liesebach 2008

8/13/2019 Liesebach 2008

1/10

Heike Liesebach, Zoltn Sink

A contribution to the systematics of the genusTilia with respect to some hybrids by RAPDanalysis

Received: 02 January 2008, Accepted: 28 March 2008

Abstract:The putative hybrid character of two Tilia varieties selected as avenuetrees (Szent IstvanandK3)should be clarified for further breeding activities. Their ancestor species should be identified by a geneticcomparison with reference material (Tilia cordata, T. platyphyllos, T. dasystyla, T.euchlora, T.europaeaandothers). RAPD marker data were evaluated by UPGMA cluster and neighbour-joining analysis. The genetic

comparison of the two selected clones with the collection of reference material offers insights into some sys-tematic relationships withinthe genus Tilia. It was supplemented by a critical discussion of methods of RAPDdata evaluation for taxonomic purposes.

Additional key words:TiliaL.; linden; taxonomy; hybrids; DNA; UPGMA cluster analysis; neighbour-join-ing; resampling

Addresses:H. Liesebach, Institute of Forest Genetics, Johann Heinrich von Thnen-Institute (vTI), FederalResearch Institute for Rural Areas, Forestry and Fisheries, Eberswalder Chaussee 3A, 15377 Waldsievers-dorf, Germany, e-mail: [email protected]. Sink, Nursery Prenor, Bke tr 1, 9707 Szombathely-Herny, Hungary

IntroductionTrees and shrubs of the genusTiliaL. (Malvaceae)

were found in the forests of the northern hemispherein Europe, Asia and Eastern North America and Cen-tral America. They are growing in temperate, sub-tropical and tropical climates and occur from moist todry regions(Muir 1984).

Representatives of these deciduous trees becameincreasingly important in municipalparks of large cit-ies and play a central role as avenuetrees (Sukopp andWurzel 2000). Growing damages of lime-trees wereobserved in European citiesbecause of theubiquitous

stress factors as salt and drought [Helsinki: Terhoand Hallaksela (2005), Warsaw: Chmielewski et al.(1998), Budapest: Sink (unpublished)]. Former

plantings were undertaken more for ornamental rea-sons. Since the 1960s, T. cordata Miller and T.platyphyllos Scop. were increasingly replaced by T.tomentosa Moench and T. euchlora K. Koch fromsoutheast Europe and the Caucasian region becauseof their better adequacy to urban conditions in Cen-tral Europe. Actually the breeding efforts focus onbetter-adapted plant material with high drought andsalt resistance and more tolerance to soil compactionfor urban locations (Karnosky et al. 1982).

In Budapest, several trees from an avenue withpositive traits in stress resistance were selected(Jmborn et al. 2001). They are about 30 years old

and originate from seedling offsprings after free polli-nation. They are putative hybrids but their parentscould notbe reconstructed. Presumed ancestorsseem

2008, vol. 59, 1322

8/13/2019 Liesebach 2008

2/10

14 Heike Liesebach, Zoltn Sink

to originate from the group of European and Cauca-sian lime-trees. One selected variety (Szent Istvn)develops morphological characteristics similar to T.dasystyla Steven or T. euchlora. The other variety

(K3) is more similar toT. platyphyllos. However, theirmorphological characteristics of twigs, leaves, flowersand fruits did not allow a doubtless taxonomic classifi-cation to fulfil the standards for traded ornamentalplants (Sink 2005). The morphological findings ofthe selected clones of interest should now be com-pleted by genetic data to get more information on theirtaxonomic status for further breeding. Thissuggestionbases on the hypothesis that phylogenetic relation-ships could be reflected by suitable genetic markersand compared to adequate reference material. Severalmethods of data evaluation should be critically re-

viewed to reconstruct the evolutionary relationships ofthe investigated group of organisms.Only some genetic studies in the genus Tilia(T.

cordata and T. platyphyllos) were carried out byisozyme markers (Maurer and Tabel 1995; Frommand Hattemer 2003). Isozyme markers in the contextof population genetic studies of outcrossing speciesare known to possess a high level of within popula-tion variation. An additionaldifficulty forcodominantnuclear markers is the polyploidisation in the phylo-geny of these lindens (Fromm and Hattemer 2003).However, sufficient population based material fromthe Caucasian species group could not be providedand thus isozyme markers did not come into ques-tion. Nuclear microsatellite markers forTiliaspeciesare not available so far. Even if they existed, theywould not be suitable for our study for the same rea-son like isozyme markers.

