VYTAUTAS MAGNUS UNIVERSITY LITHUANIAN RESEARCH CENTRE FOR AGRICULTURE AND FORESTRY Rita VERBYLAITƠ EUROPEAN ASPEN (Populus tremula L.) IN LITHUANIA: GENETIC DIVERSITY OF PLUS TREES AND POPULATIONS ASSESSED USING MOLECULAR MARKERS Doctoral dissertation Area of Biomedical Sciences Field of Ecology and Environmental Science (03B) Kaunas, 2015
122
Embed
GENETIC DIVERSITY OF PLUS TREES AND POPULATIONS ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
VYTAUTAS MAGNUS UNIVERSITY
LITHUANIAN RESEARCH CENTRE FOR AGRICULTURE AND FORESTRY
Rita VERBYLAIT
EUROPEAN ASPEN (Populus tremula L.) IN LITHUANIA: GENETIC DIVERSITY OF PLUS TREES AND POPULATIONS
ASSESSED USING MOLECULAR MARKERS
Doctoral dissertation
Area of Biomedical Sciences
Field of Ecology and Environmental Science (03B)
Kaunas, 2015
2
UDK 577.175.1(474.5) Ve-142
The research was carried out at the Lithuanian Research Centre for Agriculture and Forestry in
2006–2015. The right of doctoral studies was granted to Vytautas Magnus University jointly
with Aleksandras Stulginskis University and Lithuanian Research Centre for Agriculture and
Forestry on June 21, 2011, by the decision No. V-1124 of the Minister of Education and Science
of the Republic of Lithuania.
Scientific Supervisors
Prof. habil. dr. Remigijus Ozolin ius – Lithuanian Research Centre for Agriculture and Forestry, Area of Biomedical Sciences, Field of Ecology and Environmental Science (03B), supervised from October of 2006 till March of 2013.
Dr. Virgilijus Baliuckas – Lithuanian Research Centre for Agriculture and Forestry, Area of
Biomedical Sciences, Field of Ecology and Environmental Science (03B), supervised from
March of 2013.
Scientific Consultant
Doc. dr. Sigut Kuusien (Lithuanian Research Centre for Agriculture and Forestry, Area of
material for breeding purposes. Currently the most important application area of the genetic
markers in population biology is assessment of genetic diversity and population structure of
various organisms. RAPD markers have also been extensively used in genetic studies of forest
trees (Mosseler et al. 1992; Tulsieram et al. 1992; Grattapaglia and Sederoff, 1994; Nelson et al.
1994; Chalmers et al. 1994; Isabel et al. 1995; Nesbitt et al. 1995; Schierenbeck et al. 1997), yet
currently reliability of this method is more and more questioned (Pérez et al. 1998; Rabouam et
al. 1999). RAPD markers linked to important genetic traits are often used in plant breeding and
biotechnology, used for generating gene maps (Weising et al. 2005).
1.2.2.2. Microsatellites (SSR’s)
Microsatellite method is a very powerful technique based on PCR. This technique is also known
as analysis of simple sequence repeats (SSRs). As the name indicates, these markers are DNA
fragments composed of simple sequences and tandemly repeated motifs mostly of 1 to 5 bp in
size. These DNA fragments are amplified using PCR. Tandem repeats composed mainly of
dinucleotide CA or GA repeats are common in many eukaryote genomes. This type of repeats is
found in many different genome locations and constitutes specific, repetitive DNA class.
23
Microsatellite repeats commonly are flanked by unique sequences that are found only once in a
genome.
Microsatellite method originally was developed for human genome research (Weber and May,
1989), and later adapted to plants (Morgante and Olivieri, 1993). These markers are useful
because of high polymorphism rate – many different alleles are found in one microsatellite
locus, where each allele has a different number of tandem repeats. Different alleles in a
microsatellite locus are formed due to high mutation rate which arises because of DNA
synthesis “mistakes”. Longer microsatellite loci having more tandem repeats are more
polymorphic (Beckman and Weber, 1992). High polymorphism and co-dominant nature of these
markers makes them highly informative. For example, 8 microsatellite loci with 10 alleles each
may “generate” 83 trillions of possible genotypes. This makes them best markers of choice for
DNA fingerprinting, in forensic, human paternity analysis, and in many more research
applications. In the research of plants these markers are used for determination of mating system
and paternity, as well as for gene flow structure analyses. The microsatellite markers are
extremely useful in tree breeding programs (e.g. for identification of genetically improved seeds
in seed lot).
Microsatellites are detected in all main classes of living organisms so far, and are found more
frequently than it would be predicted solely on base composition (Tautz and Renz, 1984; Epplen
et al. 1993). SSRs are distributed more frequently in non-coding than in protein coding
sequences of the various organisms’ genomes (Wang et al. 1994; Field and Wills, 1996;
Edwards et al. 1998; Metzgar et al. 2000; Wren et al. 2000; Morgante et al. 2002). However,
many tri-nucleotide repeats associated with diseases are found using SSRs in coding sequences
of human genome (Nadir et al. 1996). Different SSR frequencies in coding and non-coding
regions are found because of the selection against frame shift mutations resulting in length
changes in non-triplet repeats in coding regions (Liu et al. 1999; Dokholyan et al. 2000).
Eukaryotic organisms have three times more repeat sequences in protein coding sequences than
prokaryotes. SSR repeat families in prokaryotes and eukaryotes are clustered in non-
homologous proteins. Eukaryotes incorporating more repeats may have an evolutionary
advantage because of faster adaptation to a changing environment (Marcotte et al. 1999).
SSRs are considered evolutionary neutral, but significant part of them is proven to be
functionally significant (Li et al. 2002a); for example, they play a role in chromatin organization
(Cuadrado and Schwarzacher, 1998; Li et al. 2000a; b; c; 2002b; Röder et al. 1998). SSRs are
also significant in DNA organization, as they allow DNA sequence to form simple and complex
loop folding patterns. These patterns can have important regulation function on gene expression
24
(Catasti et al. 1999; Fabregat et al. 2001). Simple sequence multiply mainly constitutes
centromeric and telomeric regions of chromosomes (Centola and Carbon, 1994; Murphy and
Karpen, 1995; Schmidt and Heslop-Harrison, 1996; Brandes et al. 1997; Cambareri et al. 1998;
Areshchenkova and Ganal, 1999). The different organism’s centromeres are composed of SSRs
indicating strong evolutionary link between centromere structure and function (Eichler, 1999).
SSRs are also affecting gene activity and act as transcription elements in promoter regions of
heat-shock protein gene hsp26 in Drosophila (Sandaltzopoulos et al. 1995), in Aspergillus (Punt
et al. 1990), and Phytophthora (Chen and Roxby, 1997). SSRs are acting like transcription
regulatory elements when they are found in introns (Meloni et al. 1998; Gebhardt et al. 1999;
2000). Young et al. (2000b) noticed that triplet SSRs are preferentially located in regulatory
genes related to transcription and signal transduction but are under-represented in genes of
structural proteins. As confirmed in many studies, gene translation is affected by SSRs (Ivanov
et al. 1992; Sandberg and Schalling, 1997; Henaut et al. 1998; Martin-Farmer and Janssen,
1999; Timchenko et al. 1999).
Various SSR functions and effects, their abundance and distribution are associated with their
mutation rates. SSRs mutation rates depend on species, repeat type, loci and alleles, age and sex
(Brock et al. 1999; Hancock, 1999; Ellegren, 2000; Schlötterer, 2000), but in general SSR
mutation rate is very high (10-2–10-6 events per locus per generation) as compared to point
mutations (Li et al. 2002a). SSR mutations predominantly arise as changes in repeat number.
Such high SSR mutation rates can be explained in two ways: a) DNA slippage during DNA
replication (Tachida and Iizuka, 1992) and b) recombination between DNA strands (Harding et
al. 1992).
Microsatellite sequences are usually isolated from genomic libraries by screening them with
specific repeat motifs as probes. Clones having a repeat motif are thereafter sequenced. Non
repetitive DNA sequences flanking repetitive motif region are used to design primers for PCR
amplification. Microsatellite markers are multiallelic, widely dispersed across the genome and
relatively easily scored (Morgante and Oliveri, 1993; Devey et al. 1996; Powel et al. 1996;
Barreneche et al. 1998). The microsatellite primer development scheme is presented in Figure
1.6.
25
Fig. 1.6. Scheme showing development of microsatellite primers (Figure adapted from Young et al. 2000a)
The first microsatellite primers for tree species were developed for Pinus radiata (Smith and
Devey, 1994). Later the number of available primers increased substantially; those were
developed for Quercus spp. (Dow et al. 1995; Barret et al. 1997; Isagi and Suhandono, 1997),
for Eucalyptus spp. (Byrne et al. 1996), for Pinus strobus (Echt et al. 1996), for Picea abies
(Pfeiffer et al. 1997), and for several tropical tree species (Chase et al. 1996; White and Powell,
1997; Dawson et al. 1997; Steinkellner et al. 1997). Single base pair repeat microsatellites were
discovered in pine chloroplast genomes (Vendramin et al. 1996).
26
2. MATERIALS AND METHODS
P. tremula plus trees represent an important share of national Lithuanian forest genetic
resources. Plus trees are selected by Lithuanian State Forest Service as exceptional (both by
quantitative and qualitative characteristics) representatives of autochthonous tree species; those
trees are later managed by local forest enterprises. In 2007, a total number of 137 P. tremula
plus trees were selected in 19 stands in 16 State Forest Enterprises (SFE) and monitored by
Forest State Service (Table 2.1).
2.1. Sampling of P. tremula trees for molecular analysis
All P. tremula plus trees inventoried in Lithuania in 2007 were included into RAPD, SSR and
PCR-RFLP (Narayanan, 1991) studies (n=137). The main data about those plus trees are given
in Table 2.1.
Table 2.1. Basic data about investigated Populus tremula plus trees: numbers, location (geographical coordinates), age and soil type (typology group, according to Vai ys (2006)). The presented data is provided by Lithuanian State Forest Service
a Within the State Forest Enterprises the additional trees were sampled in vicinity to the inventoried plus trees (see Table 2.1.)
Wood and leaf samples of additional 7 trees exhibiting clear P. tremulae infection and 2 sound
looking trees were collected in Biržai, Ignalina, Kretinga, Raseiniai, Rokiškis, Šal ininkai,
Taurag and Utena SFE’s in May–July, 2007. These 9 trees have been selected from the same
forest stands as plus trees and were used as reference material in PCR-RFLP study.
2.2. Assessment of tree and stand characteristics
In order to assess and compare quality traits of P. tremula plus trees, 30 additional European
aspen trees (non-plus trees) per each plus trees containing stand (n=19) were subjected for
detailed evaluation in May–July, 2007. In total, 570 such trees were evaluated. These trees were
randomly selected yet always in a close vicinity to the aspen plus trees (Table 2.1.).
30
Table 2.3. Criteria values for Populus tremula phenotypic evaluation. The scoring system is made according to assessment methodology of forest tree breeding value used by Lithuanian State Forest Service Scale, score Criteria
1 2 3 4 5
Stem form Few bends in less than 1/3 of stem length
One bend in less than 1/3 of stem length
Bend in the lower part of the stem, in more than 1/3 of stem length
Bend in the upper 1/3 part of the stem
Straight
Branch thickness
Thick branches, tubers are present
Medium thick branches
Thin branches – –
Stem presence in crown part
Stem is absent in crown part, crown starts in the lower half of the tree height
Stem is absent in crown part, crown starts in the upper ½ of the tree height
Stem is present only in the lower part of the crown
Stem is present, branching only in the upper part of the crown
Stem is present through the whole crown
Occlusion of branch wounds occurring following branch dieback at the sites of their attachment
Occlusion of wounds is very poor; big wounds are not closed. Stem with nodules.
Occlusion of wounds is poor, big scars are visible on the stem. Stem with nodules.
Occlusion of wounds is intermediate. Wounds of smaller branches are occluded, while occlusion of larger branches is not always successful. Stem is smooth, yet in some places has nodules.
Occlusion of wounds is good; the sites of branch attachment are clearly visible. Stem is smooth.
Occlusion of wounds is very good; it is hard to notice the sites of branch attachment. Stem is smooth.
31
For each tree (plus additionally selected), were assessed: (a) age (based on the latest forest
inventory data), (b) height (assessed using Haglof vertex IV height meter, m), (c) diameter
(measured at breast height using diameter calipers, cm), (d) stem straightness (score values
given in Table 2.3), (e) stem slenderness (assessed in scores, where 5 means lowest slenderness,
and 1 – the most expressed slenderness), (f) height to dry branches (assessed using Haglof
vertex IV, m), (g) height to green branches (assessed using Haglof vertex IV, m), (h) mean
crown diameter (assessed as a mean after measuring in two directions perpendicular to each
other, m), (i) crown form (irregular, umbelliferous, oval, spherical, egg-shaped, narrow
spherical), (j) branch thickness (score values are given in Table 2.3), (k) stem presence in crown
part (score values given in Table 2.3), (l) branch wound occlusion (score values given in Table
2.3), (m) presence of epicormic branches, (n) tree sanitary condition (categories given in Table
2.4).
