Vegetative propagation: linear barriers and somatic mutation affect

Forestry
An International Journal of Forest Research
Forestry 2015; 88, 612 – 621, doi:10.1093/forestry/cpv029
Advance Access publication 30 July 2015
Vegetative propagation: linear barriers and somatic mutation affect
the genetic structure of a Prunus avium L. stand
Kristjan Jarni1*, Jernej Jakše2 and Robert Brus1
1
Department of Forestry and Renewable Forest Resources, Biotechnical Faculty, University of Ljubljana, Ljubljana 1000, Slovenia
2
Department of Agronomy, Biotechnical Faculty, University of Ljubljana, Ljubljana 1000, Slovenia
*Corresponding author. Tel: +386 13203539; Fax: +386 12571169; E-mail: [email protected]
Received 11 November 2014
Microsatellite markers were used to describe the genetic structure of a natural wild cherry (Prunus avium L.)
stand in Slovenia. Based on eight analyzed loci, only 67 different multilocus genotypes (MLGs) were identified
among 217 trees, indicating a significant amount of clonal reproduction in the stand. Low spatial genetic structure (SGS) was observed in the stand when only sexually derived genets were considered (Sp ¼ 0.011), and the
kinship coefficient was only significant in the first distance class (,40 m). When both the generative and vegetative origin of trees were included, the intensity of the SGS in the stand increased (Sp ¼ 0.149). Forest paths,
streams and ditches, which represent obstacles to root growth and consequently obstruct vegetative propagation via root suckers, also affected the spatial grouping of clones in the stand. A relatively high number of somatic
mutations within clonal groups were observed, which further increased the complexity of the genetic structure
in the stand.
Introduction
Prunus avium L. is native to East Eurasia and North Africa and is the
most important European woody species of the Rosaceae family
(Russell, 2003). It is valued in forests for its high-quality wood
and ecological role. Its nutrient-rich leaves improve soil fertility
and its fruits are an important food source for many species of
animals, particularly birds, which play a key role in maintaining
the dynamic stability of the entire ecosystem. In nature, it occurs
in at least two different successional stages. Because of its ability
to sprout from its roots, it behaves as a pioneer species that can
easily colonize abandoned areas, often forming large groups or discontinuous stands. It can also be an element of climax stands,
however, where it is mostly scattered or grows in small groups
(Gömöry, 2004).
As a species with gametophytic self-incompatibility, P. avium
maintains significant genetic variability through sexual reproduction (Ganopoulos et al., 2012). However, the intrapopulation variability of cherry stands is strongly reduced by vegetative
propagation via root suckers and the formation of clonal groups
(Schueler et al., 2006; Vaughan et al., 2007a; Jolivet et al., 2011;
Jarni et al., 2012). The factors that impact genetic variability and
population structure can be broadly divided into (1) general
(direct) factors which relate to a species (or groups of species),
e.g. breeding system, life form, pollen and seed dispersal
(Vekemans and Hardy, 2004) and (2) specific (indirect) factors
which affect variability and structure indirectly through gene
flow, e.g. presence/absence of gene flow vectors such as wind,
waterways, animals and management system (Ducci and Santi,
1997; Oddou– Muratorio et al., 2004; Garcia et al., 2005; Vaughan
et al., 2007a; Brus et al., 2010; Ganopoulos et al., 2013). These
factors often interact, making it very difficult to address them
independently.
Somatic mutations have also been found to affect the genetic
variability of woody species which often propagate vegetatively
in nature, e.g. Populus tremuloides (Ally et al., 2008), Populus
nigra (Chenault et al., 2011) and Robinia pseudoacacia (Lian
et al., 2004) or in ex situ conditions, e.g. Vitis vinifera (Crespan,
2004; Vezzulli et al., 2012), Olea spp. (Lopes et al., 2009), Pinus
pinaster (Marum et al., 2009) and Picea abies (Helmersson
et al., 2008). As P. avium frequently propagates vegetatively, it
is not surprising that somatic mutations have also been confirmed in the species (Vaughan et al., 2007b). The occurrence of
mutations – permanent inclusion of random errors in DNA –
lead to differences between original and copied DNA sequences
and per se represent a basic source of genetic variability (Hamilton, 2009). In P. avium such mutations may inter alia occur in the
early phase of meristem development, when a new individual
grows from adventitious buds on roots (D’Amato, 1997; Vezzulli
et al., 2012). Thus, they can be passed on to subsequent ramet
generations. Furthermore, because plants do not sequester
their germline, these mutations can be transmitted to reproductive organs and subsequently to sexual progeny (O’Connell and
Ritland, 2004; Ally et al., 2010). As such, somatic mutations are
an important factor of ‘clonal evolution’ in plants, which
depends on (1) the age of the clone (the longer it propagates
vegetatively, the longer it is subject to stress conditions and the
# Institute of Chartered Foresters, 2015. All rights reserved. For Permissions, please e-mail: [email protected].
612
Linear barriers and somatic mutation affect the genetic structure of a Prunus avium L. stand
more mutations can accumulate), (2) various environmental
stresses and (3) genotype, since some genotypes are more susceptible to mutations than others (Pelsy, 2010). In addition,
this difference in susceptibility is also present at the level of individual loci and alleles (Ellegren, 2004).
Previous genetic analysis of P. avium in a Slovenian natural
population in Vipavska brda (Jarni et al., 2012) showed the strong
presence of vegetative propagation via root suckers. Therefore, in
this study we conducted a detailed analysis of this population in
which all P. avium trees in the stand were included. The aims of
this study were (1) to obtain in-depth insight into the spatial
genetic structure (SGS) of the species in this particular stand, (2)
to identify any anomalies in its genetic structure, such as somatic
mutations, and (3) to determine the effect of linear barriers
(e.g. forest paths, streams and ditches) on its spatial vegetative
propagation.
Materials and Methods
Study area
The 2.0 ha study area (coordinates: 45846′ 25′′ N; 13859′ 31′′ E) is a natural
P. avium stand on an Ornithogalo pyrenaici-Fagetum sylvaticae site. The
stand is located at the bottom of an 70-m wide depression and is geographically restricted by steep slopes on all four sides. Eutric brown soil is
dominant in the area. Although the stand is located in a broader extensive
farming area, we infer from historical records that forest has been continuously present on the study area for at least 200 years (Anonymous, 1997).
The structure of this stand, however, indicates that it probably experienced
large canopy openings in the past, resulting in substantial recruitment of
sexual and asexual individuals.