High variation in chloroplast DNA haplotypes wasdetected for European T. cordata populations (Fi-neschi et al. 2003). However, chloroplasts weremostly maternal inherited in Angiosperms (Harrisand Ingram 1991; Rajoraand Dancik 1992) and there-fore did not reveal taxonomic relationships in case ofpossible hybridisation between closely related spe-

cies. Secondary, it is known that several related spe-cies share their chloroplast genotypes within the ge-nus [i.e. Tilia (Brunner et al. 2001), Quercus L.(Dumolin-Lapegue et al. 1997),Betula L. (Palme et al.2004), Hedera L. (Grivet and Petit 2002)]. Thus,chloroplast markers are not appropriate for our sub-ject.

Markers widely distributed within the genomewould be the method of choice for visualisation ofphylogenetic relationships and to get information onthe origin of putative hybrids, even if the number ofreference accessions is low: species-independent

RAPD (Random Amplified Polymorphic DNA) andAFLP markers (Amplified Fragment Length Polymor-phism). Due to the fact that AFLPs need an extensivelaboratory input, the risk of somewhat lower

reproducibility was accepted. For the application ofRAPD markers no molecular information is neededfor short, arbitrary primers. As a fingerprintingmethod these markers are well established for clonal

and cultivar identification of numerous plant species.Representative examples amongst others are givenfor poplar, melon and Citrus (Rajora and Rahman2003; Tanaka et al. 2007; Rao et al. 2007). By applica-tion of a sufficient number of primers and evaluablebands, RAPDs reveal good estimations of geneticsimilarities and degree of relationship.

Methods

Plant material

The putative hybrid clone Szent Istvn (No. 1 in

Table 1) and the clone K3 (No. 2), which is veryclose toT. platyphyllosregarding morphological traits,are selected clones for urban locations with a prob-lematic taxonomic status. A collection of plant mate-rial fromthe genus Tilia was compiled to cover the pu-tative relatives, which were assumed among the Eu-ropean and Caucasian group ofTiliaspecies based onmorphological observations (T. dasystyla,T. dasystylaSteven subsp. caucasica (V. Engl.) Pigott and T. euchlora). T. euchlora is a putative hybrid ofT. cordataandT. dasystyla.

A number of individuals of available species were

included to estimate within and among populationvariation at the species level as far as possible (T.cordataandT. platyphyllos). The specimens No. 26 andNo. 27 are labelled as T. dasystyla in the Botanical Gar-den Berlin-Dahlem, but they were identified as T.cordataby morphological traits (Sink unpublished).Two other species with only 3 individuals were in-cluded into the experiment: T. tomentosa and T.amurensisRupr.

Furthermore some selected clones were involved:T. cordataWega,T. tomentosaSzeleste,T. europaeaL. Pallida, and one unknown Tilia specimen from theSpth Arboretum in Berlin.T. europaeais possibly ahybrid ofT. cordataandT. platyphyllos.

In addition, species far distant from the putativerelatives of the interesting clones were integrated asoutgroups to test the reliability of the method (Amer-ica:T. americanaL., Asia:T. henryanaSzyszyll.).

RAPD analysis

Twigs were harvested in February and March 2004.They were cultivated in greenhouse up to the devel-opment of first young leaves. Leaves were frozen inliquid nitrogen and homogenised. DNA was isolatedwith thehelp of DNeasy Plant Mini Kit (QIAGEN) ac-

cording to manufactures instructions. The samples of2050 mg leaf tissue yielded in respectively 100250ng DNA estimated by photometric assay.

8/13/2019 Liesebach 2008

3/10

A contribution to the systematics of the genusTiliawith respect to some hybrids by RAPD analysis 15

A standard protocol was used to carry out theRAPD analysis. A total of 36 arbitrary primers wereused for the amplification (Table 2). Each PCR (Poly-merase Chain Reaction) was performed in 25 l reac-

tion volume containing 100250 pg template DNA,dNTP mix (200 M each nucleotide), 0.4 M primer,reaction buffer (including MgCl2to a final concentra-tion of 2.5 mM) and 1 Unit Taq polymerase (Amer-sham Pharmacia Biotech, NJ, USA).

The amplification reactions were carried out in athermocycler (Biometra GmbH, Gttingen, Germa-ny)with the followingprogram: Initial denaturation5

minutes at 95C; 30 cycles with 30 sec 95C denatur-ation, 30 sec 35C annealing, 1 min 72C elongation;final extension 7 min 72C; hold 4C.