Additionally, the presence of mechanical injuries on the stems was also recorded. The sanitary
condition of the investigated aspen trees was evaluated according to (1991) and is
presented in Table 2.4.
Table 2.4. Scoring system for sanitary condition of Populus tremula trees (according to , 1991)
Score, points
Category descriptor
Main symptoms Additional symptoms
1 Sound looking
Leaves are green, shiny, the crown dense, the annual growth increment typical to a given species according to age, forest habitat and season.
–
2 Slightly weakened
Leaves are green, the crown slightly defoliated, the annual growth may be smaller than typical, and less than ¼ of the crown is dead.
Some localized damages on branches, presence of mechanical wounds and sporadic epicormic shoots.
3 Weakened
Leaves are smaller or lighter green, premature defoliation, the crown is thin, the dead portion of the crown composes ¼ – ½.
Localized damages are more abundant; primary colonization by stem pests, sap bleeding, and occurrence of epicormic shoots on the stem and branches.
4 Withering away
Leaves are small, lighter green or yellowish, premature defoliation or wilting, the crown is thin, the dead portion of the crown composes ½ – ¾.
Abundant and clearly visible activity of stem insect pests (exit holes, wounds, sap bleeding, sawdust on the bark and in timber); abundant presence of epicormic shoots.
5 Recently dead
Leaves are dry, wilted or prematurely defoliated, the dead portion of the crown composes > ¾, the bark is still present.
Abundant and clearly visible activity of stem insect pests and pathogens.
6 Old snags Leaves and some branches are lost, bark disrupted or crumbled away from the stem.
Visible insect exit holes as well as fungal mycelium and fruiting bodies on the stem, branches and roots.
32
2.3. DNA extraction from plant material
As extraction of good quality and clean DNA is a prerequisite for successful genetic studies, one
of our study objectives was to find the best DNA extraction method suitable for wood and leaf
tissue samples of P. tremula. For this task, we have randomly selected wood and leaf samples
from two European aspen trees.
We tested six well known DNA extraction techniques: SDS isolation, protein precipitation,
CTAB isolation, CTAB precipitation, guanidinium isothiocyanate and alkaline isolation (the full
description of these techniques is given by Milligan, 1998), and four commercially available kits
for extraction of plant genomic DNA: DNA isolation reagent for genomic DNA with Plant AC
reagent (AppliChem, Maryland Heights, USA), Nucleospin Plant Mini (Macherey-Nagel,
Düren, Germany), Genomic DNA purification kit (Fermentas, Vilnius, Lithuania) and
innuPREP Plant DNA Kit (Analytik Jena, Jena, Germany).
Before the DNA extraction samples of wood tissue (100 mg) and silica-dried leaf tissue (10 mg)
were ground using a mortar and a pestle in liquid nitrogen. The resulting powder was
immediately used for DNA extraction. The extraction of total DNA was performed in ten
different ways using the following protocols:
1. SDS isolation of total DNA (Edwards et al. 1991; Goodwin and Lee, 1993):
Transfer the ground sample material to tube and add 4 ml extraction buffer (200 mM
Tris pH 7.5, 25 mM EDTA, 250 mM NaCl, 0.5% (w/v) SDS) for each 10 mg of
tissue (e.g. for leaf sample add 400 l).
Vortex the sample for 5 sec.
Centrifuge at 12,000 x g for 1 min to pellet cellular debris.
Transfer 3 ml (300 l) of the supernatant to a new tube. Add 3 ml (300 l) of
isopropanol and incubate at 20–25°C for 2 min.
Centrifuge at 12,000 x g for 5 min.
Dry the DNA pellet at 20–25°C.
Dissolve the DNA in a 100 l TE.
Use 2.5 l of the dissolved DNA for a typical PCR reaction.
The dissolved DNA may be stored at 4°C for over one year.
2. Isolation of total DNA by protein precipitation (Fang et al. 1992; Dellaporta et al. 1983)
Transfer the ground sample material to a tube containing 1.2 ml (600 l) extraction
buffer (100 mM Tris pH 8.0, 50 mM EDTA pH 8.0, 500 mM NaCl, 2% (w/v) SDS,
33
1% (w/v) PVP-360, 0.1% (w/v) -mercaptoethanol (added immediately prior to use
in a fume hood)) and incubate at 65°C for 20 min.
Add one third of the volume potassium acetate. Shake vigorously and incubate on ice
for 5 min. Most proteins and polysaccharides are removed as a complex with the
insoluble potassium dodecyl sulfate precipitate.
Spin at 12,000 x g for 20 min at 4°C.
Pipette the supernatant into a clean micro centrifuge tube. Try to avoid as much of
the particulate material as possible. Add 0.5 vol. of isopropanol. Mix and incubate
the solution for 1 h at 4°C.
Pellet the DNA at 12,000 x g for 15 min at 4°C. Gently pour off the supernatant and
lightly dry the pellets either by inverting the tubes on paper towels for 10 min or as
long as necessary.
Incubate the DNA in 200–500 l TE at 65°C for 30 min to re-suspend it.
Transfer the solution to a micro centrifuge tube and spin for 5 min at 4°C to remove
any insoluble debris.
Transfer the supernatant to another micro centrifuge tube. Add 0.1 vol. sodium
acetate and two-thirds of the volume of cold isopropanol. Mix well, incubate at 4°C
for 1 h, and pellet the DNA for 10 min in a micro centrifuge at 4°C.
Wash the pellet with 200–500 l cold 80% ethanol for 10 min and centrifuge again
for 1 min at 4°C. Dry the pellet for 10 min in a Speed Vac.
Re-dissolve the DNA in TE using small increments (e.g. 10–100 l) depending on
the size of the pellet.
3. CTAB isolation of total DNA (Doyle and Doyle, 1987) with modifications described by
(Murray and Tompson, 1980; Saghai-Maroof et al. 1984; Rogers and Bendich, 1985;
1988; Doyle and Dickson, 1987; Fang et al. 1992; Lodhi et al. 1994; Milligan, 1998)
Heat the extraction buffer (50 mM Tris pH 8.0, 0.7 M NaCl, 10 mM EDTA, 1%
(w/v) CTAB, 0.5% (w/v) -mercaptoethanol (added immediately prior to use in a
fume hood)) to 60°C.
Immediately transfer the ground sample material to tube containing 1 ml (500 l) of
extraction buffer. Mix well.
Incubate at 60°C for 30–120 min with periodic gentle swirling.
Extract once with 1 ml (500 l) of chloroform: isoamyl alcohol. Mix gently but
thoroughly. Spin at 12,000 x g for 30 sec at 20–25°C to separate the phases.
Avoiding the interface, pipette the aqueous (top) phase into new tubes.
34
Add 0.5 vol. of 5 M NaCl. Add cold isopropanol to 40%. Mix gently to precipitate
nucleic acids. If no precipitate is visible, place at -20°C for 20 min or longer.
Spin at 12,000 x g for 1 min at 20–25°C. If no pellet or precipitate is visible, place on
ice for 20 min and spin again. In extreme case, spin for 10 min at 12,000 x g.
Gently pour off as much of the supernatant as possible without losing the nucleic
acid pellet. Add 0.5–1.0 ml of wash buffer (76% (w/v) ethanol, 10 mM ammonium
acetate) and swirl gently to wash the pellet. Let the nucleic acids sit in the wash
buffer for 15–20 min. Generally, nucleic acids will become much whiter (cleaner) at
this step.
Spin at 12,000 x g for 1 min at 20–25°C. If this is not sufficient, spin harder and
longer as before. Pour off wash buffer and allow the pellet to dry briefly (2–4 min)
by inverting the tube on a paper towel. Be careful that the pellet does not slide out.
Re-suspend DNA in re-suspension buffer (10 mM ammonium acetate, 0.25 mM
EDTA pH 8.0) in small increments (e.g. 10–100 l) depending on the size of the
pellet.
4. CTAB precipitation of total DNA (Bellamy and Ralph, 1968; Murray and Tompson,
1980; Rogers and Bendich, 1985; 1988)
Follow steps 1–6 from the protocol 3.
Add 0.1 vol. of 10% CTAB solution and mix.
Perform a second chloroform extraction as in steps 5 and 6 of protocol 3.
Add an equal volume of precipitation buffer (50 mM Tris pH8.0, 10 mM EDTA, 1%
(w/v) CTAB) to reduce the concentration of NaCl to 0.35 M. Mix gently and
incubate at 20–25°C for 30 min. Note that it is important to measure the sample
volume so that the concentration of NaCl is reduced to the proper level.
Recover the precipitated DNA by centrifugation at 12,000 x g at 20–25°C for 10–60
sec.
Re-hydrate the DNA pellet in 200 l re-suspension buffer (10 mM Tris pH 8.0, 1
mM EDTA pH 8.0, 1 mM NaCl).
Add 2 vol. of cold 100% ethanol and mix gently to precipitate the nucleic acids.
Recover the precipitated DNA by centrifugation at 12,000 x g at 4°C for 5–15 min.
Wash the DNA pellet in 200 l cold 80% ethanol and centrifuge at 12,000 x g at 4°C
for 5 min.
Re-suspend the DNA pellet in re-suspension buffer in small increments (e.g. 10–
100 l) depending on the size of the pellet.
35
At this point it may be necessary to purify the DNA further in a cesium chloride
gradient. This is especially true for those tissues that contain tannins or other
secondary compounds, although more than one chloroform extraction or a final
phenol extraction may be sufficient.
5. Guanidinium isothiocyanate isolation of total DNA (Cox, 1968; Bowtel, 1987;
Chomczynski and Sacchi, 1987; Jeanpierre, 1987; Puissant and Houdebine, 1990;
Chomczynski, 1993; Chomczynski and Mackey, 1995)
Transfer the ground sample material to the tube containing 5 ml (500 l) of
extraction buffer (6 M guanidinium isothiocyanate, 100 mM sodium acetate, pH 5.5).
Incubate at 20–25°C for 10 min. A longer incubation over 1 h with mixing may be
necessary.
Centrifuge at 12 000 x g for 10 min at 4°C to pellet the cellular debris.
Precipitate the DNA from the supernatant by adding 10 ml (1 ml) of 100% ethanol at
20–25oC. Mix by inversion and incubate at 20–25°C for 1–3 min The DNA should
become visible as a fibrous or cloudy precipitate.
Collect the DNA by centrifugation at 1000 x g for 1–2 min at 4°C.
Wash the DNA and precipitate twice with 0.5–1.0 ml of 80% ethanol.
Remove the ethanol wash and allow the DNA precipitate to dry for 5–15 min at 20–
25°C.
Dissolve the DNA to a concentration of 0.25 g/ l in TE or 8 mM NaOH; typically
this entails addition of 200 l solvent. The alkaline solvent may dissolve the DNA
faster and more completely.
If necessary, centrifuge the sample at 12,000 x g for 10 min to remove insoluble
material such as polysaccharides.
If NaOH was used to dissolve the DNA, adjust the pH of the solution to a desired pH
by adding Tris-HCl or Hepes (free acid).
6. Alkaline isolation of total DNA (Wang and Cutler 1993)
Transfer the ground sample material to the tube and add 1 ml (100 l) of 0.5 M
NaOH. Mix well.
Transfer 5 l quickly to a new tube containing 495 l storage buffer (100 mM Tris
pH 8.0, 1 mM EDTA pH 8.0). Mix well.
Use 1 l directly in a PCR reaction.
Store the isolated DNA at -20°C.
36
For DNA extraction using DNA isolation reagent for genomic DNA with Plant AC reagent
(AppliChem, Maryland Height, USA), Nucleospin Plant Mini (Macherey-Nagel, Düren,
Germany), Genomic DNA purification kit (Fermentas, Vilnius, Lithuania) and innuPREP Plant
DNA kit (Analytik Jena, Jena, Germany) follow the manufacturer’s protocols.
The concentration of the extracted DNA was measured using Biophotometer (Eppendorf) at 260
nm wave length. The purity of DNA was measured at 260/280 and 260/230 nm wave lengths.
PCR was performed in total 25 l volume, and consisted of 2 l of extracted DNA
(concentration as originally obtained), 13.8 l ddH2O, 2.5 l dNTP mix (2 mM), 2 l MgCl2, 2.5
l PCR buffer (10x), 0.2 l Taq polymerase (5 U/ l) and 1 l of each trnLUAA F (CGA AAT
CGG TAG ACG CTA CG) and trnFGAA (ATT TGA ACT GGT GAC ACG AG) primers (10
M). The chosen primers were originally described in Shaw et al. (2005). Reaction mixture was
covered with mineral oil (10 l). PCR conditions were as follows: initial denaturation step 80°C
for 5 min; denaturation step 94°C for 1 min; primer annealing step 50°C for 1 min; elongation
step 72°C for 2 min; final elongation step 72°C for 5 min. Steps 2 – 4 were repeated 35 times.
