglich
gewesen.
Mein ganz besonderer Dank gilt meinen Freunden, die mich in der Zeit unterstützt,
motiviert und immer mal wieder aufgefangen haben: Kerstin, Sabine, Frank, Micha,
Timo, Nicole, Beate; Anna, Katrin und Siobhan habt Dank für das Korrekturlesen und
das ihr Bern zu einer tollen Stadt für mich macht. Katrin, Matti, Christiane und
Thomas vielen lieben Dank für eure Freundschaft und alles was dazu gehört. Dafür
sei auch euch, Anke und Jule, von ganzem Herzen gedankt.
Und von ganzem Herzen danke ich auch meiner Familie für die Unterstützung, dass
ihr mich aufgefangen und ertragen habt, wenn es mal wieder nötig war! Vielen, vielen
lieben Dank für Alles, was ihr für mich tut!!
150
Selbständigkeitserklärung
Selbständigkeitserklärung
Hiermit erkläre ich, dass ich die vorliegende Arbeit selbständig und nur unter
Verwendung der angeführten Quellen und Hilfsmittel angefertigt habe. Wörtlich oder
inhaltlich entnommene Stellen benutzter Werke wurden als solche gekennzeichnet.
Ferner erkläre ich, dass ich an keiner anderen Stelle ein Prüfungsverfahren beantragt
bzw. die Dissertation in dieser oder anderer Form bereits an einer anderen Fakultät
zur Prüfung vorgelegt habe.
Bern, den 27.09.09
151
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement