Chloroplast DNA polymorphism reveals little geographical structure in Castanea sativa Mill. (Fagaceae) throughout southern European countries

Molecular Ecology (2000)

9

, 1495–1503

© 2000 Blackwell Science Ltd

Blackwell Science, Ltd

Chloroplast DNA polymorphism reveals little geographical structure in

Castanea sativa

Mill. (Fagaceae) throughout southern European countries

S . F INESCHI , * D . TAURCHINI , * F. VILLANI* and G. G . VENDRAMIN†*

Istituto per l’Agroselvicoltura, Consiglio Nazionale delle Ricerche, via Marconi 2, I-05010 Porano, Italy,

†

Istituto Miglioramento Genetico Piante Forestali, Consiglio Nazionale delle Ricerche, via Atto Vannucci 13, I-50134 Firenze, Italy

Abstract

The distribution of haplotypic diversity of 38 European chestnut (

Castanea sativa

Mill.)populations was investigated by PCR/RFLP analysis of regions of the chloroplast andmitochondrial genomes in order to shed light on the history of this heavily managed species.The rapid expansion of chestnut starting from 3000 years ago is strongly related to humanactivities such as agricultural practice. This demonstrates the importance of human impact,which lasted some thousands of years, on the present-day distribution of the species. Nopolymorphism was detected for the single mitochondrial analysed region, while a total of11 different chloroplast (cp) haplotypes were scored. The distribution of the cpDNAhaplotypes revealed low geographical structure of the genetic diversity. The value ofpopulation subdivision, as measured by

G

STc

, is strikingly lower than in the other species ofthe family Fagaceae investigated. The actual distribution of haplotypic diversity may beexplained by the strong human impact on this species, particularly during the Roman civ-ilization of the continent, and to the long period of cultivation experienced during thelast thousand years.

Keywords

:

Castanea sativa

, cpDNA, PCR–RFLP, phylogeography, polymorphism

Received 6 December 1999; revision received 13 April 2000; accepted 8 May 2000

Introduction

Castanea sativa

Mill. is the only European species of thegenus

Castanea

; its distribution covers the southern partof the continent. The genus

Castanea

is supposed to haveoriginated in the Asian continent; it is believed that, duringthe Tertiary, eastward migration gave rise to the Americanchestnut,

Castanea dentata

(Marsch.) Borkh., and westwardmigration resulted in the European chestnut (Zohary &Hopf 1988), although according to these authors ourknowledge of the origin of

C. sativa

is still inadequate.This genus is represented by 13 species, native to thetemperate zones of Asia, southern Europe and easternUnited States.

European chestnut is widespread in all Mediterraneancountries. Its present natural range is considered to havebeen heavily influenced by human activities because

of the length of time for which it has been cultivated.The production of good-quality fruits made chestnut animportant source of food for many human populations,particularly in rural areas. For this reason, chestnut wasalready cultivated during Roman times. The Romans areheld responsible for the great diffusion of this speciesthroughout the Mediterranean areas, as well as in centralregions of the European continent, such as Germany andFrance (Huntley & Birks 1983).

The history of chestnut after the last ice period is notcompletely clear as yet, although it is likely to havesurvived in glacial refugia located in the Mediterraneanpeninsulas, like most European tree species (Bennet

et al

.1991; Hewitt 1996; Taberlet

et al

. 1998). In the Iberian andItalian peninsulas, chestnut pollen and wood charcoalhave been found in fossil layers correlated to the end ofthe middle Pleistocene and the beginning of the upperPleistocene (450 000–100 000 years ago) (Sánchez Goñi1993). Records of

Castanea

pollen are lacking for the lastglacial period (70 000–10 000 years ago, Sánchez Goñi 1993).

Correspondence: S. Fineschi. Fax: +39 0763374330; E-mail: [email protected]

MEC1029.fm Page 1495 Thursday, September 7, 2000 3:53 PM

1496

S . F I N E S C H I

E T A L .

© 2000 Blackwell Science Ltd,

Molecular Ecology

, 9, 1495–1503

According to the Huntley & Birks (1983) fossil pollenmaps, first post-glacial records appear 9000 years ago inSpain and Greece; however, the identification of

Castanea

pollen in Ereta del Pedregal, south-eastern Spain, has stillto be checked (Sánchez Goñi 1993). The first confirmedrecord of

Castanea

pollen in the Iberian peninsula is datedat 7000 years ago in As Lamas (Galicia, north-westernSpain), and in Lucenza (Galicia) at 4000 years ago (Santos

et al.