2.4. Assessment of genetic diversity of P. tremula plus trees
2.4.1. RAPD analysis
DNA for European aspen plus tree genetic polymorphism analysis was extracted using
Nucleospin Plant Mini Kit (Macherey-Nagel, Düren, Germany). In order to generate RAPD
profiles we used DNA extracted from P. tremula leaf and wood samples simultaneously. For
RAPD analysis we selected 15 most informative primers (Table 2.5) out of the 60 tested
(Appendix 1). We have evaluated primer informativeness using several criteria: (a) overall DNA
amplification quality, (b) amplification of polymorphic fragments, (c) size differences of
polymorphic fragments, and (d) reproducibility of DNA fragments (Pivorien , 2008). PCR
conditions and fragment separation was performed as described in Žvingila et al. (2002). With
each of the 15 selected RAPD primers, PCR amplification was carried out three times using
DNA extracted from leaf and wood samples.
Table 2.5. RAPD primers used in genetic analysis of Populus tremula trees
a Adapted from the International Populus Genome Consortium IPGC (http://www.ornl.gov./sci/ipgc/ssr_resource.htm) b Adapted from Tuskan et al. (2004)
38
Fig. 2.1. Genetic linkage maps of Populus tremula female parent (Pakull et al. 2009) and Populus consensus map based on a Populus trichocarpa and Populus deltoides background, available on
http://popgenome.ag.utk.edu/cgi-bin/cmap/map_set_info, Yin et al. 2008). Figure adapted from Pakull et al. (2009)
39
Polymerase chain reaction for microsatellite analysis was performed in a total reaction volume
of 15 μl, containing 50 ng of template DNA, 5 pmol of forward fluorescent dye-labeled primer
and 5 pmol of unlabeled reverse primer (Table 2.6.), 0.2 U of Dream Taq DNA Polymerase
(Thermo Fisher Scientific, Vilnius Lithuania), 30 μM of dNTP mix (Thermo Fisher Scientific,
Vilnius Lithuania), 1x DreamTaq buffer (Thermo Fisher Scientific, Vilnius Lithuania). The PCR
was performed on GeneAmp® PCR system 9700 (Applied Biosystems (ABI), Waltham, USA)
under following conditions: 4 min at 94°C, 38 cycles of 94°C for 30 sec, 43–56°C for 45 sec,
72°C for 1 min, final extension of 10 min at 72°C.
All obtained PCR products were mixed together, taking 1 μl of each amplification product. 1 μl
of resulting 5 PCR product’s mix was dissolved in 10 μl of formamide, 0.3 μl of internal
fragment size standard solution (Gene Scan 500 LIZ standard, ABI, Waltham, USA), and
separated using ABI-PRISM® Genetic Analyzer 310 (ABI, Waltham, USA). Data analysis was
carried out using Gene-Mapper software.
2.5. Assessment of incidence of Phellinus tremulae infection in aspen stems
To check for the presence of P. tremulae mycelium in wood samples a PCR-RFLP method was
selected. According to Mallett and Myrholm (1995) and Pollastro et al. (2000) P. tremulae is a
slow growing species on growth media thus it is very likely that other, fast growing fungi may
overgrow it, and in this way false negative results may be obtained. Therefore molecular
detection method was selected.
The presence of P. tremulae in European aspen trees was determined using genomic DNA
extracted from wood samples. For amplification of fungal DNA universal primers ITS1-F
(5‘CTTGGTCATTTAGAGGAAGTAA3’) and ITS4 (5‘TCCTCCGCTTATTGATATGC3’)
were used, which are known to amplify ITS (internal transcribed spacer) region in fungi (White
et al. 1990). PCR conditions were as described by Jasalavich et al. (2000). Amplified DNA
fragments were separated on 3% agarose gel and stained with ethidium bromide. As a reference
(positive control), DNA samples of P. tremulae extracted from a fruiting body were amplified in
the same PCR and loaded on the same gels together with the investigated samples. Fungal DNA
fragments showing the same size as the positive control were cut from the gel, re-amplified with
the same universal ITS1_F and ITS4 primers and without subsequent cleaning steps were used
in the restriction analysis. PCR products were digested with restriction enzymes (Alu I, HaeIII,
TaqI and RsaI) providing identification of samples containing P. tremulae DNA. The restriction
analysis was performed as described by Jasalavich et al. (2000). The identification of other
fungal species obtained from P. tremula wood samples was not performed.
40
2.6. Identification of possible hybridization between P. tremula and hybrid aspens
Two forest stands were studied in Dubrava SFE: one natural spruce stand with an admixture of
P. tremula (75-years-old) and one 31-year-old artificially established hybrid aspen stand (P.
tremula x tremuloides, P. tremula x alba, P. tremuloides x alba). The main characteristics of the
evaluated stands are presented in Figure 2.2.
Figure 2.2. Populus tremula and hybrid aspen stands investigated in Dubrava State Forest Enterprise
The distance between the two investigated stands was approximately 130 meters with about
five-meter-high oak forest plantation in between.
In the European aspen stand leaf flushing phenology was recorded for 30 P. tremula trees, while
in hybrid aspen stand a 5,000 m2 area was defined where flushing phenology of all hybrid aspen
trees (155 in total) was assessed. Leaf flushing phenology was assessed for the whole crown of a
selected aspen trees, based on the most advanced stage and using categorical scale of five
degrees (1–5): (1) buds in winter state (no bud stretching or flushing visible), (2) leaves
emerging on 1–50% of buds, (3) leaves emerging on 51–100% of buds, (4) leaves unfolding and
petiole visible on 51–100% of buds, (5) leaves completely unfolded on 51–100% of buds.
According to this scale, the phenology assessment was carried out four times (on the 22nd, 25th
Dubrava State Forest Enterprise
Vaišvydava forest district
Block No. 49
Compartment No. 2
Compartment area 1.6 hectare
Forest soil type Ncp
Stand height 27.1 m
Stand mean diameter 29 cm
Dubrava State Forest Enterprise
Vaišvydava forest district
Block No. 39
Compartment No. 17
Compartment area 2.3 hectare
Forest soil type Ncp
Stand height 26.2 m
Stand mean diameter 28 cm
41
and 29th of April, and on the 2nd of May, 2014). Subsequently, the four categorical scores were
summarized for each assessed tree, resulting in final flushing score. In relation to the final
flushing score, evaluated trees were grouped into three flowering categories: (a) early, (b)
medium and (c) late flushing. Leaf flushing or bud burst phenology is highly correlated or
synchronized with flowering (Linkosalo, 1999). Further in the text the term ‘flowering’ will be
used instead of ‘leaf flushing’.
Seeds from two P. tremula trees and one hybrid aspen tree per flowering category were sampled
in 2014. Immediately after their collection the seeds were sown and half-sib family seedlings
were raised in a greenhouse during the same year (2014).
Ten leaves from each of 100 randomly selected seedlings in each half-sib family were collected
at the end of the first progeny vegetation season. Leaves were scanned (HP ScanJet scaner) and
measured using WinFolia Pro (S) 2004a computer software. A total of 19 traits were measured
for each leaf (reference to the trait in the text is given in brackets): 1) lamina area, 2) lamina
perimeter (perimeter), 3) vertical length of lamina (v_length), 4) horizontal width of lamina
(h_width), 5) aspect ratio (h_width / v_length), 6) form coefficient (form), 7) blade length, 8)
blade maximum perpendicular width, 9) position of maximal perpendicular width, 10)
perpendicular width at position 50% of blade length (width1), 11) perpendicular width at
position 90% of blade length (width2), 12) lobe angle at position 10% of blade length (angle1),
13) lobe angle at position 25% of blade length (angle2), 14) petiole length, 15) petiole area 16)
number of teeth, 17) average teeth height, 18) average teeth width, 19) envelope area.
2.7. Statistical evaluation
2.7.1. Assessment of genetic variation
RAPD and SSR data analysis was performed using GenAlEx 6.5 computer software (Peakall
and Smouse, 2006; 2012) and main genetic diversity indices were calculated. The number of
polymorphic fragments and their percentage in each investigated population were assessed.
Calculations of allelic frequencies were performed using binary matrix, where the presence of a
fragment (1) means AA or Aa genotypes, and the absence of fragment (0) – aa genotype (Lynch
and Milligan, 1994). Allelic frequency of each RAPD fragment was calculated using a formula:
p = 1 - q (2.1)
here p – the frequency of allele A and q – the frequency of allele a.
42
Allele frequency for codominant (SSR) data for all loci was calculated using a formula:
, (2.2)
here Nxx is the number of homozygotes for allele X (XX), and Nxy is the number of
heterozygotes containing the allele X (Y can be any other allele). N = the number of samples.
Shannon’s information index (I) (Brown and Weir, 1983) for RAPD data was calculated using a
formula:
I = -1(p ln (p) + q ln (q)), (2.3)
here for diploid binary data and assuming Hardy-Weinberg Equilibrium, q = (1 - band
frequency)0.5 and p = 1 - q.
Shannon index (I) for SSR data was calculated using a formula by Brown and Weir (1983):
I = - (pi ln)(pi)), (2.4)
here ln – natural logarithm and pi – frequency of the i-th allele.
For each of the RAPD fragments the number of different alleles (Na), as well as the number of
effective alleles (Ne) was calculated using a formula by Kimura and Crow (1964):
Ne = 1 / (p2 + q2), (2.5)
here p – the frequency of allele A and q – the frequency of allele a. Using obtained results the
mean for each investigated population was calculated.
Number of effective alleles Ne for SSR data was calculated using a formula:
Ne = 1/( pi2) , (2.6)
here pi – frequency of the i-th allele.
Observed heterozygosity (Ho) for codominant data was calculated using a formula:
, (2.7)
here N is a sample size.
Expected heterozygosity (He) for dominant data was calculated using a formula:
He = 2 p q, (2.8)
here p – frequency of allele A and q – frequency of allele a.
43
Later, using the obtained values, unbiased heterozygosity (uHe) for dominant marker data
(RAPD) was calculated using a formula:
uHe = (2N / (2N-1)) He , (2.9)
here N – sample size and He – expected heterozygosity.
For codominant marker (SSR) data expected heterozygosity (He) and later unbiased expected
heterozygosity (uHe) were calculated using the formulas 2.10 and 2.11 respectively:
, (2.10)
here pi is the frequency of the i-th allele for the population.
uHe = (2N / (2N - 1)) He, (2.11)
here N – a sample size and He – expected heterozygosity.
Using expected and observed heterozygosity measures for codominant data a fixation index (F)
was calculated using a formula:
, (2.12)
here He – expected heterozygosity and Ho – observed heterozygosity.
For codominant genetic data at a single locus, the total genetic diversity (heterozygosity) was
divided within and among populations (Hartl and Clark, 1997):
– observed heterozygosity, averaged across subpopulations
– expected heterozygosity, averaged across subpopulations
HT – total expected heterozygosity
, (2.13)
here Ho is observed heterozygosity in a subpopulation i, and k is the total number of tested
subpopulations.
44
, (2.14)
, (2.15)
here He is the expected heterozygosity within a subpopulation s, and pi,s is the frequency of the
i-th allele in a subpopulation s. The summation of the allele frequency squared is over all i-th
alleles to h the max number of alleles.
, (2.16)
here HT is the total expected heterozygosity, and pTi is the frequency of allele i over the total
population. If subpopulation sample sizes are equal then pTI=mean pi, where mean pi is the
frequency of allele i averaged over the subpopulations of equal size.
Using the mean of expected and observed heterozygosity and the total heterozygosity, Wright’s
F statistics (Wright, 1946; 1951; 1965) was calculated using equations:
, (2.17)
here FIS – the inbreeding coefficient within individuals relative to the subpopulation, –
expected heterozygosity averaged across subpopulations and – observed heterozygosity
averaged across subpopulations.
, (2.18)
here FIT –the inbreeding coefficient within individuals relative to the total, HT – total expected
heterozygosity and – observed heterozygosity averaged across subpopulations.
, (2.19)
here FST – the inbreeding coefficient within subpopulations relative to the total, HT – total
expected heterozygosity and – expected heterozygosity averaged across subpopulations.
The deviation of allele frequencies from Hardy-Weinberg’s Equilibrium (HWE) were calculated
using a formula:
45
, (2.20)
here, the sum of i to k genotypes is based on Oi, the observed number of individuals of the i-th
genotype, and Ei, the expected number for the i-th genotype. Ei is calculated as either pi2 for a
homozygous genotype or 2 pq for a heterozygous genotype. Degrees of freedom for the Chi-
Squared test was calculated using a formula DF = [Na(Na-1)]/2, where Na is the number of
alleles at the locus.
Using GenAlEx analysis software the following within-population genetic diversity indices
were calculated:
Nei’s genetic identity (I) (Nei, 1972; 1978):
,
, and , (2.21)
here I is Nei’s genetic identity, and pix and piy are the frequencies of the i-th allele in
populations x and y. For multiple loci, Ixy and Iy are calculated by summing over all loci and
alleles and dividing by the number of loci. These averaged values are then used to calculate I.
Nei’s genetic distance (Nei, 1972; 1978) was calculated using a formula:
, (2.22)
here ln is a natural logarithm and I is Nei’s identity.