PCR products were separated in 1.8% agarose gel

by electrophoresis and stained with ethidium bro-mide. The gels were photographed using the KodakEdas 290 System.

A replication of DNA extraction was carried outwith tissues from buds and cambium in one case andreplications of DNA amplification were performedwith 17 relevant primers and a subset of samples toconfirm reproducibility.

Table 1. Material collection of genusTiliaincluding selected clones

No. Species and/or cultivar Abbreviation Location in Germany Age (years) Place of origin

1 T.Szent Istvn T.SzIs Berlin, Humboldt University 3 Hungary, Budapest2 T.K3 T.plaK3 Berlin, Humboldt University 3 Hungary, Budapest

3 T. cordataWega T.cWega Berlin, Humboldt University 8 Commercial nursery clone

4 T. tomentosaSzeleste T.tSzel Berlin, Humboldt University 8 Hungary, Szeleste

5 T. europaeaPallida T.xvuPal Berlin, Humboldt University 15 Commercial nursery clone

6 T. euchlora T.xeuc1 Berlin, Botanical Garden Dahlem 29 Unknown

7 T. euchlora T.xeuc2 Berlin, Arboretum Spth ~ 70 Unknown

8 T. euchlora T.xeuc3 Berlin, Humboldt University 15 Commercial nursery clone

9 T. dasystyla T.das3 Berlin, Arboretum Spth 103 Russia, Novospasskoye

10 T. dasystyla caucasica T.cauc Tharandt, Forest botanical garden ~ 25 Iran, Aschchabad

11 T. platyphyllos T.plat1 Gttingen, Forest botanical garden ~ 130 Germany, Hxter

12 T. platyphyllos T.plat2 Gttingen, Forest botanical garden ~ 100 Germany, Hxter13 T. platyphyllos T.plat3 Gttingen, Forest botanical garden ~ 100 Germany, Hxter

14 T. platyphyllos T.plat4 Mixed forest 80100 Germany, Pritzhagen





19 T. cordata T.cor1 Berlin, Botanical Garden Dahlem 1 France

20 T. cordata T.cor2 Berlin, Botanical Garden Dahlem 22 Russia, Zhiguli

21 T. cordata T.cor3 Mixed forest 80100 Germany, Pritzhagen





26 T. cordata T.das1 Berlin, Botanical Garden Dahlem ~ 25 Caucasus

27 T. cordata T.das2 Berlin, Botanical Garden Dahlem ~ 25 Caucasus

28 T. amurensis var. amurensis T.amu1 Tharandt, Forest botanical garden ~ 30 Kasachstan, Karaganda

29 T. amurensis var. amurensis T.amu2 Tharandt, Forest botanical garden ~ 30 Russia, Moscow Pan-Kudrya-shov Botanical Garden

30 T. amurensis T.amu3 Tharandt, Forest botanical garden ~ 30 Georgia, Laghodeki (Caucasus)

31 T. tomentosa T.tom1 Berlin, Botanical Garden Dahlem 18 Hungary, Villny-Sikls



34 T. americana T.amer Berlin, Botanical Garden Dahlem ~ 50 USA

35 T. henryana var. subglabra T.henr Berlin, Botanical Garden Dahlem 18 China

36 T. sp. T. ??? Berlin, Arboretum Spth ~ 50 Unknown

8/13/2019 Liesebach 2008

4/10


Data evaluation

Clear and well resolved fragments were scoredmanually to construct a zeroand one matrix of allwell

resolved bands, where one represented the pres-ence of a certain band and zero their absence. Thepairwise DICE coefficient was calculated (similarity= 2a/(2a + b + c)), whereas a is the number ofmatches and b and c are the numbers of mis-matches comparing two specimens. The correspond-ing distance matrix (distance = 1 similarity) wasused to execute a hierarchical cluster analyses(UPGMA, unweighted pair group method usingarithmetic averages) and as an alternative the neigh-bour-joining (NJ) analysis for construction ofdendrograms. The UPGMA clustering method pro-duces branch lengths proportional to genetic dis-tances, whereas the neighbour-joining method mini-mises the total length of the tree.

Several resampling methods for the RAPD frag-ments as well as for theTiliataxa were carried out totest the robustness of the trees (program package

FreeTree by Pavlicek et al. (1999), N=1000replicateddatasets for all tests). The software TreeView (Page1996) was used for graphical displays of NJ trees. Areal scaled dendrogram of UPGMA cluster analysiswas built with the SAS statistic package (SAS Insti-tute Inc. 2003) using proc distance, proc iml, proccluster and proc tree.