Plant Mini Kit (Qiagen) according to the supplier’s protocol. All individuals
were genotyped at nine nuclear microsatellite loci using two multiplex
PCR reactions (Vaughan and Russell, 2004) as well as three simplex PCRs.
In multiplex reactions the following primers were combined: Multiplex–A:
EMPaS12 and EMPaS14; Multiplex–B: EMPa004, EMPa005, EMPaS02,
EMPaS06. Loci EMPa15, UDP98-412 and PceGA34 were analyzed in
simplex reactions. The forward primers of each primer pair were labelled
with fluorescent dyes FAM, HEX or NED (Applied Biosystems). The Multiplex
PCRs were performed according to the protocol of Vaughan and Russell
(2004). The simplex PCR for locus EMPa15 followed the protocol of Clarke
and Tobutt (2003). The amplification of the remaining two simplex PCRs
was carried out in a total volume of 15 ml containing 0.8 mM dNTPs,
0.4 mM of primer, 0.6 U of Taq DNA polymerase, 2.5 mM of MgCl2, 1× PCR
buffer and 5 ml of genomic DNA. They were amplified by the following thermocycler programme: 948C for 2 min, followed by 35 cycles of 948C for 30 s,
Ta8C for 45 s (the annealing temperature (Ta) for locus UDP98-412 was 608C
and 558C for PceGA34), and 728C for 60 s with an elongation step of 728C for
5 min. The same volume of formamide and 3 ml of GeneScanTM 500 ROX
standard as internal standard (Applied Biosystems) were added to the
PCR products. They were genotyped with a capillary sequencer
(ABI3130XL, Applied Biosystems) and the resulting flowgrams were
scored using Peak ScannerTM 1.0 (Applied Biosystems) software.
Micro-Checker software (Van Oosterhout et al., 2004) was used to detect
genotyping errors resulting from the presence of null alleles, stutter peaks
and large allele dropout. At locus EMPaS14, where we identified only
three alleles, we detected a pattern in which the shorter allele (198 bp)
was amplified normally, while we noticed some very low peaks on the
curve at the plausible location of the second allele (values below 200 fluorescent units) on the location of the longer allele (213 bp), suggesting the
probable existence of an allele. Although the locus analysis did not detect
the presence of null alleles or large allele dropouts, we were unable to get
clear and unambiguous results for this locus in three repetitions. Because
such errors can affect the findings, we excluded locus EMPaS14 from
further analysis.
Geographic and tree characteristic measurements
Genetic analysis included all 217 P. avium trees with diameter at breast
height (DBH) .5 cm. The DBH of these trees was measured with a p-tape
with 1-mm precision. The diameter distribution of P. avium in the stand
was tested for normality with the Kolmogorov–Smirnov test. Differences
in average diameters in individual clonal groups were tested with the
Kruskal– Wallis test. The tests were performed with IBMw SPSSw Statistics.
The precise location of trees and linear barriers to vegetative propagation
(forest paths, streams and ditches) were determined with a Trimble
GeoXT device (Trimble Navigation, Ltd.). GPS Pathfinder Office Software
(Trimble Navigation, Ltd.) was used to transfer the data to a PC and for postprocessing. Clone surfaces were calculated by forming polygons from the
outer trees of individual clones and calculating the area with MapSource
(Garmin Ltd.). Several clonal groups containing multiple ramets were identified. For each clonal group (altogether 28 groups, see Table 1), spatial coordinates were averaged for the mid-point of the group, which represented
the hypothetical location of the generative progenitor. After genetic analysis was completed, the five to six largest DBH trees in the seven most numerous multilocus lineages (MLLs) were cored (38 trees in total) with an
increment borer to determine age. The cores were prepared prior to analysis
with established dendrochronological methods (Stokes and Smiley, 1968).
Counting annual rings was performed with a Nikon SMZ80 stereoscopic
microscope with 10× magnification.
DNA extraction and SSR analysis
For genetic analysis young leaves or cambium from all 217 trees were collected and put directly on silica gel until DNA extraction. Approximately
20 mg of dried sample was ground in an automatic grinding mill (TissueLyser LT, Qiagen) and afterwards total DNA was extracted with the DNeasyw
Data analysis
Identification of MLGs and MLLs followed standardized methods proposed
by Arnaud-Haond et al. (2007) with GenClone 2.0 software (Arnaud-Haond
and Belkhir, 2007). When taking into account departures from Hardy–
Weinberg equilibrium (using FIS), the probability (pgen) of occurrence of
each observed genotype was estimated according to Young et al. (2002):
pgen (FIS ) =
l
[( fi gi ) × (1 + zi FIS(i) )]2h ,
(1)
i=1
where l is the number of loci, h is the number of heterozygote loci, f and g are
the allelic frequencies of the alleles f and g at the ith locus and zi ¼ 1 (or 21)
if the ith locus is homozygous (or heterozygous). When the same genotype
is detected n times in a sample of N trees, the probability that the repeated
genotype is derived from a distinct sexual reproductive event was estimated following Parks and Werth (1993):
psex (FIS ) =
N
i=n
N! pgen (FIS ) i 1 − pgen (FIS ) N−i .
i!(N − i)!
(2)
The significance of psex was considered from the first reencounter (n ¼ 1). To
ascertain the uniqueness of MLGs with missing data (e.g. unamplified loci),
such MLGs were examined on a case-by-case basis after removing the
missing loci from the entire dataset. Based on the recalculated psex estimates, these MLGs were either classified as being unique or were assigned
to another MLG into a MLL. A similar approach was used to assign MLGs that
differed at only one locus into MLLs in order to account for somatic
613
Forestry
Table 1 Basic indicators of Prunus avium trees in the Vipavska brda stand: breast diameter, distances among trees of the same genotype as revealed
by microsatellite analysis and age structure of the seven analysed largest clone groups
DBH (cm)
mean
min
max
Distance between trees
(m)
NR3
Mean
SD
dmax
d
neighb
16
11
9
22
21
9
12
3
4
2
2
115
2
6
8
5
2
6
2
6
6
4
2
3
6
4
7.1
2
22
24.45
23.97
18.67
20.74
29.13
29.19
28.39
19.87
20.63
23.50
21.35
21.45
21.90
26.82
22.90
28.60
22.85
24.42
25.05
25.88
18.77
37.83
23.65
16.27
24.05
23.48
23.99
16.27
37.83
5.86
9.48
8.73
6.53
6.70
3.14
8.64
3.06
4.25
4.24
0.35
5.92
2.26
3.29
6.93
4.13
0.35
4.83
3.61
7.73
6.97
8.81
2.76
6.60
3.75
4.21
5.12
0.35
9.48
29.71
23.50
10.26
20.77
30.05
17.63
34.20
19.19
26.54
2.83
5.02
20.19
3.00
11.86
55.72
13.47
7.27
25.95
2.74
40.59
9.97
8.14
5.57
7.44
15.51
9.09
17.54
2.74
55.72
3.44
3.61
2.07
2.11
3.20
3.07
5.34
7.36
7.37
2.83
5.02
3.45
3.00
1.79
7.86
4.78
7.27
6.64
2.74
6.87
2.71
4.17
5.57
2.90
3.35
3.70
4.32
1.79
7.86
Area1 (m2)
334
218
37
162
374
82
357
3
76
/
/
133
/
18
214
72
/
211
/
281
36
26
/
188
33
13
/
/
374
Age2 (years)
Mean
Interval
70.804
61.804
67.504
75.804
70.004
67.334
73.504
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
66 –78
45 –71
59 –90
70 –79
60 –78
58 –74
69 –81
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
dmax ¼ maximum distance between ramets; d
neighb ¼ average distance between nearest neighbours.