Dendrograms were constructed with PAST version 2.17c software (Hammer et al. 2001) using
RAPD and SSR data and UPGMA method (Sneath and Sokal, 1973) (paired group algorithm,
similarity measure – Euclidean distance), also on the basis of Nei’s genetic distance matrices
(Nei, 1978). The significance of clusters was assessed using bootstraps (Suzuki and Shimodaira,
2004). Mantel test between genetic distance matrix and geographic distance matrix was
performed with GenAlEx software (Peakall and Smouse, 2006; 2012) using 9,999 permutations.
PCA (principal component analysis) based on 5 SSR markers was performed with PAST version
2.17c software (Hammer et al. 2001) using two options: within group and among groups.
Broken stick method (Macarthur, 1957) was used to determine how many components are
important. Also, a load of each SSR marker in each of important components was analyzed.
46
Molecular variation in Lithuanian P. tremula populations was assessed by hierarchical AMOVA
(Excoffier et al. 1992; Mengoni and Bazzicalupo, 2002) using GenAlEx v. 6.5 (Peakall and
Smouse, 2006; 2012). To assess genetic distance FST and RST fixation indices were calculated
and total variation was partitioned into among- and within-population effects. The probability P
(rand =data) for the fixation indexes was based on standard permutation across the full data set
with 9,999 permutations.
Bayesian clustering approach was implemented using computer analysis software STRUCTURE
version 2.1 (Pritchard et al. 2000) to estimate the most likely number of clusters (K) into which
the SSR and RAPD multilocus genotypes were assigned with certain likelihoods. The
population priors were not used. A Markov chain with 100,000 and 50,000 iterations for SSR
and RAPD analysis, respectively, following a burn in period of respective 100,000 and 50,000
iterations was used. Each run was replicated 10 times. The most likely number of clusters was
identified by the delta K ( K) criterion. Number of clusters was determined by calculation based
on the second order rate of change of the likelihood ( K) (Evanno et al. 2005).
2.7.2. Correlation between occurrence of certain RAPD fragments and presence of DNA of P. tremulae
For correlation assessment a binary data matrix was used. SAS (SAS 9.4 package, by SAS
Institute Inc., Cary, NC, USA) procedure GENMOD (generalized linear models) with model
options of link function ‘logit’ and the binomial distribution variance function was used for
estimation of locus band presence or absence effect for wood infection with P. tremulae. Only
polymorphic fragments were used for the evaluation of data. In order to check which loci were
the most discriminating the sites or populations, stepwise discriminant analysis was done to
select a subset of the fragments for use in discriminating among the classes. The STEPDISC
SAS procedure was applied to select RAPDs contributing most to the differentiation of
individuals grouped by wood infection with P. tremulae. The least significance level 0.05 was
used for the selection of single RAPD markers. Detrended correspondence analysis performed
with PAST version 2.17c software (Hammer et al. 2001) was applied to detect pattern in wood
infection in the two groups of trees (infected and free of pathogen DNA) by RAPD markers best
discriminating those groups.
47
2.7.3. Analysis of European aspen and hybrid aspen leaf parameters
SAS STEPDISC procedure (slentry=0.05) was used to select leaf traits which could be used in
discriminating European and hybrid aspens. Principal component analysis (PCA) using Pearson
correlation and distance-based biplot was applied for selected leaf traits. The attribution of aspen
individuals to one or another taxa was performed by discriminant procedure in PAST version
2.17c software (Hammer et al. 2001). For analysis of variance PROC MIXED (mixed model
equations) and the REML (restricted maximum likelihood) option in SAS (SAS 9.4 package, by
SAS Institute Inc., Cary, NC, USA) was used. Variance components were calculated for half-sib
families and progeny seedlings using the following model:
ijmijiijm etfy )(
here ijmy is the value of a single observation, is the grand mean, if refers to the random
effect of a family i, )(ijt is the random effect of tree j in family i, and ijme is the random error
term.
48
3. RESULTS AND DISCUSSION
3.1. General evaluation of investigated P. tremula stands and plus trees
Evaluation of sanitary condition of P. tremula stands from which plus trees originate, showed
that stands of the best condition were found in Ignalina, Jurbarkas, Anykš iai, Raseiniai and
Utena SFEs. The highest frequency of trees with P. tremulae basidiocarps was found in
Rokiškis, Marijampol , K dainiai, Biržai and Taurag SFEs (Table 3.1). The only forest stand
in which fruiting bodies of P. tremulae have not been observed was located in Ignalina SFE,
although this might be due to a relatively young age of this stand (37-yrs-old).
Investigated plus trees of P. tremula grow in old and very old aspen stands (Table 3.1.). This old
stand age (today in Lithuania, timber crop rotation of P. tremula is 40 years) is one of the main
reasons for frequent occurrence of damages caused by an aspen trunk rot fungus P. tremulae.
The plus trees are rather old – only six trees haven’t reached the age of 40 years at the time of
our investigations (Table 2.1.). One plus tree had fruiting bodies of P. tremulae in the upper
third of its trunk, while other plus trees showed no external symptoms of the trunk rot disease
and thus were assigned for molecular detection of the infection in their wood samples. Several
wood samples were placed onto nutrient media for fungal isolation according to methodology by
Bakys et al. (2009). Pure cultures of P. tremulae have been isolated from some of the collected
wood samples (data not shown), although the frequency of isolation of this fungus was far lower
compared to the frequency of detection of its DNA in the wood samples. These results indicated
that in many cases false negative results showing incidence of infection may be obtained by the
isolation method (Chakravarty and Hiratsuka, 2007).
Tabl
e 3.
1. M
ean
char
acte
rist
ics
of in
vest
igat
ed P
opul
us tr
emul
a st
ands
. Ass
essm
ent m
ade
in M
ay–J
uly,
200
7. F
or e
xpla
natio
n of
diff
eren
t var
iabl
es
see
sect
ion
2.2.
Sta
nd n
ames
giv
en a
ccor
ding
to n
ames
of r
espe
ctiv
e St
ate
Fore
st E
nter
pris
e
Forest stand name
Stand age, years
Tree height, m
Stem diameter(dbh), cm
Stem straight- ness, score
Stem slender- ness, score
Stem height to dry branches, m
Stem height to green branches, m
Crown diameter, m
Branch thickness, score
Stem presence, score
Branch wound occlusion, score
Sanitary condition, score
No. (%) of trees with fruiting bodies of
Phellinus tremulae
Any
kšia
i 61
31
.3
39.9
4.
0 3.
9 12
.6
18.4
8.
7 3.
3 3.
0 3.
5 1.
1 4
(13.
3 %
)
Birž
ai
66
30.0
41
.0
3.7
3.6
10.8
17
.8
7.1
3.7
3.5
3.6
1.2
10 (3
3.3%
) Ig
nalin
a 37
21
.4
23.9
3.
7 3.
9 3.
5 9.
4 5.
7 3.
5 3.
6 4.
7 1.
2 0
(0%
) Ju
rbar
kas
51
29.6
34
.9
3.7
3.5
13.2
17
.1
5.5
3.8
3.4
4.2
1.2
1 (3
.3%
) K
aiši
ador
ys 1
a 59
27
.7
38.2
3.
9 3.
5 8.
9 14
.2
5.8
3.5
3.3
3.6
1.4
4 (1
3.3%
) K
aiši
ador
ys 2
a 55
26
.5
38.2
3.
9 3.
9 7.
1 15
.1
5.9
3.4
3.3
3.7
1.4
3 (1
0%)
Kda
inia
i 62
27
.5
38.5
3.
6 3.
6 9.
9 15
.0
8.3
3.1
3.0
2.7
1.5
16 (5
3.3%
) K
retin
ga
58
28.7
34
.7
3.9
4.0
8.5
14.9
6.
9 3.
9 3.
8 3.
7 1.
6 3
(10%
) K
urš
nai
67
29.7
40
.2
3.7
3.8
7.1
14.4
5.
9 2.
9 3.
0 3.
4 1.
4 2
(6.7
%)
Mar
ijam
pol
46
29
.5
31.6
4.
0 3.
8 8.
7 19
.2
3.6
3.5
2.9
3.3
1.8
6 (2
0%)
Pakr
uojis
1a
55
25.2
34
.1
3.6
3.3
9.9
15.1
7.
0 3.
6 3.
2 3.
5 1.
5 3
(10%
) Pa
kruo
jis 2
a 67
27
.7
35.2
3.
2 3.
1 13
.2
15.8
7.
2 3.
2 3.
0 3.
8 2.
2 8
(26.
7%)
Ras
eini
ai
47
28.9
31
.6
3.9
3.8
12.1
16
.7
6.1
3.8
3.7
4.6
1.2
2 (6
.7%
) R
okiš
kis 1
a 62
28
.5
37.5
3.
3 3.
6 6.
0 16
.3
5.9
3.1
3.0
2.9
1.8
21 (7
0%)
Rok
iški
s 2a
67
28.5
41
.7
2.8
3.2
8.7
14.7
7.
8 2.
8 3.
0 3.
3 1.
8 21
(70%
) Ša
kiai
65
27
.5
41.6
3.
7 3.
6 8.
8 15
.1
6.7
2.9
3.0
3.3
1.3
5 (1
6.7%
) Ša
lin
inka
i 57
27
.8
38.9
3.
3 3.
2 8.
6 14
.4
8.0
3.2
3.0
3.6
1.6
7 (2
3.3%
) Ta
urag
67
30
.9
42.5
4.
2 4.
3 11
.4
17.4
8.
3 4.
2 4.
0 3.
7 1.
2 13
(43.
3%)
Ute
na
62
28.7
29
.8
3.7
3.7
8.9
16.3
5.
8 3.
7 3.
2 3.
8 1.
2 1
(3.3
%)
Mea
n of
P.
trem
ula
plus
tr
eesb
60
29.5
40
.6
4.5
4.4
13.2
16
.5
7.8
4.0
4.0
4.5
1.0
1
a Stat
e Fo
rest
Ent
erpr
ises
, in
whi
ch tw
o fo
rest
stan
ds w
ere
inve
stig
ated
: Kai
šiad
orys
1, B
da fo
rest
ry d
istri
ct, b
lock
no.
287
; Kai
šiad
orys
2, P
ravi
eniš
ks f
ores
try
dist
rict,
bloc
k no
. 89,
Pak
ruoj
is 1
and
2, L
inku
va fo
rest
ry d
istri
ct, r
espe
ctiv
e bl
ock
nos.
39 a
nd 5
2; a
nd R
okiš
kis 1
and
2, K
amaj
ai fo
rest
ry d
istri
ct, r
espe
ctiv
e bl
ock
nos.
206
and
208.
b se
e A
ppen
dix
2.
50
Fig. 3.1. Fragments of amplified fungal DNA extracted from European aspen wood. M – molecular size marker, 1 to 20 – tested DNA samples, K – PCR control samples (left, sample without DNA, and right,
sample with plant DNA), G – amplification product of Phellinus tremulae DNA (positive control)
The presence of P. tremulae DNA in wood samples of P. tremula plus trees was assessed using
molecular methods – PCR-RFLP (Figure 3.1.). Gel electrophoresis allowed preliminary
identification of plus trees infected with this parasitic fungus. The amplified bands of similar
size to that produced by P. tremulae – positive sample (positive control, Figure 3.1.) were cut
out and re-amplified. The re-amplified PCR products were digested with restriction enzymes
allowing identification of samples containing P. tremulae DNA. The results of identification
based on the RFLP profiling are presented in Table 3.2. The presented method of detection of P.
tremulae DNA in aspen wood proved to be a valuable tool for identification of infected trees at
early stages of disease development. Such screening for infection could be useful in resolving
which trees should be selectively felled before trunk rot disease reaches its advanced stages
(Allen et al. 1996).
The results showed that 73 out of 137 (53.3%) European aspen plus trees were infected with P.
tremulae. The mean age of infected (60 years) and sound-looking (59 years) trees was almost
identical. The highest proportion of infected trees was found in Šakiai (80.0%), Anykš iai
(66.7%), Kaišiadorys (60.0%) and Marijampol (60.0%) stands (see Table 3.2). According to
the results of PCR-RFLP profiling, the healthiest plus trees were found in Kretinga and Rokiškis
(16.7% and 25.0% infected trees, respectively), followed by Ignalina and Kurš nai stands both
with 40% of infected trees.
Quality assessment data in the investigated aspen stands (Table 3.1.) showed that plus trees were
selected in forest stands of good and very good quality of phenotypic traits used in forest
51
selection and breeding: almost all stands could be given high quality indices which are close to
the indices given for the plus trees (see Table 3.1 and Appendix 2). The maximum difference in
quality indices was seen between a stand and a plus trees in K dainiai and Rokiškis 1 stands
(Table 3.1.).
Table 3.2. Presence (+) and absence (–) of Phellinus tremulae DNA in wood samples of Populus tremula plus trees as assessed by RFLP profiling
a for detailed description of DNA extraction methods see section 2.3.
Subsequently, we tested quality of the extracted DNA by PCR amplification (Figure 3.2).