Results

Out of the 36 random primers tested, the amplifi-cations with 17 primers resulted in suitable and re-producible banding patterns. The remaining 19 prim-ers did not produce PCR products or reliable resolvedbands. The data evaluation for banding patterns of 17primers yielded in 403 polymorphic bands that were

scored in a zero and one matrix.The fragment lengthsranged between 230 bp and 2200 bp, and the meannumberof bands persuccessful primerequals to 23.7.An example for RAPD banding patterns is given forprimer L 12 in Figure 1.

The DICE similarity coefficients were used to cal-culate the corresponding distance matrix. As ex-pected, maximum distances occur between the twooutgroup specimensT. americanaandT. henryanato allotherTiliaobjects (distances 0.75 0.88). The mini-mum distances (0.01 0.09) were observed betweenthe threeT.euchloraspecimens. All other distances

range between these extremes.The clones of special interest Szent Istvn andK3 should be regarded for their single pairwise dis-tances to the next relatives among the Tilia collection.

Table 2. Random primers used for RAPD analysis ofTiliaspecimens

Primer designation Sequence (53) Primer designation Sequence (53)

L 01 GGC ATG ACC T L 19 GAG TGG TGA C

L 02 TGG GCG TCA A L 20 TGG TGG ACC AL 03 CCA GCA GCT T OPA 03 AGT CAG CCA C

L 04 GAC TGC ACA C OPA 09 GGG TAA CGC C

L 05 ACG CAG GCA C OPA 13 CAG CAC CCA C

L 06 GAG GGA AGA G OPB 10 CTG CTG GGA C

L 07 AGG CGG GAA C OPD 11 AGC GCC ATT G

L 08 AGC AGG TGG A OPE 01 CCC AAG GTC C

L 09 TGC GAG AGT C OPE 02 GGT GCG GGA A

L 10 TGG GAG ATG G OPE 03 CCA GAT GCA C

L 11 ACG ATG AGC C OPE 05 TCA GGG AGG T

L 12 GGG CGG TAC T OPE 11 GAG TCT CAG G

L 13 ACC GCC TGC T OPE 16 GGT GAC TGT G

L 14 TGT ACA GGC T OPE 20 AAC GGT GAC C

L 15 AAG AGA GGG G OPG 14 GGA TGA GAC C

L 16 AGG TTG CAG G OPG 15 ACT GGG ACT C

L 17 AGC CTG AGC C OPK 20 GTG TCG CGA G

L 18 ACC ACC CAC C OPY 07 AGA GCC GTC A

Table 3. Minimum pairwise DICE distances ofTiliaclones ofinterest to referencematerial (AbbreviationsseeTable 1)

Szent Istvn K3

T.xeuc1 0.477 T.plat1 0.432

T.xeuc2 0.481 T.plat2 0.472

T.xeuc3 0.506 T.SztIs 0.522T.plaK3 0.522 T.plat3 0.533

T.das2 0.566 T.plat5 0.566

8/13/2019 Liesebach 2008

5/10

8/13/2019 Liesebach 2008

6/10


euchlora, the clone Szent Istvn,T. dasystylaand thehybrid T. europaea Pallida (maximum distance0.55).

Two clusters at the above-species level were ascer-

tained: One forT. cordataand T. amurensis(distance0.63) and the other for aT. platyphyllos T. dasystylacomplex (distance 0.65). ThisT. platyphyllos T. dasy-stylacomplex consists ofT. platyphyllos,T. dasystyla,T.dasystyla caucasica, the clone Szent Istvn and all hy-brids containingT. platyphyllosorT. dasystyla.

The speciesT. tomentosa,T. americanaand T. hen-ryana as well as the unknown specimen from theSpth Arboretum are singular in the dendrogramwithout a grouping.

The robustness of the above described UPGMAtree topology was evaluated with several resampling

methods offered in the package FreeTree and by com-paring with the NJ dendrogram. Firstly the rearrange-ment of taxa (modifying the input order) waschecked. No influence on the structure was observed,all nodes were confirmed with a percentage 100% inthe UPGMA as well in the NJ tree.