Areas were not calculated where clone contained fewer than three ramets (N , 3).
2
Age was determined in the seven largest multilocus lineages (MLLs) only. In these groups we analyzed the five or six thickest trees.
3
Number of ramets in MLL.
4
Number of trees for which the age was determined.
5
This group became one of the largest MLLs only after the inclusion of mutated individuals, and was not included in the analysis of age.
1
mutations or scoring errors. The genotypic richness (R) of all P. avium trees
was estimated as R ¼ (G 2 1)/(N 2 1) where N is the number of sampled
trees and G the number of MLLs (Dorken and Eckert, 2001). Observed heterozygosity (HO) was estimated using Arlequin software (Excoffier et al.,
2005).
In order to describe the frequency distribution of clonal membership,
the Pareto model was used as proposed by Arnaud-Haond et al. (2007).
This model takes the form N≥X ¼ aX2b where N≥X is the number of
sampled ramets belonging to MLLs containing X or more ramets. The parameter b indicates the scaling of partitioning of the ramets among MLL
size classes. Spatial descriptors (1) maximum distance between ramets
(dmax ) and (2) average distance between nearest neighbours (d
neighb )
were computed for each MLL. Existence of spatial aggregation of clonemates was estimated with an aggregation index (AC) as proposed by
Arnaud-Haond et al. (2007): AC ¼ (psg 2 psp)/psg where psg is the average
614
probability of clonal identity of all sample unit pairs and psp the average
probability of clonal identity among pairwise nearest neighbours. The significance of AC was estimated by a 10 000 permutation test. The relationship between MLL size (number of ramets (NR)), maximum distance
(dmax ), average distance between nearest neighbours (d
neighb ) and DBH
was assessed with Spearman’s correlation coefficients performed with
IBMw SPSSw Statistics software.
SGS was assessed using SPAGeDi software (Hardy and Vekemans, 2002).
The kinship coefficient Fij (Loiselle et al., 1995), which measures correlations
in the frequencies of homologous alleles in pairs of individuals, is calculated
as
k )( pj − p
k )
( pi − p
Fij = l k
(
p
(1
−
p
)
+
1/(2n
− 1))
k
l
k k
(3)
Linear barriers and somatic mutation affect the genetic structure of a Prunus avium L. stand
k is average frequency of allele k at locus l in a population with n indiwhere p
viduals and pi and pj are frequencies of allele k in individuals i and j (Hardy
and Vekemans, 2002). To visualize SGS, kinship coefficients were plotted
against 40-m distance classes on a logarithmic scale to generate spatial
genetic autocorrelograms. To verify the significance of Fij, confidence
intervals (95 per cent) were obtained by permuting individuals among locations 10 000 times. The intensity of SGS was assessed with ‘Sp’ statistics
(Vekemans and Hardy, 2004) using SPAGeDi software, where Sp ¼ bF/
(1 2 F1). In our case, F1 was the kinship coefficient Fij between adjacent
individuals (pairs of individuals separated ,40 m). The impact of vegetative
propagation on SGS was calculated as SGS by per centclonal ¼ 100 – 100×
Spind/Spall (Schueler et al., 2006), where Spind is the Sp statistic calculated
without clonal propagules and Spall is the Sp statistic including all trees.
Results
Clonality
Among the 217 trees, we were able to genotype 211 trees fully at
the eight microsatellite loci, and these clustered into 64 MLGs. All
ramets within a MLG were associated with a psex value below
1023. The six remaining trees had one (5) and two (1) loci
missing. By sequentially removing the missing loci before reanalysing the data, it was possible to assign three of these trees to previously identified MLGs (psex , 0.01) and three to new MLGs. By
sequentially removing the mismatched loci for MLGs differing at
only one locus (Figure 1) and in one case at two loci (in this particular case the difference in allele length was only 2 × 2 bp and the
spatial distance to the nearest neighbour belonging to the same
MLL was 10.1 m), we clustered all 67 MLGs into 59 MLLs (psex ,
0.01) of which 26 MLL≥2.
The gentle slope of the Pareto distribution b ¼ 0.63+0.10 (r 2 ¼
0.769, P , 0.001) indicates a skewed distribution within the population, mostly consisting of a few prevalent clonal lineages and
many small ones. MLL size (NR) ranged from one to 22 ramets,
whereas 85 per cent of MLLs consisted of six or less ramets. The
genotypic richness (R) was 0.27.
The estimated aggregation index (AC) was 0.70 (P , 0.001),
showing a significant level of spatial aggregation of ramets belonging to the same MLLS. The geographic size of MLLs, as measured by
the maximum distance between two ramets (dmax ), ranged from
2.7 to 55.7 m, while intra-MLL average distance between nearest
neighbours (d
neighb ) ranged from 1.8 to 7.8 m. Maximum distance
(dmax ) was significantly related to MLL size (NR), rs ¼ 0.78, P ,
0.001, while we did not find a relationship between NR and
(d
neighb ) or between NR and DBH.
Natural and artificial barriers to vegetative propagation
of P. avium in the stand
Natural barriers identified and located in the study area included
two streams, which dry up in summer, and two ditches; forest
paths represented barriers of anthropogenic origin (Figure 2).