Chloroplast tRNR L-F intergenic spacer amplification revealed that the most effective extraction
technique was protein precipitation, while CTAB precipitation, guanidinium isothiocyanate
method and AppliChem DNA isolation reagent resulted in no amplification at all thus
confirming DNA quantification results. DNA extracted from wood tissue by alkaline isolation
gave no amplification as well, even though some DNA was present in the template. It is possible
that there were some PCR inhibitory substances left in the DNA solution, or that extracted DNA
was degraded or oxidized. Other methods (CTAB and SDS) gave satisfactory results: DNA
concentration on average varied from 88 ng/ l for wood tissues to 245 ng/ l for leaf tissues. The
same success of DNA extraction as for protein precipitation technique was obtained using
commercial DNA extraction kits (Nucleospin Plant Mini (Macherey-Nagel, Düren, Germany);
Genomic DNA purification kit (Fermentas, Vilnius, Lithuania) and InnuPREP Plant DNA Kit
(Analytik Jena, Jena, Germany)).
54
Fig. 3.2. PCR products obtained after amplification using trnLUAA F - trnFGAA primers. M stands for molecular size marker (MassRuller DNA ladder mix, Fermentas), number 1 to 8 designates plant
samples that corresponds to DNA extracted using protein precipitation method from wood (1–2) and leaf (3–4) as well as alkaline isolation method from wood (5–6) and leaf samples (7–8); K1 is negative DNA
control (fungal DNA sample)
In Table 3.4 we compared six DNA extraction methods that proved to be suitable for DNA
extraction from aspen wood and leaf tissues. Here we describe the specific conditions applied in
the DNA extraction, present the number of cleaning steps, as well as the time required for
extraction. Some methods (protein precipitation and CTAB) require a long time for extraction,
but they involve several steps that are not laborious (e.g. cooling, long centrifugation,
precipitation) and in total amount of effort are comparable to commercial DNA extraction kits.
In our study, we used several DNA extraction techniques with different number of cleaning
steps. We started with the alkaline isolation technique, that hardly requires any cleaning step,
and finally we tested several commercial DNA extraction kits, containing several different
cleaning steps as well as enzymatic RNA and protein digestion. Our results suggested that for
large-scale investigations the best methods for DNA extraction are CTAB and protein
precipitation, although these require more time compared to the expensive commercially
available kits. Despite of being time-consumable these two methods yielded good quality DNA
suitable for PCR (see Table 3.4. and Figure 3.2.).
55
Table 3.4. Comparison of six DNA extraction methods/commercial kits that proved to be suitable for DNA extraction from wood and leaf tissues of Populus tremula
DNA extraction method Time required for DNA isolation
No. of cleaning steps DNA quality
Cost
SDS 25 min per sample one alcohol precipitation step
medium low
Protein precipitation ~ 4–4.5 h (possibility to handle several samples at a time)
three steps, protein and 2x alcohol precipitation
high low
CTAB ~ 2.5 h (possibility to handle several samples at a time)
two steps, chloroform and alcohol precipitation
high low
Nucleospin Plant Mini (Macherey-Nagel)
30 min per sample four steps, precipitation (binding to membrane), 3x wash
high high
Genomic DNA purification kit (Fermentas)
25 min per sample three steps, chloroform precipitation, ethanol precipitation
high medium
InnuPREP Plant DNA Kit (Analytik Jena)
40 min per sample five steps, precipitation (binding to the membrane), protein digestion and 3x wash
high high
The size of the amplification product is one of the quality measurements of the extracted DNA,
and is crucial for some molecular investigations (Deguilloux et al. 2003; Rachmayanti et al.
2009). The amplified P. tremula DNA fragment is 820 bp in size (Figure 3.2.), thus we can
conclude that CTAB or protein precipitation techniques are well suited for DNA extraction from
aspen wood and leaf tissues. The SDS method is relatively fast, yet it results in DNA of
comparably low quality and might not be recommended for the routine use. All of the three
commercially available kits produced good results in our study and proved to be effective in
molecular research including DNA extraction from wood tissue samples. The usefulness of
commercial DNA extraction kits was also confirmed by other authors (Deguilloux et al. 2003;
Rachmayanti et al. 2006; 2009). These kits are easy to handle, simple and fast, yet significantly
more expensive.
3.3. Genetic diversity of P. tremula plus trees
One of the objectives of the present study was to assess genetic diversity of European aspen plus
trees and to establish fingerprints that would allow distinguishing different tree genotypes. After
performing amplification of DNA extracted from all 137 European aspen plus trees with 15
random decamer primers (Table 2.5.) we obtained 292 DNA bands. The amplification patterns
obtained with primer Roth A05 are presented in Figure 3.3.
56
Fig. 3.3. Amplification products obtained using Roth A05 primer. Lanes 1 to 8 show Populus tremula plus tree samples, lines M show molecular size markers (on the left – 1 kb MassRuler™ DNA Ladder
Mix (Fermentas); on the right – DNA Ladder 100bp plus (AppliChem))
Three independent PCR amplifications were performed using separately extracted DNA from
wood and leaf tissues. Independent amplification was performed in order to avoid any
inaccuracies and to make sure that all amplified bands were of the tree origin. The results of the
RAPD amplification are shown in Table 3.5. A total of 282 out of 292 amplified bands (96.6%)
were polymorphic. On average, 19.5 bands were produced with each primer. As presented in
Table 3.5, the most informative primer was Roth A05 (27 bands), followed by primers Roth
A01 and Roth B12, both resulting in 25 bands. The least number of bands was produced using
primers Roth B17 and Roth B13 (12 and 14 bands, respectively). Fragment size varied from 250
to 3500 bp. The widest size range (390–3500 bp) was obtained with the primer Roth 17005
(Table 3.5.). Different number of RAPD fragments was obtained in RAPD profiles of the
examined European aspen plus trees. The amount of DNA fragments characteristic to one
genotype varied from 93 (tree No.155 from Šal ininkai) to 160 (tree No. 094 from Raseiniai).
On average, 132.32 loci per genotype were obtained.
57
Table 3.5. Number of amplified and polymorphic DNA bands produced by 15 RAPD primers, their size range and discrimination power in a study of genetic diversity in Lithuanian Populus tremula populations
Primer used Number of
amplified bands
Number of
polymorphic bands
Amplified band
size range
Discrimination
power
Roth A01 25 23 380–3000 bp 0.9890
Roth A03 23 23 550–3500 bp 0.9878
Roth A04 20 19 470–3500 bp 0.9912
Roth A05 27 27 380–3000 bp 0.9918
Roth A09 22 21 400–3000 bp 0.9875
Roth A19 20 20 400–2500 bp 0.9932
Roth B01 15 15 470–2000 bp 0.9871
Roth B03 18 16 460–3000 bp 0.9928
Roth B12 25 23 480–3000 bp 0.9923
Roth B13 14 14 520–2500 bp 0.9919
Roth B17 12 11 400–1900 bp 0.9827
Roth 17002 20 20 250–3000 bp 0.9921
Roth 17005 18 18 390–3500 bp 0.9910
Roth 17009 15 14 800–3000 bp 0.9811
Roth 37001 18 18 520–3000 bp 0.9920
Using RAPD in this study we established the fingerprints unique to each European aspen plus
tree. These fingerprints could be very useful in future genetic studies, or in breeding
experiments, and will allow using genetic material more efficiently. Distinguishing of individual
trees especially in young age is very difficult or even impossible task using morphological
criteria, thus the genetic fingerprints of individual trees can help to overcome this problem.
Using RAPD data, we assessed the genotype specific RAPD profiles for P. tremula plus trees.
Primer A19 distinguished all European aspen plus trees, while other primers – only in
combination with each other. This result is similar to Liu and Furnier’s (1993), where one
primer distinguished all 110 P. tremuloides individuals. Such primers are valuable in research
activities because only one PCR step is required to identify certain individual. However, primers
able to distinguish large number of individuals are rare and more often identification of
genotypes are achieved combining the results of several primers. The example of such
identification is the study of Sanchez et al. (2000), where identification of 89 P. tremula
genotypes was performed combining four different primers. In our study, 21 combinations of
two primers distinguished all assessed genotypes of European aspen trees (Table 3.6).
58
Table 3.6. Primer combinations valid for distinguishing all investigated European aspen genotypes
The discriminating power of primer or particular set of primers was defined by Kloosterman et
al. (1993) as PD=1- (Pi)2, where Pi represents the frequency of each genotype. In our RAPD
study, PD for RAPD primers ranged from 0.9827 (primer Roth B17) up to 0.9932 (primer Roth
A19). Different banding patterns obtained for all European aspen plus trees also confirmed their
sexual origin. Sanchez et al. (2000) demonstrated that 36% of P. tremula trees, originating from
the same stand had the same RAPD amplification profile, however, such results were obtained
probably due to extensive planting of aspen in Spain decades ago. Meanwhile, in Lithuania P.
tremula has never been artificially propagated, which evidently resulted in more diverse aspen
stands due to sexual reproduction.
3.3.1. Nei’s genetic distance between P. tremula plus trees
Based on RAPD data (binary data matrix, showing presence or absence of RAPD fragment in
amplification profile), genetic distances between investigated European aspen trees were
calculated according to Nei’s method (1978). The maximum genetic distance (0.141) was
assessed between tree No. 096 (Raseiniai) and tree No. 150 (Utena). Genetically most similar
trees No. 104 and 105 (genetic distance 0.012) are both located in Marijampol SFE. The
average genetic distance between European aspen plus trees in Lithuania was 0.0964. RAPD
based dendrogram of genetic distance among European aspen trees was constructed using
UPGMA grouping method (Figure 3.4.). High RAPD polymorphism rate (96.6%) of
investigated European aspen confirms significance of the plus trees, specifically utilized in
forest tree breeding. High genetic diversity ensures genotypic plasticity of trees adapting to
changing environment. According to Leibenguth and Shogi (1998), high polymorphism creates
conditions to reproductive and adaptive heterosis. Constructed 137 European aspen plus trees
RAPD profile library could be used for identification of these plus trees in successive selection
and breeding programs.
Fig.
3.4
. Eur
opea
n as
pen
tree
s clu
ster
ed u
sing
UPG
MA
met
hod,
by
algo
rith
m P
aire
d gr
oup,
sim
ilari
ty m
easu
re E
uclid
ean
dist
ance
(Boo
t N=
1,00
0),
base
d on
RAP
D d
ata
(Cop
hene
tic c
orre
latio
n 0.
4729
)
60
3.3.2. Correlation between genetic distances and geographic distribution pattern of P.tremula plus trees
96.6% of all obtained RAPD amplification products were polymorphic, revealing considerable
genetic diversity among P. tremula plus trees. Nevertheless, the correlation between genetic and
geographic distances obtained using Mantel test was weak (r=0.178, correlation significance
p=0.001). The results of Mantel test are showed in Figure 3.5. Weak genetic differentiation
among P. tremula trees from different geographic origins probably occurs due to biological
characteristics of this species. P. tremula is wind pollinated and dispersed, and that ensures high
gene migration rate (Lexer et al. 2005; Hall et al. 2007). The other possible reason for weak
genetic differentiation could be explained by only a small amount of P. tremula genome used
for this study.
Fig.3.5. Correlation between RAPD genetic distances and geographic location data of European aspen plus trees (Mantel test, 9,999 permutations)
3.4. Population structure of P. tremula in Lithuania
The aim of this study was to infer the number of European aspen populations in Lithuania and to
assess their genetic differences based on DNA marker data. A total of 314 P. tremula trees
(including 137 plus trees) from 16 SFE’s in Lithuania were included in this study.
In this study we used five microsatellite primers that amplifies genetic loci in 5th, 7th, 8th, 15th
and 16th chromosomes (Figure 2.1). In the analysis of the obtained SSR data we also considered
the RAPD results. In this particular case, RAPD analysis data should be treated with caution
because of the variable numbers of trees per population included in our analysis (2 to 30
individuals, see Table 2.1).
61
3.4.1. Genetic relatedness of P. tremula populations by PCA
Genetic relatedness of European aspen populations based on 5 microsatellite loci was analyzed
by principal component analysis. Figures 3.6 and 3.7 show PCA of European aspen genetic
variation within and among populations.
Fig. 3.6. A – PCA (principal component analysis) scatter diagram with arrows indicating SSR marker load based on 5 SSR marker correlation matrixes (within group/population option). B – PCA screen plot
with Broken Stick. C, D – PCA loadings indicating SSR marker load coefficients on each of two most important (as indicated by Broken Stick) components explaining accordingly 24.0% and 17.9% of total
variance
62
Fig. 3.7. A – PCA (principal component analysis) scatter diagram with arrows indicating SSR marker load based on 5 SSR marker correlation matrixes (among group/population options). B – PCA screen
plot with Broken Stick. C, D, E – PCA loadings indicating SSR marker load coefficients on each of first three most important (as indicated by Broken Stick) components explaining accordingly 28.7%, 27.0%
and 17.9% of total variance
63
As demonstrated in Figures 3.6 and 3.7, the first two components are important in within group
analysis and three components are important in between group analysis. GCPM 1608 and
GCPM 1532 SSR markers have the largest load in PCA and are the most important in within
group analysis, while WPMS 14, GCPM 1608 and PMGC 2607 markers are most
discriminating among populations.