, unpublished data). Chestnut pollen appears for thefirst time in Italy 5500 years ago. A rapid expansion ofchestnut began by 2000 years ago, and by 1000 years ago itsdistribution was similar to the present-day one, includingFrance and southern Germany (Huntley & Birks 1983).

Pollination in the genus

Castanea

is described as anemo-philous, entomophilous, or a combination of both (Kubitzki

et al

. 1993; Harlow

et al

. 1996; Gellini & Grossoni 1997).Entomophilous trees are characterized by flowering afterthe appearance of leaves, the production of heavy pollen,and by scented flowers. All these traits are observed inEuropean chestnut; however, it has been hypothesizedthat the role played by insects may be passive and indir-ect, resulting in the dislodging of pollen from the anthers,but without visiting the female flowers (Harlow

et al

. 1996).The limited seed dispersal by weight, which is typical oflarge and heavy seeds (Harlow

et al

. 1996), is partiallybalanced by animals and humans who also contribute totheir diffusion.

The distribution of genetic diversity of wild and cul-tivated European chestnut has been intensively studied usingnuclear markers such as isozymes and random amplifiedpolymorphic DNA (RAPD) (Fineschi 1988; Pigliucci

et al

.1990; Fineschi & Malvolti 1991; Villani

et al

. 1991a,b; Fineschi

et al

. 1994a,b; Müller-Starck

et al

. 1994; Villani

et al

. 1994,1997, 1999). In particular, data obtained by Villani

et al

.(1997) on populations sampled throughout the Mediter-ranean countries showed: (i) a higher genetic diversity inTurkish and Spanish populations as compared to Italianand French ones, and (ii) a high level of genetic differenti-ation among eastern Turkish and European populations.

During the last decade, several studies have reportedresults on the chloroplast DNA (cpDNA) variation ina wide range of plants, including trees (for a review seeSoltis

et al

. 1992; Petit 1999). cpDNA polymorphism andits geographical distribution have been described in someEuropean species of the family Fagaceae: oaks, particu-larly

Quercus robur

L. and

Q. petraea

(Matt.) Liebl. (Ferris

et al

. 1993; Kremer & Petit 1993; Petit

et al

. 1993; Ferris

et al

. 1995; Dumolin-Lapègue

et al

. 1997a; Ferris

et al

.1998); beech,

Fagus sylvatica

L. (Demesure

et al

. 1996), andrecently one species from the austral hemisphere, thesouthern beech,

Nothofagus nervosa

(Phil.) Dim. Et Mil.(Marchelli

et al

. 1998).Cytoplasmic genomes, both chloroplasts and mitochon-

dria, are maternally inherited in most angiosperms; studies

on organelle DNA markers demonstrated that when seedflow is less efficient than pollen flow, as is the case inmany angiosperms, organelle polymorphisms are highlystructured in comparison with the nuclear ones (Petit

et al

. 1993). Phylogeography analysis assumes particularrelevance for the conservation of forest genetic resources;the history in the post-glacial period is said to have beenthe main factor in shaping the actual distribution of thediversity.

Chloroplast markers were used to: (i) analyse the distri-bution of haplotypic diversity in chestnut stands throughoutthe Southern-European distribution range; (ii) shed lighton the possible migration processes after the last glaci-ation; and (iii) understand the role of human impact onthe actual distribution of the cytoplasmic diversity.

Material and methods

Winter buds were collected from 38 populations, bothnatural and naturalized, from Turkey, Portugal, Spain,France and Italy (Table 1). According to Zohary & Hopf(1988) ‘the spontaneously growing chestnuts in Italy,southern France, Spain and adjacent places in western andcentral Europe do not represent genuinely wild populationsbut naturalized elements derived from the cultivatedchestnuts brought by man’, hence the description ‘natural-ized’. The sample size was five individuals per populationwhen possible, with three as minimum sample size. Treeswere sampled at a minimum distance of 100 m apart. Thetotal number of analysed individuals was 181. Details ofthe sampling and site locations are given in Table 1.