Trees from the same MLLs typically do not bridge proximal barriers.
The beds of the two streams, which are on average 3 –4 m wide and
extend down to the bedrock, represent a strong barrier to vegetative propagation via root suckers. Forest paths and natural ditches
represent slightly weaker barriers (Figure 2); they hamper but do
not prevent vegetative propagation. The soil on forest paths and
the bottom of natural ditches is shallower than the surrounding
soil, but complete absence of developed soil (i.e. bedrock visible
on the surface) occurs only sporadically. Paths and ditches in the
stand are important infrastructure for water discharge during
heavy rainfall. This causes erosion and prevents the development
of a deeper, more developed soil type.
Diameter and age structure of the stand
The diameter distribution shows normal distribution and hence
even-sized structure of P. avium individuals in the stand (data not
shown). We did not find deviation from the normal distribution
with the Kolmogorov– Smirnov test (P . 0.05). DBH of all 217
P. avium individuals was in the 8.8 –50.1 cm interval with an arithmetic mean of 24.5 cm (data not shown).
Average DBH of trees in the seven largest MLLs was 18.6 –
29.2 cm (Table 1). Differences in average DBH in these groups
were confirmed with the Kruskal–Wallis test H(6) ¼ 24,74, P ,
0.001. The average age of the 5 –6 largest DBH trees in the
groups was 61 –76 years. Intra-group age differences were 9 –31
Figure 1 Frequency distribution (%) of genetic distances (differences in allele length) among ‘Prunus avium’ trees in the Vipavska brda stand. The dashed
line represents the threshold below which different MLGs (after excluding the slightly different loci) originated from the same MLL.
615
Forestry
Figure 3 Spatial correlograms: kinship coefficient values (solid line, Fij) are
presented for all ramets of P. avium trees for each 40 m distance class on
a logarithmic scale. Dashed lines denote 95% confidence interval.
Figure 4 Spatial correlograms: kinship coefficient values (solid line, Fij)
are presented for one ramet per clonal group of P. avium trees for each
40 m distance class on a logarithmic scale. Dashed lines denote 95%
confidence interval.
Figure 2 Spatial distribution of P. avium trees in the Vipavska brda stand in
relation to natural barriers and pathways. The symbol groups represent
different genotypes, with the exception of black diamonds (V), which
represent ‘unique’ trees that do not belong to any clonal group. In the
case of forest paths, the thicker line indicates a broader and more sunken
path, while the thinner line represents a narrower and less sunken path.
years and the total span of ages across all analyzed trees was
45 –90 years.
SGS of the stand
Microsatellite markers showed that vegetative propagation of
P. avium had a large impact on its SGS in the stand (Figures 3 and 4).
The consequence of spatial proximity of identical genotypes is
the non-random distribution thereof. The kinship coefficient (Fij)
calculated for all ramets reaches its maximum in the first distance
class (,40 m). Below this distance, the value of the Fij coefficient
is significantly larger than expected and drops steeply as the distance increases. When only one ramet per clone was used in analysis, the degree of kinship between trees in the stand decreases
significantly. Comparison of Fij for the first distance class (,40 m)
616
shows that the coefficient is more than four times lower if only
one ramet is included and that significant SGS is detected only in
the first distance class (,40 m) (Figure 4). The fluctuation of Fij at
the largest distance classes in both figures is random due to the
small number of individuals compared.
Total SGS intensity evaluated with Sp statistics (Vekemans and
Hardy, 2004) shows that the degree of kinship between trees in
the stand strongly correlates to their spatial proximity. Taking
into account both generative reproduction and vegetative propagation, the stand has strong SGS (Sp ¼ 0.149 (Table 2)). Sp values
for individual loci are high and range from 0.054 (EMPaS12) to
0.302 (EMPa004). Using only one ramet per clone, the Sp value
for the entire stand decreases significantly (Sp ¼ 0.011) but still
indicates significant SGS (Vekemans and Hardy, 2004). Sp values
for individual loci range from 0.012 (EMPa004, PceGA34) to 0.061
(EMPaS02). The contribution of vegetative propagation to SGS is
large (93 per cent) and ranges from 2 per cent (EMPaS12) to 96
per cent (EMPa004) in individual loci.
Somatic mutations
Five individual trees and two groups (with two individuals each),
were differentiated from the other, typically more populous
groups at one of the eight analyzed loci and in one case differences were observed at two loci (Table 3). In all cases, the
trees (groups) were located immediately adjacent to each
other or were intermixed (Figure 2). These results indicate the
Linear barriers and somatic mutation affect the genetic structure of a Prunus avium L. stand
Table 2 Spatial genetic structure of all P. avium trees and one ramet per clonal group in the Vipavska brda stand
All trees (N ¼ 217)
One ramet per clonal group (N ¼ 59)
Locus
b
F1
Sp
b
F1
Sp
%clonal
EMPaS02
EMPa004
EMPa005
EMPaS06
EMPaS12
EMPa015
UDP98-412
PceGA34
all loci
20.099***
20.274***
20.172***
20.127***
20.052***
20.093***
20.124***
20.124***
20.135***
0.174
0.092
0.120
0.083
0.039
0.072
0.052
0.102
0.092
0.119
0.302
0.196
0.138
0.054
0.100
0.131
0.138
0.149
20.057 n.s.
0.012 n.s.
20.015 n.s.
0.013 n.s.
0.053 n.s.
0.046 n.s.
0.023 n.s.
0.012 n.s.
0.011 n.s.
0.062
0.020
0.003
0.022
0.000
0.006
0.005
0.031
0.019
0.061
0.012
0.015
0.013
0.053
0.046
0.023
0.012
0.011
49
96
92
91
2
54
82
91
93
b ¼ regression slope of pairwise kinship coefficient Fij on the logarithm of geographical distance; F1 ¼ average Fij in the first distance class; Sp ¼ Sp statistics,
indicator of SGS (Vekemans and Hardy, 2004), %veg ¼ impact (proportion) of vegetative propagation on the SGS.