3.4.2. Genetic parameters of assessed SSR loci of evaluated P. tremula populations
All five microsatellite loci used in this study were polymorphic and revealed 7 to 14 alleles
(Table 3.7). The locus GCPM 1532 was least polymorphic, while locus WPMS 14 had the
highest number of alleles (Table 3.7). The effective number of alleles was considerably lower
than the observed number of alleles for all the loci investigated. Observed heterozygosity was
higher than expected heterozygosity in all loci except GCPM 1532, where expected
heterozygosity exceeded observed. The Fixation Index (also called the Inbreeding coefficient)
was found to be close to zero at WPMS 16, GCPM 1532 and WPMS 14 loci, thus indicating a
random mating. PMGC 2607 and GCPM 1608 loci had Inbreeding coefficient of -0.41 and -0.59
(Table 3.7), which suggests negative assortative mating or selection for heterozygotes. All loci
used in this study except PMGC 2607 show significant population differentiation based on FST
index (Table 3.7). This indicates that all microsatellite loci selected for this study are suitable for
population genetic structure studies. Based on RST fixation index, populations were significantly
differentiated at all the loci as well, except for PMGC 2607 locus (Table 3.7).
Table 3.7. The mean statistics of microsatellite loci with ± standard error. The differentiation showed in this table is among populations. Na – number of different alleles; Ne – effective number of alleles; I – Shannon's information index; Ho – observed heterozygosity; He – expected heterozygosity; uHe – unbiased expected heterozygosity; F – inbreeding coefficient; Rst – fixation index; Fst – fixation index
Index WPMS 16 PMGC 2607 GCPM 1532 GCPM 1608 WPMS 14 Na 10 9 7 13 14
Ne 4.171±0.333 2.238±0.090 2.809±0.181 2.240±0.071 2.947±0.223
I 1.547±0.080 0.890±0.046 1.175±0.056 0.948±0.041 1.315±0.086 Ho 0.760±0.054 0.759±0.032 0.577±0.050 0.869±0.036 0.747±0.051
He 0.731±0.026 0.544±0.016 0.620±0.026 0.546±0.015 0.627±0.031
for Anykš iai population indicates clear negative assortative mating, or selection against
homozygotes. Mean Inbreeding coefficient value (-0.23) shows, that P. tremula in Lithuania
could be experiencing selection pressure against homozygotes (Table 3.9). This could be
expected from the biology of the species’, as it reproduces not only via seed, but also by root
suckers, and to retain high diversity it needs to favor heterozygotes. High genetic diversity rate
is also crucial for P. tremula as for pioneer species. Petit and Hampe (2006) noticed that
heterozygote advantage is an expected genomic feature of long lived and widely distributed in
different environments forest tree species. As P. tremula is distinguished by longevity (further
enhanced by clonal reproduction) (Wühlisch, 2009) and vast amount of seed production every
year (Reim, 1929), there exists high potential to “filter” fit gene combinations via selection
(Lindtke at al. 2012).
Tabl
e 3.
9. T
he in
tra-
popu
latio
n ge
netic
div
ersi
ty in
dice
s with
stan
dard
err
ors o
ver t
he lo
ci, o
f Eur
opea
n as
pen,
ass
esse
d us
ing
SSR
data
. N –
no.
of
sam
pled
tree
s; N
a –
num
ber o
f diff
eren
t alle
les;
Ne
– ef
fect
ive
num
ber o
f alle
les;
I –
Shan
non ’
s inf
orm
atio
n in
dex;
Ho
– ob
serv
ed h
eter
ozyg
osity
;
He
– ex
pect
ed h
eter
ozyg
osity
; uH
e –
unbi
ased
exp
ecte
d he
tero
zygo
sity
; F –
fixa
tion
inde
x
Popu
latio
n N
N
a N
e I
Ho
He
uHe
F A
nykš
iai
17
2.60
0±0.
400
2.20
1±0.
156
0.81
6±0.
086
0.96
5±0.
024
0.53
8±0.
028
0.55
4±0.
029
-0.8
09±0
.086
B
iržai
18
6.
000±
1.04
9 3.
438±
0.44
2 1.
393±
0.15
0 0.
861±
0.04
5 0.
691±
0.03
5 0.
712±
0.03
6 -0
.266
±0.1
18
Igna
lina
19
5.60
0±0.
812
3.38
8±0.
586
1.31
9±0.
175
0.77
3±0.
080
0.67
1±0.
050
0.69
0±0.
051
-0.1
88±0
.179
Ju
rbar
kas
18
4.60
0±0.
600
2.72
0±0.
302
1.13
8±0.
119
0.83
1±0.
039
0.61
6±0.
036
0.63
4±0.
037
-0.3
57±0
.063
K
aiši
ador
ys
21
5.60
0±0.
927
2.79
2±0.
333
1.19
8±0.
162
0.69
3±0.
059
0.61
8±0.
050
0.63
3±0.
052
-0.1
53±0
.134
K
dain
iai
22
5.20
0±0.
917
3.08
7±0.
541
1.22
2±0.
183
0.85
9±0.
042
0.64
0±0.
054
0.65
5±0.
055
-0.3
73±0
.111
K
retin
ga
22
5.60
0±0.
678
3.55
1±0.
804
1.32
5±0.
181
0.68
0±0.
086
0.67
2±0.
053
0.68
8±0.
054
-0.0
58±0
.185
K
urš
nai
21
5.00
0±0.
447
2.94
1±0.
450
1.22
5±0.
132
0.72
4±0.
046
0.63
2±0.
047
0.64
7±0.
048
-0.1
65±0
.103
M
arija
mpo
l
17
5.60
0±0.
927
3.67
6±0.
621
1.36
4±0.
208
0.77
3±0.
095
0.68
7±0.
061
0.70
8±0.
063
-0.1
60±0
.193
Pa
kruo
jis
30
6.40
0±1.
364
2.59
0±0.
393
1.15
0±0.
154
0.71
9±0.
127
0.58
8±0.
043
0.59
8±0.
044
-0.2
23±0
.209
R
asei
niai
18
3.
600±
0.24
5 2.
224±
0.25
7 0.
927±
0.08
3 0.
538±
0.14
1 0.
529±
0.04
7 0.
550±
0.05
0 0.
002±
0.24
7 R
okiš
kis
16
3.80
0±0.
663
2.24
3±0.
536
0.90
0±0.
175
0.55
3±0.
069
0.48
4±0.
074
0.50
1±0.
076
-0.1
70±0
.104
Ša
kiai
20
6.
000±
1.04
9 3.
588±
0.68
8 1.
383±
0.18
5 0.
829±
0.06
6 0.
685±
0.04
9 0.
703±
0.05
0 -0
.235
±0.1
43
Šal
inin
kai
20
6.00
0±0.
707
2.32
8±0.
176
1.13
4±0.
077
0.56
0±0.
097
0.56
0±0.
037
0.57
4±0.
038
-0.0
07±0
.159
Ta
urag
17
5.
600±
1.03
0 2.
918±
0.28
0 1.
261±
0.15
8 0.
776±
0.06
8 0.
640±
0.04
6 0.
659±
0.04
8 -0
.223
±0.0
88
Ute
na
18
4.60
0±0.
510
2.40
8±0.
245
1.04
5±0.
100
0.74
4±0.
106
0.56
8±0.
042
0.58
4±0.
043
-0.2
91±0
.150
M
ean
19.1
63
5.11
3±0.
216
2.88
1±0.
119
1.17
5±0.
039
0.74
2±0.
022
0.61
4±0.
013
0.63
1±0.
013
-0.2
30±0
.040
67
Fig. 3.8. Observed heterozygosity in investigated European aspen populations assessed using SSR
In addition to SSR the mean genetic diversity indices for all investigated European aspen
populations were also calculated using RAPD data (Table 3.10). The effective number of alleles
exceeded the number of different alleles for populations with low number of assessed individuals (5
or less) (Table 3.10). Shannon’s information index varied from 0.192 in Kurš nai up to 0.401 in
K dainiai closely followed by Pakruojis (0.402) population. The lowest expected heterozygosity
values were calculated for Kurš nai (0.127), Taurag (0.136), Marijampol (0.188) and Anykš iai
(0.196) populations. The number of assessed trees in Kurš nai, Marijampol and Taurag
populations was low (5, 5 and 2 respectively) (Table 3.10), whereas the expected heterozygosity in
Anykš iai population was also among the lowest assessed using SSR data (Table 3.9).
68
Table 3.10. The within population genetic diversity indices with standard error over the loci of European aspen, assessed using RAPD data. N – number of sampled trees; Na – number of different alleles; Ne – effective number of alleles; I – Shannon’s information index; He – expected heterozygosity; uHe – unbiased expected heterozygosity
3.4.4. Nei’s genetic distances between populations of P. tremula
In order to assess genetic relatedness of P. tremula populations in Lithuania we calculated Nei’s
genetic distances between them, according to SSR data. Pairwise genetic distances for P. tremula
populations are presented in Table 3.12. Assessed mean genetic distances among European aspen
populations in Lithuania is 0.125. The most similar populations are from Jurbarkas and Kurš nai
(genetic distance was 0.035). Slightly greater genetic distance was assessed between Pakruojis –
Šal ininkai (0.036) and Šal ininkai – Taurag (0.038) populations (Table 3.12), while the most
distinct populations are from Anykš iai and Rokiškis (0.388). Anykš iai – Utena (0.313),
Marijampol – Rokiškis (0.293) and Anykš iai – Biržai (0.258) populations also are among the
most different P. tremula populations (Table 3.12).
Somewhat similar results were obtained and using RAPD data for population genetic distance
assessment. European aspen pairwise Nei’s genetic distance data are showed in Table 3.12. Mean
genetic distance between P. tremula populations in Lithuania according to RAPD is 0.131. The two
most similar populations are from Pakruojis and Šal ininkai (0.031), confirming Nei’s genetic
distance calculated between these two populations using SSR data. The other genetically similar
populations according to RAPD data are Pakruojis – Utena and Pakruojis – Kaišiadorys (0.037).
Genetic distance between these populations according to SSR data is close to the mean genetic
distance and is equal to 0.103 and 0.123 respectively (Table 3.12). The most different populations
according to RAPD are Kurš nai – Taurag (0.292), Marijampol – Taurag (0.252) and Rokiškis –
Taurag (0.245). Compared to SSR data, these populations are quite similar, Nei’s genetic distances
among them are 0.068, 0.098 and 0.162 respectively. However, only two samples from Taurag
population were used in RAPD analysis, possibly influencing the results (Table 2.1).
Tabl
e 3.
12. P
airw
ise
Nei
’s g
enet
ic d
ista
nces
bet
wee
n in
vest
igat
ed P
opul
us tr
emul
a po
pula
tions
. The
upp
er n
umbe
r in
dica
tes
gene
tic d
ista
nce
calc
ulat
ed u
sing
SS
R da
ta (b
old)
and
the
low
er n
umbe
r is c
alcu
late
d us
ing
RAPD
dat
a (I
talic
)
Bir
žai
0.25
8 0.
153
Igna
lina
0.22
0 0.
117
0.05
00.
136
Jurb
arka
s 0.
209
0.11
1 0.
054
0.13
4 0.
040
0.10
9
Kai
šiad
orys
0.
196
0.14
7 0.
093
0.09
2 0.
068
0.10
2 0.
118
0.11
2
Kda
inia
i 0.
238
0.10
1 0.
062
0.11
4 0.
066
0.07
2 0.
118
0.07
1 0.
057
0.08
4
Kre
tinga
0.
212
0.10
8 0.
085
0.10
8 0.
068
0.10
1 0.
133
0.08
4 0.
068
0.09
2 0.
076
0.05
9
Kur
šna
i 0.
229
0.23
5 0.
065
0.19
9 0.
054
0.20
7 0.
035
0.22
4 0.
131
0.16
2 0.
129
0.20
6 0.
106
0.19
4
Mar
ijam
pol
0.
213
0.19
4 0.
078
0.15
4 0.
051
0.13
9 0.
106
0.15
4 0.
089
0.08
8 0.
053
0.11
4 0.
062
0.13
5 0.
114
0.18
8
Pakr
uojis
0.
253
0.14
9 0.
079
0.09
1 0.
077
0.10
5 0.
066
0.11
2 0.
123
0.03
7 0.
126
0.08
1 0.
122
0.08
8 0.
061
0.18
5 0.
163
0.10
0
Ras
eini
ai
0.17
4 0.
133
0.12
20.
147
0.12
20.
111
0.07
90.
096
0.13
40.
143
0.15
70.
077
0.15
60.
108
0.12
0 0.
238
0.16
60.
175
0.09
00.
135
Rok
iški
s 0.
388
0.18
7 0.
191
0.11
5 0.
199
0.12
1 0.
164
0.15
4 0.
241
0.05
8 0.
249
0.12
1 0.
188
0.10
6 0.
117
0.18
7 0.
293
0.11
3 0.
111
0.06
3 0.
238
0.18
2
Šaki
ai
0.25
7 0.