Total DNA was extracted from frozen material usingthe CTAB method described by Doyle & Doyle (1990) andmodified by Dumolin

et al

. (1995).cpDNA was amplified using the universal primers

described by Taberlet

et al

. (1991), Demesure

et al

. (1995) andDumolin-Lapègue

et al

. (1997b) (Table 2). Reactions werecarried out in a total volume of 25

µ

L consisting of about20 ng of template DNA, 2.1 m

m

of MgCl

2

, 100

µ

m

of eachdNTP, 0.2

µ

m

of each primer and 2.5 U of

Taq

polymerasewith the respective 1

×

PCR buffer (

Taq

polymerase and10

×

buffer purchased from Gibco BRL (Life Technologies).DNA amplification was performed in a Genius TechneDNA thermal cycler (Cambridge, UK). An initial 4 mindenaturation at 94

°

C was followed by 30 cycles of 94

°

Cfor 45 s, annealing at different temperatures dependingon the primers used for 45 s, and extension at 72

°

C for 2–4 min depending on the length of the fragments to beamplified. Amplification cycles were followed by a final10 min extension at 72

°

C. Details on the amplification foreach fragment are given in Table 2.

Aliquots (10

µ

L) of the amplification products weredigested with 5 U of each restriction enzyme (Table 2) forat least 3 h and at most 8–12 hours overnight.


CH

LO

RO

PL

AS

T D

NA

PO

LY

MO

RP

HIS

M IN

EU

RO

PE

AN

CH

ES

TN

UT

1497

© 2000 B

lackwell Science L

td,

Molecular Ecology

, 9, 1495–1503

Table 1

Details on sample size, location, country, and number of individuals per haplotype detected in the analysed

Castanea sativa

populations

Population number Country*

Population name

Location Longitude Latitude

Haplotypes

A B C D E F G H I J K Sample size

1 E Castanyet 2.63 41.89 0 0 0 0 0 0 1 0 4 0 0 52 E Valverde –6.13 40.23 0 0 0 0 3 2 0 0 0 0 0 53 E Medulas –6.09 42.45 0 0 0 1 0 0 0 0 4 0 0 54 E Viladrau 2.40 41.83 0 0 0 0 0 0 0 0 5 0 0 55 E Aracena –5.90 37.92 0 0 0 1 0 0 0 0 4 0 0 56 E Prades 0.99 41.34 0 0 0 1 3 0 0 0 0 1 0 57 E Navedo –5.10 43.05 0 0 0 0 0 0 0 0 5 0 0 58 E Anaz –3.08 43.36 0 0 0 0 0 0 0 1 4 0 0 59 E Navasfrias –6.16 40.29 0 0 0 0 0 0 0 0 5 0 0 5

10 E Hervas –5.20 40.26 0 1 0 0 2 0 0 0 2 0 0 511 P Guarda –6.60 40.54 0 0 0 0 0 0 0 0 5 0 0 512 P Bornes –6.92 41.54 0 0 0 0 0 0 0 0 5 0 0 513 P Braganca –6.10 41.80 0 2 0 0 0 0 0 0 3 0 0 514 P Vila Pouca –6.99 41.50 0 0 0 0 0 0 0 2 3 0 0 515 F Colognac 3.74 44.05 0 0 0 0 0 0 0 3 2 0 0 516 F Valmale 3.89 44.27 0 0 0 0 0 0 0 0 5 0 0 517 F Nojaret 3.89 44.32 0 0 0 0 0 0 0 0 5 0 0 518 I Triste 12.02 42.70 0 0 0 0 0 0 2 0 3 0 0 519 I S. Pietro 11.97 42.80 0 0 0 0 0 0 0 0 5 0 0 520 I Allumiere 11.87 42.15 1 1 0 1 0 0 0 0 0 2 0 521 I Cavriglia 11.48 43.52 0 2 0 0 0 2 0 0 0 0 1 522 I Loritto 10.33 46.14 0 0 0 0 0 0 0 0 0 0 5 523 I Borgo Chiesanuova 10.79 45.15 0 0 0 0 0 0 0 0 0 0 3 324 I Monreale 13.29 38.08 1 0 0 0 0 0 0 1 2 1 0 525 I Ucrìa 14.89 38.05 0 0 4 0 0 0 0 0 0 1 0 526 I Pitarrone 15.07 37.80 0 0 0 0 0 0 0 0 5 0 0 527 I Castellaro 14.11 37.81 0 0 0 0 1 0 1 0 2 1 0 528 I Belvedere 15.87 39.63 0 0 0 0 0 0 0 0 2 0 2 429 I Zomaro 16.15 38.31 0 0 0 0 0 0 0 0 0 2 2 430 TK Hopa 41.57 41.39 0 0 0 0 0 0 0 0 1 0 3 431 TK Sinop 35.05 42.00 0 0 0 0 0 0 0 0 4 0 0 432 TK Akcakoca 31.16 41.07 0 0 0 0 0 0 0 0 5 0 0 533 TK Bursa 29.08 40.12 0 0 0 0 0 0 0 0 5 0 0 534 TK Kemalpasa 27.35 38.40 2 1 0 0 0 0 0 0 0 0 0 335 TK Oluk 30.00 37.00 0 5 0 0 0 0 0 0 0 0 0 536 TK Golcuk 29.79 40.72 0 0 0 0 0 0 0 0 2 0 2 437 TK Sapanca 30.25 40.65 0 1 0 0 0 0 0 0 2 0 2 538 TK Karadere 30.84 40.75 0 0 0 0 0 0 0 0 3 0 2 5Total sample size 4 13 4 4 9 4 4 7 102 8 22 181Percentage 0.022 0.072 0.022 0.022 0.050 0.022 0.022 0.039 0.564 0.044 0.122