Significance of regression slope ‘b’ were tested with a one-side Mantel test with 10 000 permutation (n.s. P . 0.05; *0.01 , P , 0.05; **0.001 , P , 0.01;
***P , 0.001).
Table 3 Allelic variation observed at individual microsatellite loci in separate clonal groups of P. avium species in the Vipavska brda stand
N
Sample material
EMPaS02
EMPa004
EMPa005
EMPaS06
EMPaS12
EMPa015
UDP98-412
PceGA34
1
18
1
1
11
1
2
2
7
2
5
1
3
1
Leaf
Leaf/cambium
Cambium
Leaf
Leaf/cambium
Cambium
Cambium
Leaf
Leaf
Leaf/cambium
Leaf
Cambium
Leaf
leave
141
141
141
141
146
146
146
146
141
141
146
146
146
146
194
194
194
194
184
182
192
192
190
190
184
184
184
184
249
249
249
249
247
247
245
247
249
249
258
258
247
247
206
206
206
206
206
206
206
206
206
206
206
206
206
206
138
138
138
138
138
138
145
145
136
136
138
138
138
138
223
223
223
221
225
225
225
225
225
223
225
223
225
225
113
117
117
117
123
123
119
119
123
123
117
117
117
117
140
140
140
140
132
132
140
140
161
161
134
134
153
153
148
148
148
148
148
148
148
148
146
146
148
148
148
148
194
194
196
194
194
194
192
192
192
192
184
184
194
194
258
258
258
258
258
258
258
258
258
258
258
258
249
249
224
224
224
224
224
224
208
208
206
206
208
208
210
210
145
145
145
145
145
145
145
145
138
138
145
145
138
145
225
225
225
225
225
225
225
225
256
251
254
251
225
225
117
119
119
119
123
123
123
123
123
123
119
119
119
119
151
151
151
149
140
140
151
151
165
165
140
140
159
159
The original genotype (italicized) is determined as that containing the largest number of ramets. When both genotypes have the same number of ramets,
the original is the one whose particular allele is more frequent. (N) is the number of ramets in the group. Bold values indicate variation of alleles at individual
loci.
possible occurrence of somatic mutations at individual loci.
Given that the stand age is relatively high and that the life expectancy of P. avium is 80 – 100 years, we assume that many of
the founder trees of the existing clonal groups are no longer
present in the stand, which may be indicated by lying trees and
deadwood identified in the area. For mutations, we therefore
assumed the genotype that was better represented in the
stand is the original genotype. When both genotypes were
equally numerous, the genotype whose particular allele was
the more frequent allele in the total population was assumed
to be the original genotype.
At loci EMPa004 and EMPa005, the subgroups differed from the
original by one tandem repeat (2 bp) (Table 3). An insertion at locus
EMPa004 created a new 196 bp allele that had not previously
existed in the stand, while a deletion created a 182 bp allele that
had existed in the stand but was relatively infrequent. At locus
EMPa005, deletion created a 245 bp allele which had not previously
existed in the stand. At locus EMPaS12, the new allele, which is 7 bp
longer than the original one, is relatively common in the stand,
while at loci EMPa015 and UDP98-412, we found mutations on
both alleles at the locus. At locus EMPa015, deletion of one
tandem repeat (2 bp) occurred on the shorter alleles, and at the
larger locus, the differences between the original and new alleles
are 3 and 5 bp, respectively. The new 223 bp allele is relatively
common in the stand, while the 251 bp allele is new. At locus
UDP98-412, the new 113 bp allele is rare (identified on only one
more tree in the stand), while the 117 bp allele is the most
common.
617
Forestry
Discussion
Diameter and age structure of the stand
DBH analysis of P. avium shows homogenous structure while age
analysis shows uneven-aged structure of the stand (Table 1),
which means that the stand origin is not related to a single past
event that could have triggered regeneration, such as a large
natural or artificial disturbance, although the vegetative origin of
trees and DBH structure might indicate this. Our analysis showed
that DBH structure is a poor indicator of age for P. avium, as had
already been reported by Vaughan et al. (2007b), but on the
other hand, tree diameter can be more strongly correlated with
age if the trees do not grow in a dense stand but in a more open
area (Gömöry, 2004). One of the reasons for the low correlation
between DBH and age in our case is poor tending treatment in
the past, evident in the shaded and exceptionally small crowns
of many adult trees, which can affect radial growth. In the past
15 –25 years, the annual increment of such trees was only a
single layer or several layers of cells, compared with trees with
more growing space and hence larger radial growth. Poor tending
is reflected not only in small crowns but also in poor tree vitality, exceptionally slow growth and relatively high age for the species.
SGS of the stand
Recent molecular marker studies have shown significant SGS for
P. avium, which confirms the species’ limited pollen and seed dispersal (Schueler et al., 2006; Vaughan et al., 2007a; Jolivet et al.,
2011). Schueler et al. (2006) used microsatellite markers and
S-alleles in Germany and found that P. avium trees in a stand are
genetically more similar than expected by chance up to a distance
of 85 m. Our study also confirms the strong and significant SGS
of P. avium, which could be the consequence of substantial vegetative propagation in combination with high conspecific density
(Jolivet and Degen, 2011). Indeed, at least 69 per cent of all analyzed trees in the stand have vegetative origin, and in addition,
the MLL size (NR) distribution (data not shown) and Pareto index
(b ¼ 0.63) indicate dominance of some large clonal lineages and
the additional presence of many small ones. Calculated for all
trees, kinship coefficients (Fij) have the highest values in the first distance class (,40 m) and significantly (four times) lower values
when taking into account only trees of generative origin (one
ramet/clone; Figures 3 and 4). Similarly, a decrease in the Fij of
trees of purely generative origin was reported by Vaughan et al.
(2007a), who confirmed significant SGS up to a distance of 120 m.
The difference in the ‘length’ of SGS between our study and that
of Vaughan et al. (2007a) likely occurred due to the difference
in population density (Jolivet et al., 2011). While the density of
P. avium in our stand was 110 trees ha21, in the two British populations it was 17 and 24 trees ha21, respectively. The effect of the
density of P. avium trees on SGS was also confirmed by simulations,
where higher density yields a ‘shorter’ SGS (Jolivet and Degen, 2011).
On the other hand, the simulations showed that the kinship coefficient is greatest at lower densities. This was in contrast with real
populations (Jolivet et al., 2011), where the highest coefficients
were in more dense populations. Large Fij was also revealed in our
study, most probably resulting from extensive clonal propagation,
which is known to inflate the strength of SGS at shorter distances
(Jolivet and Degen, 2011). Asuka et al. (2004) confirmed the effect
618
of tree density on SGS for Fagus crenata (Blume) where high densities
lead to lower SGS due to the strongly overlapping seed shadow of different mothers. As shown by simulation (Jolivet and Degen, 2011),
this could also apply for P. avium, for which the overlapping of
seed shadows and thinning effect could lead to the reduction of
SGS through generations.