143
0.05
30.
099
0.05
40.
112
0.07
90.
101
0.10
20.
040
0.11
00.
094
0.05
80.
097
0.07
9 0.
151
0.11
10.
102
0.07
40.
044
0.14
40.
137
0.18
10.
080
Šal
inin
kai
0.24
4 0.
168
0.08
20.
081
0.09
50.
117
0.05
10.
131
0.12
40.
053
0.13
70.
091
0.15
60.
096
0.07
0 0.
188
0.17
50.
102
0.03
60.
031
0.06
30.
151
0.12
70.
076
0.10
70.
067
Tau
rag
0.
226
0.20
6 0.
054
0.18
6 0.
064
0.19
7 0.
054
0.18
1 0.
056
0.21
7 0.
070
0.16
1 0.
098
0.15
8 0.
068
0.29
2 0.
098
0.25
2 0.
050
0.20
2 0.
086
0.18
1 0.
162
0.24
5 0.
092
0.21
0 0.
038
0.21
1
Ute
na
0.31
3 0.
168
0.13
10.
091
0.10
40.
121
0.06
50.
136
0.17
50.
059
0.22
90.
103
0.21
00.
116
0.08
8 0.
179
0.23
80.
131
0.10
30.
037
0.15
70.
159
0.13
50.
093
0.13
30.
062
0.09
80.
053
0.11
0 0.
223
Popu
latio
n
Anykšiai
Biržai
Ignalina
Jurbarkas
Kaišiadorys
Kdainiai
Kretinga
Kuršnai
Marijampol
Pakruojis
Raseiniai
Rokiškis
Šakiai
Šalininkai
Taurag
Tabl
e 3.
13. P
airw
ise
F ST v
alue
s (W
righ
t, 19
46; 1
951;
196
5) fo
r all
inve
stig
ated
Pop
ulus
trem
ula
popu
latio
ns c
alcu
late
d us
ing
SSR
data
. Sig
nific
antly
diff
eren
t es
timat
es b
y pr
obab
ility
P (r
and
dat
a) b
ased
on
999
perm
utat
ions
are
show
n in
bol
d
Bir
žai
0.07
1
Ig
nalin
a 0.
064
0.01
1
Jurb
arka
s 0.
066
0.01
5 0.
011
Kai
šiad
orys
0.
062
0.02
4 0.
018
0.03
4
Kda
inia
i 0.
071
0.01
5 0.
015
0.03
1 0.
016
Kre
tinga
0.
062
0.01
9 0.
015
0.03
4 0.
018
0.01
9
Kur
šna
i 0.
069
0.01
7 0.
014
0.01
1 0.
036
0.03
4 0.
026
Mar
ijam
pol
0.
061
0.01
6 0.
011
0.02
7 0.
023
0.01
2 0.
014
0.02
7
Pakr
uojis
0.
081
0.02
3 0.
021
0.02
1 0.
038
0.03
6 0.
034
0.02
0 0.
042
Ras
eini
ai
0.07
4 0.
048
0.04
6 0.
036
0.04
9 0.
062
0.05
5 0.
047
0.06
0 0.
040
R
okiš
kis
0.13
6 0.
066
0.06
7 0.
062
0.08
5 0.
085
0.06
6 0.
048
0.08
9 0.
046
0.08
8
Ša
kiai
0.
071
0.01
2 0.
012
0.02
1 0.
026
0.02
6 0.
013
0.02
0 0.
023
0.02
2 0.
055
0.06
5
Šal
inin
kai
0.08
2 0.
026
0.02
7 0.
017
0.04
0 0.
042
0.04
5 0.
024
0.04
7 0.
012
0.03
2 0.
054
0.03
3
T
aura
g
0.06
8 0.
014
0.01
5 0.
015
0.01
6 0.
019
0.02
4 0.
018
0.02
3 0.
016
0.03
7 0.
060
0.02
2 0.
014
Ute
na
0.09
8 0.
039
0.03
2 0.
022
0.05
3 0.
064
0.05
7 0.
028
0.06
1 0.
035
0.05
7 0.
057
0.03
9 0.
035
0.03
4
Popu
latio
n
Anykšiai
Biržai
Ignalina
Jurbarkas
Kaišiadorys
Kdainiai
Kretinga
Kuršnai
Marijampol
Pakruojis
Raseiniai
Rokiškis
Šakiai
Šalininkai
Taurag
Fig.
3.1
0. C
lust
erin
g of
Pop
ulus
trem
ula
popu
latio
ns a
ccor
ding
to N
eis‘
s gen
etic
dis
tanc
es. C
lust
erin
g ca
lcul
ated
usi
ng U
PGM
A m
etho
d an
d Pa
ired
gro
up
algo
rith
m a
nd E
uclid
ean
dist
ance
sim
ilari
ty m
easu
re (B
oot N
=10
00).
A –
dend
rogr
am g
ener
ated
usi
ng S
SR d
ata
(Cop
hene
tic C
orre
latio
n 0.
9198
); B
—
dend
rogr
am g
ener
ated
usi
ng R
ADP
data
(Cop
hene
tic C
orre
latio
n 0.
9379
)
74
Pairwise FST values (Wright, 1946; 1951; 1965) for aspen populations (SSR data) are shown in
Table 3.13. Regardless of the significance of the many estimates (larger than 0.017 in Table
3.13), genetic differentiation of the assessed populations we consider as small or moderate
(exceeding 0.05). The average FST value between Rokiškis and other populations is 0.072, while
average FST value for Anykš iai population is even greater and reached 0.076, indicating the
reduced gene flow to Anykš iai and Rokiškis populations. In similar studies, comparable FST
results were obtained after studying five isozyme loci in P. tremula populations from France,
Austria, Southern and Northern Sweden. Ingvarsson (2005) found that FST values differed from
0.040 to 0.161, averaging to 0.117. Limited gene flow among populations (FST=0.11) has been
observed in the study encompassing P. tremula populations from numerous European countries
using chloroplast SSR markers (Petit et al. 2003b). In our study average FST value was 0.037
indicating more pronounced gene flow, yet limited geographic range among the analyzed
Lithuanian populations should be considered. Lithuania is a relatively small country with no
geographic barriers interrupting gene flow among European aspen populations.
To reveal the similarities between assessed P. tremula populations according to calculated Nei’s
genetic distances, genetic dendrograms were constructed (Figure 3.10). In the dendrogram based
on SSR data, Lithuanian aspen populations are grouped into three separate clusters. P. tremula
population from Anykš iai is clustered separately. The first dendrogram cluster, composed of 8
populations contains both pairs of populations most similar according to Nei’s genetic distance
(Jurbarkas – Kurš nai and Pakruojis – Šal ininkai, see Table 3.12). These two pairs form a
separate sub-cluster. The second cluster is composed of 4 populations from Kaišiadorys,
K dainiai, Kretinga and Marijampol . The third cluster contains three populations from
Raseiniai, Rokiškis and Utena. These three populations can be characterized by high to
moderate genetic distance, e.g. they differ more from the other investigated populations.
Reliability estimates are showed at the branching points of the dendrogram (Figure 3.10). Three
dendrogram clusters are formed with high reliability measure, as the lowest reliability value is
80% and was obtained separating the first and the second clusters (Figure 3.10).
Genetic dendrogram, constructed using genetic distances assessed by RAPD data, shows
grouping of P. tremula populations into two main clusters (Figure 3.10). Anykš iai, Ignalina,
Jurbarkas, K dainiai, Kretinga and Raseiniai populations form separate cluster 1, while Biržai,
Kaišiadorys, Marijampol , Pakruojis, Rokiškis, Šakiai, Šal ininkai and Utena form the second
clade (Figure 3.10).The distinct clade is formed by Kurš nai and Taurag populations. These
75
two populations can be characterized by low number of assessed individuals (2 for Taurag and
5 for Kurš nai, see Table 2.1 in Materials and methods).
3.4.5. Correlation between genetic distances and geographic distribution pattern of P.tremula populations
Obtained microsatellite data were used in Mantel test to examine correlation between genetic
distances and geographic distances of assessed P. tremula populations. The result of Mantel test
is 0.068 with significance level 0.01, when based on 9,999 permutations (Figure 3.11). Results
of Mantel test for SSR data indicate an even weaker correlation than for RAPD data (Figure 3.5)
between genetic and geographic distances of investigated European aspen populations. The
correlations from Mantel test were much stronger for separate loci (data not shown).
Fig. 3.11. Correlation between genetic distances and geographic locations of Populus tremula populations based on SSR data (Mantel test, 9,999 permutations)
3.4.6. P. tremula within- vs. among-population genetic diversity
Microsatellite data were also used to assess genetic diversity within individuals vs among
populations of P. tremula. The multilocus AMOVA (Excoffier et al. 1992; Mengoni and
Bazzicalupo, 2002) using microsatellite distance matrix revealed higher among population
variance for the RST index than for the FST index (Figure 3.12). As presented in Figure 3.12, 96%
and 95% (FST and RST respectively) of variation is explained by variation within aspen
individuals, while 4% and 5% (FST and RST respectively) – because of the variation among
different populations. According to Lexer et al. (2005) and Hall et al. (2007), variation among
European aspen populations based on neutral markers is approximately 1%. The rest of the
variation is between individuals. This might be because of the aspen biology (as a wind
76
pollinated and dispersed tree), that ensures high gene flow and diminishes spatial genetic
structure among populations.
Fig. 3.12. The partition of the variance based on the multilocus AMOVA carried out for FST (left) and RST (right) estimates with GenAlEx ver. 6.5 (Peakall and Smouse, 2006; 2012)
Fig. 3.13. Distribution of Populus tremula populations in principal coordinate axes using pairwise population matrix of Nei’s unbiased genetic distance based on SSR data
PCA of European aspen population distribution based on SSR data is showed in Figure 3.13.
The first axis explains 39.7%, the second axis explains 24.7% and the third axis explains 13.3%
of the assessed genetic variation. In total, the first three axes expalains 77.7% of genetic
variation. PCA indicates that Anykš iai and Rokiškis populations are more genetically distinct
77
from the other assessed P. tremula populations. This result confirms higher FST values
calculated for these two populations.
Fig. 3.14. Distribution of Populus tremula populations in principal coordinate axes using pairwise population matrix of Nei’s unbiased genetic distance based on RAPD data
PCA of distribution of European aspen populations based on RAPD data is showed in Figure
3.14. The first axis explains 26.9%, the second axis explains 15.8% and the third axis explains
12.4% of the assessed genetic variation. In total, the first three axes explains 55.1% of variation.
All investigated European aspen populations form single cluster, except for Kurš nai and
Taurag populations. However, these two populations are represented by a low number of
investigated individuals and might be excluded from the interpretation of the results.
3.4.7. Clustering of P. tremula populations using Bayesian approach
In order to cluster P. tremula populations based on their genetic differences, we used Bayesian
clustering approach. The projected number of populations was calculated by logarithmic
possibilities for the number of clusters in which the analyzed individuals group together. Using
calculated logarithmic possibilities we employed Delta K method ( K) to infer the true number
of genetic clusters. K calculations were based on the second order rate of change of the
likelihood (Evanno et al. 2005).
78
Fig. 3.15. K calculations of the second order rate of change of likelihood, based on SSR (A) and RAPD data (B)
K method based on SSR data suggests, that P. tremula individuals should be grouped in three
clusters, while RAPD data analysis revealed two clusters (Figure 3.15). Based on K
calculations, we attributed P. tremula individuals to each of the three clusters grouped according
to SSR data (Figure 3.16) and two clusters in relation to RAPD data (Figure 3.17). In Figure
3.18 similar calculations with 3 most discriminating populations SSR markers (as indicated by
PCA among population option, see Figure 3.7) are presented. Results of Bayesian clustering
approach confirms clustering results of P. tremula populations according to Nei’s genetic
distance analysis, where three main clusters of assessed populations were obtained for SSR, and
two – for RAPD data (Figure 3.10).
79
Fig. 3.16. Proportion of Populus tremula trees attributed to 3 clusters (as estimated by K from simulation summary of STRUCTURE program output; 100,000 runs) by 5 SSR markers
Fig. 3.17. Proportion of Populus tremula trees attributed to 2 clusters (as estimated by K from simulation summary of STRUCTURE program output; 50,000 runs) by RAPD markers
80
Fig. 3.18. Proportion of Populus tremula trees attributed to 3 clusters (as estimated by K from simulation summary of STRUCTURE program output; 100,000 runs) by 3 SSR markers most
discriminating P. tremula populations (as indicated in PCA analysis of marker load in 3 components, section 3.4.1)
The results from both clustering methods (UPGMA, based on Nei’s genetic distances (3.4.4.
section) and Bayesian clustering approach) indicate that existing provenance regions could be
revised, as both molecular markers revealed latitudinal trend. Site ecological as well as forest
inventory data may also be used in addition to justify the borders of provenance regions and
well-founded use of forest reproductive material in practical forest selection and breeding.