The comparison of UPGMA and NJ dendrogramconfirms the stability of all clusters at the specieslevel (Figs 2, 3):T. cordatawith 10 specimens (smalldifferences in arrangement ofT. cordataNo. 5, 6 and 7within population Pritzhagen), T. amurensis with 3specimens,T. platyphylloswith 9 specimen (small dif-ferences in arrangement ofT. platyphyllosNo. 7 and 8within population Pritzhagen)and T. tomentosa with 4specimens. Thehybridgroup consisting ofT. euchlo-ra, T. dasystyla, T. Szent Istvn and T. europaeaPallida is present in both dendrograms, but withchanging positions ofT.Szent Istvn andT. euro-paeaPallida. The most obvious difference betweenboth dendrograms was found above the species level.The hybrid group meets the T. cordata cluster in the NJtree just beforeT. amurensiscomes together withT.cordata, whereas the hybrid group is combined withthe T. platyphylloscluster in theUPGMAdendrogram.

To assess the stability of clustering a resampling of

Tiliaspecimen was carried out for the UPGMA as wellas the NJ tree. The method of jackknifing over individ-uals determines whether their skipping can change thetopology of others branches. The number of skippedtaxa was successive increased, starting from oneskipped taxon, up to no further change of jackknifingpercentages was observed. This was the case for n=6skipped taxa, even though the authors of the softwarerecommended only 1 to 3 depending from the totalnumber. The fusion between theT. platyphyllosgroupand the hybrid group is confirmed in the UPGMAdendrogram with 100% (Fig. 2). In contrast, the asso-

ciation of the hybrid group to theT. cordatacluster inthe NJ tree is very weak with only 12% (Fig. 3).Additionally, two methods of resampling the

RAPD fragments were applied. The common used

bootstrapping (sampling with replacement) producesrepeated data sets that contain some of the RAPDfragments more than one times whereas other RAPDfragments are not present. The jackknifing of RAPD

fragments randomly drops a part of the fragments andretains the other part, commonly 50% (202 out of 403RAPD fragments). For each dendrogram the resultsof both resampling methods arevery similar (data notshown). However, remarkabledifferences werefoundbetween the UPGMA and NJ dendrograms. The fu-sion of theT. platyphylloscluster and the hybrid groupis highly supported with 68 resp. 70% (bootstrap-ping, jackknifing) in the UPGMA dendrogram. TheNJ tree supports the association of the hybrid groupto theT. cordatacluster with only 6%. Thus, abovespecies level the results of the NJ tree construction

method are not reliable within the genusTilia.

Discussion

The RAPD technique could be applied to any ob-jects without prior information on DNA sequences.This relatively simple method with short arbitraryprimers could involve sometimes lower reproduci-bility of banding patterns compared to other PCRbased techniques with longer and more specific prim-ers. The major disadvantage, however, is the commondominant nature of the data obtained (Nybom 2004).This disadvantage is rather important in populationgenetic studies than in a taxonomic investigation.

Several possibilities are available for calculation ofdistance matrices from binary datasets. Comparisonsof several similarity coefficients were reported e.g. byDuarte et al. (1999) and by da Silva Meyer et al.(2004). They concordantly suggest the use of coeffi-cients that do not include negative co-occurrences,because the negative co-occurrences do not necessar-ily mean that the regions of the DNA are identical.The reported distance levels of clustering within pop-ulations (0.18 0.22), at the species level (0.40 0.55) or between related species (0.63 0.65) are a

good consensus for the populations and species inthis study, but they are not to generalise. They de-pend on thesimilarity coefficient, on thecollectedref-erence material and on the markers used. They can-not be carried forward to other subjects.

Starting from the DICE similarity matrix two me-thods of tree construction were tested: the clusteranalysis method UPGMA and the neighbour-joiningmethod(Saitou and Nei 1987). Their respective algo-rithms are responsible for their different advantagesand disadvantages. The UPGMA method is directlyusing the genetic distances for clustering and to scale

the branch lengths. For this reason it is only suitablefor datasets consisting of objects with relatively con-stant rates of evolution. The NJ method works with-out such a precondition, therefore it is adequate to

8/13/2019 Liesebach 2008

7/10


phylogenetic studies covering higher taxonomic lev-els (Mohammadi and Prasanna 2003). The NJ algo-rithm minimises the total length of the tree. That im-plies possible changes of tree topology, if OTUs (Op-erational Taxonomic Units) were added or dropped.An example of such a change is given by Nei (1996)for vertebrates.