The significant positive correlation between NR and dmax and the
absence of connection between NR and d
neighb may indicate that
clonal growth in this stand is an expansion process rather than
one that leads to the densification of clonal patches. However,
the general SGS profile of the stand fits the ‘isolation by distance’
model, where the Fij between closer trees is higher than expected
and lower among distant trees (Rousset, 1997, 2000; Hardy and
Vekemans, 1999). When calculated for all ramets, the profile is
even more explicit, which means vegetative propagation is a significant factor in SGS formation. This is also confirmed by comparison of SGS intensity expressed with Sp statistics, which shows that
the contribution of vegetative propagation to SGS in the stand
averages 93 per cent (Table 2). These values are significantly
higher than those shown by Schueler et al. (2006), who obtained
22 per cent. Direct comparison of Sp values (Table 2: Schueler et al.
(2006) one ramet per clone Sp ¼ 0.009, all ramets Sp ¼ 0.012)
shows that the SGS intensity in our population is slightly and 12
times higher, respectively. Jolivet et al. (2011) arrived at similar
Sp values for trees of generative origin, particularly in stands
with lower density. In comparison to the study of Vaughan et al.
(2007a), in the case of one ramet per clone, our SGS intensity is
three-times lower than that in managed forest, and when all
ramets are included, our SGS exceeds that in unmanaged forest.
Somatic mutations
The most common cause of mutations at microsatellite loci is polymerase slippage during DNA replication (Eisen, 1999). In P. avium,
which often propagates vegetatively, such mutations can occur in
the early phase of meristem development, when the new plant
sprouts from adventitious buds on roots (D’Amato, 1997; Vezzulli
et al., 2012). There is a high probability that this phenomenon is
present in our stand. Analysis of eight microsatellite loci showed
that some individuals differ in just one of the eight loci (in one
case at two loci), mainly with just one allele difference (Table 3).
Such multilocus combinations can hardly be obtained in the
sexual cross of highly heterozygous individuals. The probability
that two randomly selected trees in our stand have the same genotype at seven (six) loci and not at one (two), is psex , 0.01. Although
we have not studied this in detail, we did not observe any phenotypic differences between mutant and non-mutant trees in the
field, most probably because mutations were found at microsatellite loci, which are considered as neutral markers.
Somatic mutations have frequently been reported for woody
species, which often propagate vegetatively in nature (Ally et al.,
2008; Chenault et al., 2011; Lian et al., 2004). However, despite
the high mutability of microsatellite loci (estimated frequency is
10 – 2 – 10 – 6 mutations per locus per generation, see Li et al.,
2002), simultaneous mutations of both alleles at the same locus
are very rare. A rough estimation for this could be the square of
the probability mentioned above. They have been confirmed for
R. pseudoacacia (Lian et al., 2004), and in the process of somatic
embryogenesis, also in Pinus pinaster (Marum et al., 2009). In our
case simultaneous mutations at loci EMPa015 and UDP98-412
Linear barriers and somatic mutation affect the genetic structure of a Prunus avium L. stand
were detected. However, despite the fact that we performed two
additional PCRs on ramets carrying potential mutations, we
cannot completely exclude the possibility of genotyping error,
nor the possibility that mutations did not occur at the same time
and that the tree harbouring only one mutated allele is no longer
present.
Somatic mutations of P. avium have previously been reported by
Vaughan et al. (2007b). They detected groups of trees differentiated at one of 13 SSR loci and none at S-loci. Their study confirmed
mutations at loci EMPaS06, EMPaS14, EMPa018, PceGA34, and, like
our study, at loci EMPa004 and EMPa015. For R. pseudoacacia Lian
et al. (2004) inter alia found that longer alleles tend to mutate into
shorter alleles, which supposedly inhibits the perpetual ‘growth’ of
microsatellites. Our results showed that mutations at ‘shorter’ loci
were bi-directional (deletions and insertions which lead to loosing
or gaining length), while all mutations at the ‘longer’ loci EMPa005
and EMPa015 were deletions. A similar mutation pattern was
shown in a microsatellite study of British cherries (Vaughan et al.,
2007b). However, for all different MLGs (67), heterozygosity
amounted to HO ¼ 0.744, which is only slightly higher than that
for MLLs (59), HO ¼ 0.735. Our results show slightly increased heterozygosity caused by somatic mutations, which has also been
confirmed by Vaughan et al. (2007b). In addition, if we include
mutated ramets into the MLLs, the average number of trees in
the MLLs increases from 3.2 to 3.7. The growing surface of the
MLLs remains the same, as the newly included trees are mostly
within the MLL area (Figure 2). Based on our results and the
results of other studies, we can conclude that somatic mutations
at SSR loci are relatively frequent, especially in the tree species
that easily propagate vegetatively. What is more, we believe that
they have often been overlooked and treated as an error, especially
in cases where non-exhaustive sampling strategies are employed
when there is not a complete overview of the genetic structure of
the samples.
Natural and artificial barriers to vegetative propagation
of P. avium in the stand
There are little data and few studies on the root system of P. avium.
Kutschera and Lichtenegger (2002) report that P. avium has a wide
root system and describe a specimen from South Tyrol, growing on
a southern slope on brown soil that had a horizontal root system
diameter of 24 m. The measurement involved a single tree and
makes it difficult to infer maximum root system sizes of P. avium
and the associated distances between individual root suckers in
one root system and one generation. In our study the distance
maxima between individuals of the same clone (dmax ) is 56 m. In
Germany, Schueler et al. (2006) confirmed a dmax of the same
genotype of 74 m, but in their analysis, they were treated as
single individuals. Gömöry (2004) reports about the same genotypes even at a distance of 240 m, but due to the poor accuracy
of the isozyme markers used, the author notes that this should
be taken with some reservations. However, it appears that the
established distance of 56 m between individuals in our case is
too large for one generation, i.e. to sprout from the root system
of the same tree.
Despite the relatively large spatial dimension, trees within MLLs
remain strongly spatially grouped (Figure 2), which to a certain
extent is due to streams and ditches that our study shows as efficient natural barriers to vegetative propagation via root suckers.
We recorded only one case where trees in the same group overcame a barrier and expanded to the other side of a ditch (see red
squares in centre of Figure 2). Forest paths are a weaker barrier to
vegetative propagation, particularly smaller and rarely used skidding trails and mud tracks. Yet these barriers may have been
bridged 50 or more years ago, perhaps before the creation of
such paths, when the paths were shallower and hence less effective in stopping vegetative propagation. Although root growth and
expansion was not examined in our study, it should be noted
that vegetative propagation via root suckers is strongly dependent
on the physical, chemical and biological properties of the soil (see
Kutschera and Lichtenegger, 2002). The findings of our study
should therefore aid further studies on different sites and with
different tree species.