Referring to the relatively small area of Lithuania for the photoperiodic and temperature
gradients, where no geographic boundaries are present and most of the altitudinal variation is
within 150 meters limit, we do not expect any significant effects of the present-day natural
selection on P. tremula. European aspen field trials until now are absent in Lithuania because of
commercial considerations, revealing additional opportunities in the future in order to improve
provenance transfer and promote breeding of P. tremula.
81
3.5. Correlation between presence of P. tremulae DNA and certain RAPD fragments
Correlation between presence of P. tremulae mycelium in wood of European aspen (Table 3.2)
and RAPD fragments was determined using SAS procedure GENMOD. Only 5 RAPD
fragments were important in differentiation of trees by infection of P. tremulae (A05-850, B13-
1100, B17-850, 17002-1750 and 17009-1600). Only fragments A05-850 (p=0.0278) and B13-
1100 (p=0.0222) were significant in group differentiation according to SAS procedure
GENMOD. The cumulative frequency of A05-850 and B13-1100 fragments in investigated P.
tremula populations is showed in Figure 3.19. Loci B17-850, 17009-1600 and 17002-1750 were
close to 5% of statistical significance.
Fig. 3.19. Distribution of investigated Populus tremula populations in Lithuania and cumulative frequency per category (%) in RAPD loci B13-1100 (on the left) and A05-850 fragments
PCA based on detrended correspondence analysis using 5 important RAPD loci was performed
and revealed the distribution of the infected and non-infected trees in two axes (Figure 3.20).
Only B13-1100 locus was important in individual differentiation by presence of P. tremulae
DNA in aspen wood (explaining 7% of variation). The convex hulls of two sample groups
82
(infected and uninfected by P. tremulae) are overlapping, but the frequencies in two loci
indicate quite good possibilities for using these RAPDs loci in tree breeding at early ontogenesis
stages.
Fig. 3.20. Distribution in two axes of European aspen uninfected (indicated as dashed line and diamond shape) and infected by Phellinus tremulae (indicated as solid line and cross shape) trees based on the
results of detrended correspondence analysis using RAPD loci B13-1100, A5-850, B17-850, 17002-1750 and 17009-1600
Additionally, we calculated the correlation whether tree individual heterozygosity, RAPD
marker polymorphism, and provenance region, population or forest soil type had an effect on
presence of P. tremulae in European aspen. Pearson correlation between tree individual
heterozygosity by 5 SSR loci and presence or absence of P. tremulae DNA in wood samples
was estimated close to zero (-0.04). The same correlation between tree individual RAPD marker
polymorphism and presence of fungal DNA also resulted in similar estimate (-0.03). Percentage
of P. tremulae infected trees was slightly lower on temporarily moisturized or overmoistured
and eutrophic sites.
Provenance region and population influence on the presence of P. tremulae in wood of assessed
P. tremula was tested using SAS procedure MIXED (REML option), but the results showed that
these two factors have no effect on distribution of this fungal pathogen.
RAPD markers are known to amplify anonymous regions of genome, thus allowing to catch
some of the genetic variation underlying the resistance mechanisms only by chance. To increase
our chances in capturing correlation, we have used 15 highly informative primers (Table 3.5).
However, statistically significant correlation between the presence of P. tremulae and RAPD
loci was established for only two of them (B13-1100 p=0.0222 and A05-850, p=0.0278).
Obtained results suggest that search for correlation without prior screening of the resistance and
83
recognition of genetic mechanisms underlying the resistance of P. tremula is unreliable. The
chances of inferring genetic markers associated with European aspen resistance to P. tremulae
using RAPD are negligible.
The subsequent step of this study could be the identification of genome regions, represented by
B13-1100 and A05-850 fragments. The identification of these regions could lead to marker
assisted breeding, or QTL involved in identification of the resistance mechanisms.
We can conclude, that the infection of European aspen plus trees with P. tremulae doesn’t
depend on population, provenance region or individual tree heterozygosity. RAPD loci
correlated to the infection of European aspen by P. tremulae should be investigated further to
confirm or reject their value in assessing the resistance of aspen to this fungal pathogen.
3.6. Hybridization between European and hybrid aspen
The aim of this study was to assess the possible hybridization between European and hybrid
aspen by leaf morphology of P. tremula progenies. Tree flowering phenology was assessed to
reveal if European and hybrid aspen phenologies overlap thus allowing the taxa to hybridize.
The assessment of European and hybrid aspen flowering phenology revealed that hybrid aspens
are characterized by earlier phenology compared to European aspen trees. The percentage of late
flowering category trees in European aspen group was 48% and in hybrid aspen group 9%; the
percentages in medium flowering category was 35% vs 25% in European and hybrid aspens
respectively, while early flowering group was comprised of 17% of European and 66% of
hybrid aspen trees (Figure 3.21). European and hybrid aspen trees attributed to the same
flowering category were flowering at the same time.
Fig. 3.21. The proportion of European and hybrid aspen trees in each of early, medium and late flowering phenology categories
84
Two P. tremula and one hybrid aspen tree per flowering category were sampled (n=9), and leaf
morphological parameters of their half-sib progenies (n=900) were analyzed using SAS
STEPDISC procedure; ten leaves per each of the progeny tree (e.g. 9000 leaves). As the result
of SAS STEPDISC, we selected 7 leaf morphological trait parameters (out of 19) best
discriminating between P. tremula and hybrid aspen groups. Selected leaf trait parameters were:
PCA based on Pearson correlation and distance-based biplot was applied for 7 selected leaf trait
parameters and revealed that the first two components in diagram explain 97.8% of the variation
(Figure 3.22). Perimeter together with v_length, h_width and width2 forms one group in the
component pattern; the other parameter group is represented by leaf lobe angle traits (angle1
and angle2), while leaf form coefficient (form) is clustered separately.
Fig. 3.22. Principal component pattern of seven selected leaf trait parameters (form, v_length, perimeter, h_width, width2, angle1, and angle2), that best discriminates between European and hybrid aspens
Mother trees of 9 half-sib families were clustered together using Euclidean distances according
to 7 important leaf trait parameters (Figure 3.23). In the resulting dendrogram two separate
clades are formed: in the upper clade P. tremula mother trees of all three flowering categories
together with early flowering hybrid aspen mother tree are grouped; the second clade is
comprised of two hybrid aspen mother trees and early and medium flowering P. tremula trees.
The dendrogram revealed that most closely related mother trees are late flowering P. tremula
trees and early flowering P. tremula with late flowering hybrid aspen tree.
85
Fig. 3.23. Clustering of European and hybrid aspen mother trees using Euclidean distance according to
7 selected leaf trait parameters (form, v_length, perimeter, h_width, width2, angle1, angle2) of their offspring
Multivariate permutation (n=10,000) of the offspring, belonging to both (European and hybrid
aspens) groups, using Euclidean distance measure revealed, that hybrid and European aspen
trees differ significantly in means of seven selected leaf trait parameters. However, discriminant
analysis of European and hybrid aspen groups using seven important trait parameters classified
correctly only 61% of individuals, though F value was 15.4 and p<0.0001 (Hotelling’s
t2=108.3) (Figure 3.24).
Fig. 3.24. Discriminant analysis of European and hybrid aspen group trees, using seven selected leaf trait parameters. Grey color indicates European aspen and white color – hybrid aspen group trees
86
Variance components estimates presented in Table 3.14 show that seedling influence on all the
trait parameters is several times higher than that of the family. Family influence is also
Progenies of two aspen maternal trees had significantly more seedlings classified into hybrid
group (tree No. 9 – 76% and tree No. 28 – 47%) compared with the remaining P. tremula
mother trees. The same results were obtained also by clustering mother trees according to their
progeny leaf trait parameters (both trees were clustered together with hybrid aspen mother trees
in the lower clade of the dendrogram, see Figure 3.23).
The leaf variability of European and hybrid aspens’ one-year-old seedlings is very high,
therefore it was complicated to trace possible hybrid seedlings based only on leaf morphology
analysis. Obviously, to trace possible hybrids in obtained half-sib families, additional
application of molecular methods would be required. Phenological structure of aspen and hybrid
aspen stands is helpful in initial phase of screening to predict possible gene flow, because of the
timing in flowering phenology – European aspen trees had the highest number of progenies
classified to the hybrid aspen group. This was expected, as the hybrid aspen group is
characterized by earlier flowering time than European aspen.
87
CONCLUSIONS
1. The best methods for DNA extraction from various tissues of Populus tremula (giving
the highest amounts of good quality DNA) are CTAB and protein precipitation, and
commercially available kits Nucleospin Plant Mini, Genomic DNA purification kit and
innuPREP Plant DNA Kit;
2. Weak correlations between genetic and geographic distances were obtained using both
RAPD and SSR markers(r=0.178 and r=0.068, respectively), indicating that the largest
part of molecular variance can be attributed to a within-population variation;
3. RAPD data allowed genotype clustering into groups that are geographically closer to
each other compared to SSR-based cluster groups;
4. Microsatellite analysis revealed that individuals of P. tremula should be grouped into
three clusters, while RAPD analysis clearly showed only two clusters. The results
obtained by both clustering methods (UPGMA and Bayesian) indicate that existing P.
tremula provenance regions in Lithuania may be revised as both molecular markers
revealed latitudinal trend;
5. Population, provenance region and individual heterozygosity have little or no influence
on infection of P. tremula plus trees with trunk rot fungus Phellinus tremulae. Two
RAPD loci among the 282 loci tested showed association with the wood infection and
thus could be used for development of valuable molecular markers in tree breeding at
early ontogenesis stages;
6. Phenological structure of P. tremula and hybrid aspen stands is helpful for initial
prediction of possible gene flow among aspen populations: trees with the earliest
flowering phenology produced the highest proportion of hybrid seedlings.
88
Conclusions on the defended statements:
Statement Conclusion Comment
1. Genetic differentiation between Lithuanian populations of European aspen correlates with their geographic distribution
Rejected Weak correlations between genetic and geographic distances using both RAPD and SSR markers’ data (0.178 and 0.068) were obtained indicating that the largest part of molecular variance can be attributed to within population variation
2. Genetic diversity among local (Lithuanian) European aspen populations is low because of a high gene migration rate among them
Accepted Only 4% of an assessed genetic variation is attributed to differentiation among P. tremula populations, while 96% of the variation lies within populations
3. The relationship between susceptibility of European aspen to trunk rot caused by Phellinus tremulae and tree genetic properties can be revealed using RAPD analysis
Partially accepted Five RAPD fragments were important in differentiation of P. tremula trees by presence of P. tremulae infection, and only two of them, A5-850 and B13-1100, were significant (at p 0.05) in tree differentiation into susceptibility categories
4. Hybridization between hybrid and European aspens can be revealed by morphological leaf traits of their progenies at juvenile age
Partially accepted It is complicated to trace possible hybridization between hybrid and European aspens based only on leaf morphology and flowering phenology. To trace possible hybrids in half-sib families of P. tremula, application of molecular methods must be considered
89
REFFERENCES
1. Aagaard J.E., Krutovskii K.V., and Strauss S.H. 1998. RAPD markers of mitochondrial
origin exhibit lower population diversity and higher differentiation than RAPDs of nuclear
origin in Douglas fir. Molecular Ecology 7: 801–812.
2. Abbott R., Albach D., Ansell S., Arntzen J.W., Baird S.J.E., Bierne N. et al. 2013.
Hybridization and speciation. Journal of Evolutionary Biology 26: 229–246.
3. Allen E., Morrison D., and Wallis G. 1996. Common Tree Diseases of British Columbia.
Canadian Forest Service, Pacific Forestry Centre, Victoria, British Columbia, Canada.
4. Lietuvos mišk statistika. 1937. Lietuvos Mišk Departamentas, Kaunas. (In Lithuanian).
5. Lithuanian Statistical Yearbook of Forestry. 2014. State Forest Service, Kaunas.
6. Areshchenkova T., and Ganal M.W. 1999. Long tomato microsatellites are predominantly
associated with centromeric regions. Genome 42: 536–544.
7. Arnold M. 1997. Natural hybridization and evolution. Oxford University Press, New
York, NY.
8. Arnold M., and Martin M. 2010. Hybrid fitness across time and habitats. Trends in
Ecology and Evolution 25: 530–536.
9. Asif M.J., and Cannon C.H. 2005. DNA extraction from processed wood: a case study for
the identification of an endangered timber species (Gonostylus bancanus). Plant
Molecular Biology Reporter 23: 185–192.
10. Bakys R., Vasaitis R., Barklund P., Thomsen I.M., and Stenlid J. 2009. Occurrence and
pathogenicity of fungi in necrotic and non-symptomatic shoots of declining common ash
(Fraxinus excelsior) in Sweden. European Journal of Forest Research 128: 51–60.
– no polymorphic DNA fragments, – few polymorphic DNA fragments, – clear polymorphic DNA fragments; * – most DNA fragments are similar in size, ** – most DNA fragments are similar in size and scorable at some extent, *** – DNA fragments are easy to score; – DNA fragments are not reproducible, – most of the DNA fragments are reproducible yet indistinct, – DNA fragments are reproducible and easy to score; + – no amplification, ++ – low quality amplification, +++ – good quality amplification. Primer evaluation carried out according to Pivorien (2008)