Regarding theTiliamaterial in this study, concor-dant sub-trees at the population and species levelwere found with both tree constructionmethods. Yet,the high sensitivity of the NJ method against the skip-ping of taxa (OTU based jackknifing) leads to sub-stantial doubts. In contrast, UPGMA regarding thereal genetic distances is much more stable against theskipping of taxa. The aspect of stability of tree topol-ogy against the random omitting of taxa seems to be

very important due to the fact that the collected refer-ence material might be incomplete. The collected ref-erence material was selected with respect to the tar-get of the study; nevertheless, it was depended on theavailability in botanical gardens. Since there is no rea-son to assume largely different rates of evolutionwithin the genusTilia, the UPGMA tree is the morereliable tree in this case and should be considered asthe basis for conclusions.

The UPGMA dendrogram as a whole is in a goodagreement with the taxonomy of theTiliagenus. Re-garding the clustering above species level, T. platy-phyllos and T. dasystyla are belonging to the subsectionTrabeeculares and T. cordata and T. amurensis aregrouped together in the subsection Reticulares (Muir1984). These both subsections form the section

Fig. 2. Dendrogram of UPGMA clusteranalysis based on DICE similarities of 36 Tilia specimens (Abbreviationsee Table 1),numbers indicating jackknife values over theTiliaobjects (percentage of n=1000 replicated datasets)

8/13/2019 Liesebach 2008

8/10


Anastraea, whereasT. tomentosa, T. henryanaandT.americanaare members of the section Astrophilyra(Muir 1984).

The most problematic part of the dendrogram isthe hybrid group consisting ofT. euchlora,T. dasy-styla,T.Szent Istvn andT. europaeaPallida.

The cloneT. europaeaPallida has been regarded

as a hybrid ofT. platyphyllosandT. cordata. Its seed-lings are very variable. Partially they develop morpho-logical traits asT. platyphyllos, but never asT. cordata.Pigott & Sell (1995) concluded from their observa-

tions thatT. platyphyllosis one parent, but there is noevidence forT. cordataas the other parent.

T. euchlora, commonly accepted as a possibly hy-brid betweenT. cordataandT. dasystyla, exists only inculture. Another source (Wikipedia, http://en.wiki-pedia.org/wiki/Tilia, unfortunately no reference gi-ven) listedT. euchloraas the hybrid ofT. dasystyla

andT. platyphyllos.T. euchloraseems to be one com-mercial clone with numerous ramets in Europe. Itsseeds do not germinate (Muir 1984). One putativeancestorT. cordata(orT. platyphyllos) is a diploid spe-

Fig. 3. Dendrogram of NJ analysis based on DICE similarities of 36Tiliaspecimens (Abbreviation see Table 1), numbers in-dicating jackknife values over theTiliaobjects (percentage of n=1000 replicated datasets)

8/13/2019 Liesebach 2008

9/10


cies with 2n=82 chromosomes (Uhrikova andSchwarzova 1980; Pigott 2002), whereas the otherancestor T. dasystyla is a tetraploid species with4n=164 chromosomes (Pigott and Francis 1999).

The hybrid T. euchlorais probably triploid and itssterility could be explained from the ploidy level.

Rametsof clonesshouldbe genetically identicalperse. However, the small genetic distances based onRAPD markers between the 3 specimens in this studycould not be regarded as definitive arguments for oragainst a single clone in this special case.

It is alleged that RAPD patterns are not alwayscomplete reproducible from laboratory operationcauses (Nybom 2004). This might be one cause forourresults. Butadditionally twoother facts shouldberegarded. One reason comes from the real existing

ecosystem tree. Numerous microorganisms likebacteria and fungi are located withinplant tissues, theso-called endophytes. These endophyte spectra coulddiffer in dependence from age, substrate and otherenvironmental factors as it was detected for a singlepoplar clone (Ulrich et al. 2007). Thus, short arbi-trary primers could amplify DNA from the tree, but aswell from endophytes. The other fact is, that there isno observation up to now how variable RAPD pat-terns of a definitive clone could become after at least140 years of vegetative propagation, when somaticmutations proceeded within this period.

Perhaps both hybrids,T. euchloraandT. euro-paeaPallida, do not descend fromT. cordata. Thisperspective is in accordance with the cluster analysisin our study. The taxon name T. euchlora is occupiedfrom the specimen firstly described by K. Koch 1866and its vegetative offspring. It is assumed that theclone Szent Istvn is a sister ofT. euchlora. The pu-tative parents could be T. dasystylaand T. platyphyllos.