In addition to vegetative propagation via root suckers, P. avium
tends to exhibit strong vegetative propagation from stumps
(Russell, 2003), but our study did not detect this due to the
‘absence’ of cutting. Vegetative propagation through buds on
fallen trees has not yet been reported. Although we did not explicitly evaluate the possibility of clonal establishment originating from
downstream movement of small branches, the results of the
genetic analysis and spatial distribution of ramets show that this
phenomenon did not occur in our study area. In addition, many
lying trees were detected across both streams but no rooted
shoots were found on lying logs, and as our results show, such
events have also not been the case in the past; we did not find a
single tree of the same genotype growing on both banks of a
stream.
Conclusion
Prunus avium is a fast growing species with a relatively short production period and highly valuable and desirable wood. Increasing
its share in forests is strongly promoted. Our research shows that
more than two-thirds of the analyzed trees in the stand have a
MLG identical to at least one other tree in the stand, indicating
vegetative origin. The consequence of the spatial proximity of identical genotypes is non-random spatial distribution, which results in
significant SGS. Trees of the same clone are strongly spatially
grouped, and additionally, grouping is affected by forest paths,
streams and ditches, which represent obstacles to root growth
and consequently obstruct vegetative propagation via root
suckers. Rare somatic mutations that inter alia produce completely
new alleles previously not present in the stand also contribute to
the complexity of the genetic structure. All of these findings
(except somatic mutations) indicate that caution is necessary in
P. avium management when determining seed stands for seed production as well as in the protection of its genetic diversity.
Acknowledgements
We thank three anonymous reviewers, Berthold Heinze and Gary Kerr for
providing helpful critiques that have greatly improved the paper.
Conflict of interest statement
None declared.
619
Forestry
Funding
This work was supported by the Slovenian Research Agency (P4-0059 and
V4-1438) and the Pahernik Foundation.
References
Ally, D., Ritland, K. and Otto, P. 2008 Can clone size serve as a proxy for clone
age? An exploration using microsatellite divergence in Populus tremuloides.
Mol. Ecol. 17, 4897– 4911.
Ganopoulos, I.V., Aravanopoulos, F.A. and Tsaftaris, A. 2013 Genetic
differentiation and gene flow between wild and cultivated Prunus avium:
An analysis of molecular genetic evidence at a regional scale. Plant
Biosyst. 147, 678– 685.
Garcia, C., Arroyo, M., Godoy, A. and Jordano, P. 2005 Mating patterns, pollen
dispersal, and the ecological maternal neighbourhood in a Prunus mahaleb
L. population. Mol. Ecol. 14, 1821 –1830.
Gömöry, D. 2004 Mutual links of demographic and genetic processes in a
wild cherry populations during the colonization. Biologia 59, 493–500.
Hamilton, M.B. 2009 Population Genetics. Wiley– Blackwell, 407 pp.
Ally, D., Ritland, K. and Otto, S.P. 2010 Aging in a long-lived clonal tree.
PLoS Biol. 8, e1000454. doi: 10.1371/journal.pbio.1000454
Hardy, O.J. and Vekemans, X. 1999 Isolation by distance in a continuous
population: reconciliation between spatial autocorrelation analysis and
population genetics models. Heredity 83, 145–154.
Anonymous 1997 Slovenija na vojaškem zemljevidu 1763–1787 (1804) ¼
Josephinische Landesaufnahme 1763–1787 (1804) für das Gebiet der
Republik Slowenien. Opisi 3. zvezek. Znanstvenoraziskovalni center SAZU,
Arhiv Republike Slovenije.
Hardy, O.J. and Vekemans, X. 2002 SPAGeDi: a versatile computer program
to analyse spatial genetic structure at the individual or population levels.
Mol. Ecol. Notes 2, 618– 620.
Arnaud-Haond, S. and Belkhir, K. 2007 GENCLONE: a computer program to
analyse genotypic data, test for clonality and describe spatial clonal
organization. Mol. Ecol. Notes 7, 15 –17.
Arnaud-Haond, S., Duarte, C.M., Alberto, F. and Serrao, A. 2007
Standardizing methods to address clonality in population studies. Mol.
Ecol. 16, 5115– 5139.
Asuka, Y., Tomaru, N., Nisimura, N., Tsumura, Y. and Yamamoto, S. 2004
Heterogeneous genetic structure in a Fagus crenata population in an
old-growth beech forest revealed by microsatellite markers. Mol. Ecol. 13,
1241– 1250.
Brus, R., Galien, U., Božič, G. and Jarni, K. 2010 Morphological study of the
leaves of two European black poplar (Populus nigra L.) populations in
Slovenia. Period. Biol. 112, 317– 325.
Chenault, N., Arnaud-Haond, S., Juteau, M., Valade, R., Almeida, J.L., Villar,
M. et al. 2011 SSR-based analysis of clonality, spatial genetic structure
and introgression from the Lombardy poplar into a natural population of
Populus nigra L. along the Loire River. Tree Genet. Genomes 7, 1249 –1262.
Helmersson, A., Jansson, G., Bozhkov, P.V. and Von Arnold, S. 2008 Genetic
variation in microsatellite stability of somatic embryo plants of Picea abies:
A case study using six unrelated full-sib families. Scand. J. For. Res. 23,
2 – 11.
Jarni, K., De Cuyper, B. and Brus, R. 2012 Genetic variability of wild cherry
(Prunus avium L.) seed stands in Slovenia as revealed by nuclear microsatellite loci. PLoS One 7, e41231. doi: 10.1371/journal.pone.0041231.
Jolivet, C. and Degen, B. 2011 Spatial genetic structure in wild cherry
(Prunus avium L.): II. Effect of density and clonal propagation on spatial
genetic structure based on simulation studies. Tree Genet. Genomes 7,
541 – 552.
Jolivet, C., Höltken, A.M., Liesebach, H., Steiner, W. and Degen, B. 2011
Spatial genetic structure in wild cherry (Prunus avium L.): I. variation
among natural populations of different density. Tree Genet. Genomes 7,
271–283.
Kutschera, L. and Lichtenegger, E. 2002 Wurzelatlas mitteleuropäischer
Waldbäume und Sträucher. 2nd edn. Stocker, 604 pp.
Clarke, J.B. and Tobutt, K.R. 2003 Development and characterization of
polymorphic microsatellites from Prunus avium ‘Napoleon’. Mol. Ecol.
Notes 3, 578– 580.
Li, Y.C., Korol, A.B., Fahima, T., Beiles, A. and Nevo, E. 2002 Microsatellites:
genomic distribution, putative function and mutational mechanisms: a
review. Mol. Ecol. 11, 2453– 2465.
Crespan, M. 2004 Evidence on the evolution of polymorphism of
microsatellite markers in varieties of Vitis vinifera L. Theor. Appl. Genet.
108, 231– 237.
Lian, C., Oishi, R., Miyashita, N. and Hogetsu, T. 2004 High somatic instability
of a microsatellite locus in a clonal tree, Robinia pseudoacacia. Theor. Appl.
Genet. 108, 836– 841.
D’Amato, F. 1997 Role of somatic mutations in the evolution of higher
plants. Caryologia 50, 1– 15.
Loiselle, B.A., Sork, V.L., Nason, J. and Graham, C. 1995 Spatial genetic
structure of a tropical understory shrub, Psychotria officinalis (Rubiaceae).
Am. J. Bot. 82, 1420– 1425.
Dorken, M.E. and Eckert, C.G. 2001 Severely reduced sexual reproduction in
northern populations of clonal plant, Decodon verticillatus (Lythraceae).
J. Ecol. 89, 339– 350.
Ducci, F. and Santi, F. 1997 The distribution of clones in managed and
unmanaged populations of wild cherry (Prunus avium). Can. J. Forest Res.
27, 1998 –2004.
Eisen, J.A. 1999 Mechanistic basis for microsatellite instability. In:
Microsatellites. Evolution and Applications. Golstein, D.B. and Schlötterer,
C. (eds). Oxford University Press, pp. 34 –48.
Ellegren, H. 2004 Microsatellites: simple sequences with complex evolution.
Genetics 5, 435–445.
Excoffier, L., Laval, G. and Schneider, S. 2005 Arlequin ver. 3.0: an integrated
software package for population genetics data analysis. Evol. Bioinformatics
Online 1, 47– 50.
Ganopoulos, I., Aravanopoulos, F.A., Argiriou, A., Kalivas, A. and Tsaftaris, A.
2012 Genome and population dynamics under selection and neutrality: an
example of S-allele diversity in wild cherry (Prunus avium L.). Tree Genet.
Genomes 8, 1181– 1190.
620
Lopes, T., Capelo, A., Brito, G., Loureiro, J. and Santos, C. 2009 Genetic
variability analyses of the somatic embryogenesis induction process in
Olea spp. using nuclear microsatellites. Trees 23, 29 –36.
Marum, L., Rocheta, M., Maroco, J., Oliveira, M.M. and Miguel, C. 2009
Analysis of genetic stability at SSR loci during somatic embryogenesis in
maritime pine (Pinus pinaster). Plant. Cell Rep. 28, 673–682.
O’Connell, L.M. and Ritland, K. 2004 Somatic mutations at microsatellite loci
in western redcedar (Thuja plicata: Cupressaceae). J. Hered. 95, 172–176.
Oddou–Muratorio, S., Demesure– Musch, B., Pélissier, R. and Gouyon, P.H.
2004 Impacts of gene flow and logging history on the local genetic
structure of a scattered tree species, Sorbus torminalis L. Crantz. Mol. Ecol.
13, 3689 –3702.
Parks, J.C. and Werth, C.R. 1993 A study of spatial features of clones in a
population of Bracken fern, Pteridium aquilinum (Dennstaedtiaceae).
Am. J. Bot. 80, 537–544.
Pelsy, F. 2010 Molecular and cellular mechanisms of diversity within
grapevine varieties. Heredity 104, 331– 340.
Linear barriers and somatic mutation affect the genetic structure of a Prunus avium L. stand
Rousset, F. 1997 Genetic differentiation and estimation of gene
flow from F – statistics under isolation by distance. Genetics 145,
1219 – 1228.
Vaughan, S.P. and Russell, K. 2004 Characterization of novel microsatellites
and development of multiplex PCR for large-scale population studies in wild
cherry, Prunus avium. Mol. Ecol. Notes 4, 429– 431.
Rousset, F. 2000 Genetic differentiation between individuals. J. Evolution
Biol. 13, 58 –62.
Vaughan, S.P., Cottrell, J.E., Moodley, D.J., Connolly, T. and Russell, K. 2007a
Distribution and fine-scale spatial-genetic structure in British wild cherry
(Prunus avium L.). Heredity 98, 274– 283.
Russell, K. 2003 EUFORGEN Technical Guidelines for Genetic Conservation
and Use for Wild Cherry (Prunus avium). International Plant Genetic
Resources Institute, 6 pp.
Schueler, S., Tusch, A. and Scholz, F. 2006 Comparative analysis of the
within-population genetic structure in wild cherry (Prunus avium L.) at the
self-incompatibility locus and nuclear microsatellites. Mol. Ecol. 15,
3231– 3243.
Vaughan, S.P., Cottrell, J.E., Moodley, D.J., Connolly, T. and Russell, K. 2007b
Clonal structure and recruitment in British wild cherry (Prunus avium L.).
Forest Ecol. Manag. 242, 419– 430.
Vekemans, X. and Hardy, O.J. 2004 New insights from fine-scale spatial
genetic structure analyses in plant populations. Mol. Ecol. 13, 921–935.
Stokes, M.A. and Smiley, T.L. 1968 An Introduction to Tree-ring Dating.
University of Arizona Press, 73 pp.
Vezzulli, S., Leonardelli, L., Malossini, U., Stefanini, M., Velasco, R. and Moser,
C. 2012 Pinot blanc and Pinot gris arose as independent somatic mutations
of Pinot noir. J. Exp. Biol. 63, 6359–6369.
Van Oosterhout, C., Hutchinson, W.F., Wills, D.P.M. and Shipley, P. 2004
MICRO-CHECKER: software for identifying and correcting genotyping
errors in microsatellite data. Mol. Ecol. Notes 4, 535–538.
Young, A.G., Hill, J.H., Murray, B.G. and Peakall, R. 2002 Breeding system,
genetic diversity and clonal structure in the sub-alpine forb Rutidosis
leiolepis F. Muell. (Asteraceae). Biol. Conserv. 106, 71 –78.
621