The other clone of interest K3 is closest associ-ated toT. platyphyllosas it was shown by the singlepairwise distances to the next relatives among theTiliacollection. The selected clone K3 meets the T.platyphylloscluster in the UPGMA dendrogram at a

distance of 0.54. This is just in range of the assumedwithin-species variation. Nevertheless, a small con-tribution ofT. dasystylawithin the pedigree should benot excluded, because of the genetic similarity to T.Szent Istvn (see table 3) and some morphologicaltraits.

Further breeding effort for Tilia avenue treesshould focus on crossings betweenT. dasystylaas oneparent andT. cordataand T. platyphyllosas the otherparent. Offspring material from both hybrid familiesshould be tested for stress resistance traits and fortheir adequacy to urban environments. Furthermore,

the ability for vegetative propagation should includeinto the selection criterions.Parents and offspring material from these both hy-

brid families could be a good addition to the reference

material for further taxonomic investigation at theRAPD marker level and a better characterisation ofthe present selections Szent Istvn and K3. More-over, a perspective for more specific DNA markers

than RAPDs forspecies identification could be the de-velopment of SCAR markers (Sequence Character-ized Amplified Region) based on single specificRAPD bands (Scheef et al. 2003). This would be astep to more reliable results with lower number ofmarkers, but the suitability of SCAR markers to iden-tify hybrids is not self-evident and should be tested inevery special case.

Acknowledgements

The authors would like to acknowledge the sup-

port of all colleagues from the Spth Arboretum inBerlin, from the Botanical Garden in Berlin-Dahlem,from the Forest Botanical Garden in Gttingen andfrom the Forest Botanical Garden in Tharandt whocontributed to the collection of reference material.For her excellent laboratory guidance we are verygrateful to Ms Elke Ewald. This study is a part of thePhD thesis of Zoltn Sink at the Berlin HumboldtUniversity supervised by Prof. Hans-Heinrich Jesch(2005).

References

Brunner I.,Brodbeck S.,Bchler U.,Sperisen C. 2001.Molecular identification of fine roots of treesfrom the Alps: reliable and fast DNA extractionand PCR-RFLP analyses of plastid DNA. Molecu-lar Ecology 10: 20792087.

Chmielewski W., Dmuchowski W., Suplat S. 1998.Impact of urban environmental pollution ongrowth, leaf damage, and chemical constituentsof Warsaw urban trees. General Technical Report Pacific Southwest Research Station, USDA For-est Service No. PSW-GTR-166: 215219.

Duarte J.M., dos Santos J.B., Melo L.C. 1999. Com-

parison of similarity coefficients based on RAPDmarkers in the common bean. Genetics and Mo-lecular Biology 22: 427432.

Dumolin-LapegueS., Demesure B., Fineschi S., CorreV.L., Petit R.J. 1997. Phylogeographic Structureof White Oaks Throughout the European Conti-nent. Genetics 146: 14751487.

Fineschi S., Salvini D., Taurchini D., Carnevale S.,Vendramin G.G. 2003. Chloroplast DNA varia-tion ofTilia cordata(Tiliaceae). Canadian Journalof Forest Research 33: 25032508.

Fromm M., Hattemer H.H. 2003. Inheritance of allo-

zymes and hybridization in two European Tiliaspecies. Heredity 91: 337344.GrivetD.,Petit R.J. 2002. Phylogeography of thecom-

monivy(Hedera sp.) in Europe: genetic differenti-

8/13/2019 Liesebach 2008

10/10


ation through space and time. Molecular Ecology11: 13511362.

Harris S.A., Ingram R. 1991. Chloroplast DNA andBiosystematics: The Effects of Intraspecific Di-

versity and Plastid Transmission. Taxon 40:393412.

Jmborn B.E., Sink Z., Ifju Z. 2001. The selectionand Cultivation ofTilia Clones Tolerating Pol-luted Urban Environment. International Journalof Horticultural Sciences 7: 8286.

Karnosky D.F., Gerhold H.D., Collins W.H. 1982.METRIA Projects on Species Trials and CultivarTesting. Journal of Arboriculture 8: 178181.

Maurer W.D., Tabel U. 1995. Conservation of geneticresources ofTiliatree species in Rheinland-Pfalz(Germany). In: Population genetics and genetic

conservation of forest trees. Baradat Ph., AdamsW.T., Mller-Starck G. (eds.) SPB Academic Pub-lishing bv, Amsterdam, The Netherlands, pp.421426.

Mohammadi S.A., Prasanna B.M. 2003. Analysis ofGenetic Diversity in Crop PlantsSalient Statis-tical Tools and Considerations. Crop Science 43:12351248.

Muir N. 1984. A survey of the genusTilia. Plantsman5: 206242.

Nei M. 1996. Phylogenetic analysis in molecular evo-lutionary genetics. Annual Review of Genetics30: 371403.

Nybom H. 2004. Comparison of different nuclearDNA markers for estimating intraspecific geneticdiversity in plants. Molecular Ecology 13:11431155.

Page R.D.M. 1996. TREEVIEW: An application todisplay phylogenetic trees on personal comput-ers. Computer Applications in the Biosciences12: 357358.

Palme A.E., Su Q., Palsson S., Lascoux M. 2004. Ex-tensive sharing of chloroplast haplotypes amongEuropean birches indicates hybridization amongBetula pendula, B. pubescensandB. nana. Molecular

Ecology 13: 167178.Pavlicek A., Hrda S., Flegr J. 1999. FreeTree Free-

ware Program for Construction of PhylogeneticTrees on the Basis of Distance Data and Boot-strap/Jackknife Analysis of the Tree Robustness.Application in the RAPD Analysis of GenusFrenkelia. Folia Biologica 45: 9597.

Pigott C.D. 2002. A review of chromosome numbersin the genusTilia(Tiliaceae). Edinburgh Journalof Botany 59: 239246.

Pigott C.D., Francis B. 1999. The taxonomic status ofTilia dasystyla in Crimea, Ukraine. Edinburgh

Journal of Botany 56: 161173.

Pigott C.D., Sell P. 1995. Nomenclature of the Euro-pean species ofTilia. I. Tilia europaea L. Kew Bulle-tin 50: 135139.

Rajora O.P., Dancik B.P. 1992. Chloroplast DNA in-

heritance in Populus. TAG Theoretical and Ap-plied Genetics 84: 280285.

Rajora O.P., Rahman M.H. 2003. Microsatellite DNAand RAPD fingerprinting, identification and ge-netic relationships of hybrid poplar (Populus canadensis) cultivars. TAG Theoretical and Ap-plied Genetics 106: 470477.

Rao M., Soneji J., Chen C.X., Choi Y., Huang S.,Gmitter F. 2007. Identification of zygotic andnucellar seedlings from citrus rootstock candi-dates using RAPD. Acta Horticulturae 738:255260.

Saitou N., Nei M. 1987. The neighbor-joining me-thod: a new method for reconstructing phylogen-etic trees. Molecular Biology and Evolution 4:406425.

SAS Institute Inc. 2003. SAS for Windows.Scheef E.A., Casler M.D., Jung G. 2003. Development

of Species-Specific SCAR Markers in Bentgrass.Crop Science 43: 345349.

da Silva Meyer A., Franco Garcia A.A., Pereira deSouza A., Lopes de Souza Jr. C. 2004. Compari-sonof similarity coefficientsused forcluster anal-ysis with dominant markers in maize (Zea maysL.). Genetics and Molecular Biology 27: 8391.

Sink Z. 2005. Untersuchungen zu genotypspezifi-schen Reaktionen und zur taxonomischen Zuge-hrigkeit bei der In-vitro-Regeneration vonstresstoleranten Sorten der Gattung Tilia, ein-schlielich in Ungarn selektierter Klone fr diesptere Verwendung im urbanen Raum. CullivierVerlag, Gttingen, 144 pp.

Sukopp H., Wurzel A. 2000. Changing climate andthe effects on vegetation in central European cit-ies. Arboricultural Journal 24: 257281.

Tanaka K., Nishitani A., Akashi Y., Sakata Y., NishidaH., Yoshino H., Kato K. 2007. Molecular charac-

terization of South andEast Asian melon, CucumismeloL., and the origin of Group Conomon var.makuwa and var. conomon revealed by RAPDanalysis. Euphytica 153: 233247.

Terho M., Hallaksela A.-M. 2005. Potential hazardcharacteristics of Tilia, Betula, and Acer trees re-moved in the Helsinki City Area during20012003. Urban Forestry & Urban Greening 3:113120.

Uhrikova A., Schwarzova T. 1980. Chromosomenumber reports LXIX. Taxon 29: 729730.

Ulrich K., Ulrich A., Ewald D. 2007. Diversity of

endophytic bacterial communities in poplargrown under field conditions. FEMS Microbiol-ogy Ecology 63: 169180.

liesebach 2008

Documents