*E, Spain; P, Portugal; F, France; I, Italy; TK, Turkey.

ME

C1029.fm

Page 1497 T

hursday, Septem

ber 7, 2000 3:53 PM

1498

S . F I N E S C H I

E T A L .

© 2000 Blackwell Science Ltd,

Molecular Ecology

, 9, 1495–1503

Restriction fragments were separated by electrophoresisin 8% polyacrylamide gels using Tris Borate EDTA buffer(1

×

) at constant 300 V for 90–120 min. After electrophoresis,gels were stained with ethidium bromide and photographedunder UV light with Polaroid 667 film using a MP4 PolaroidLand Camera. A 100 bp ladder from Gibco BRL/Life Techno-logies was used as size marker.

Polymorphism occurred as point mutations and inser-tion/deletions. Polymorphic fragments were labelled bydecreasing order of fragment length as visualized in thepolyacrylamide gels and as described by Demesure

et al

.(1996). Haplotypes were defined according to differentcombinations of length variants.

The level of population subdivision using unorderedalleles (

G

STc

) for cytoplasmic genomes was calculatedaccording to Pons & Petit (1995) using the software

haplodiv

. The level of population subdivision for orderedalleles (

N

STc

) was calculated using the program

haplonst

(Pons & Petit 1996). These two parameters may becompared using an analytical test (Pons & Petit 1996). Apermutation test was also designed to confirm the differ-ence: 1000 random permutations of haplotypes identitieswere made, maintaining the haplotype frequencies andthe matrix of pairwise haplotype differences as in theoriginal study (Burban

et al

. 1999). The distribution of

values obtained by permutation was compared with theobserved values. For the

N

STc

analysis, a distance matrixderived from the pairwise number of mutational dif-ferences between haplotypes was used. According toPons & Petit (1996), significantly higher values for

N

STc

than for

G

STc

indicate the existence of a phylogeographicstructure.

The ratio of seed to pollen flow was calculated usinga modified equation of Ennos (1994), in which

F

ST

wassubstituted by

G

ST

:

where

G

STb

and

G

STc

indicate the level of populationsubdivision based on nuclear and cytoplasmic markers,respectively. In this study,

G

STb

is taken from Villani

et al

.(1991a, 1997, 1999 and personal communication).

Results

One mitochondrial and nine chloroplast fragments, eachdigested with 1–3 restriction enzymes, were analysedto identify polymorphism. Restriction analysis revealed

Table 2 Details on primers, restriction enzymes, amplification conditions, and indication of polymorphic sites

Gene Primer pairs and sequence CodeAnnealing temperature (°C)

Extension time (min)

Restriction enzymes Polymerase Reference

Chloroplast primersTrnC CCAGTTCAAATCTGGGTGTC CD 58 4 TaqI No 1TrnD GGGATTGTAGTTCAATTGGTTrnD ACCAATTGAACTACAATCCC DT 54.5 2 AluI Yes 1TrnT CTACCACTGAGTTAAAAGGGpsaA ACTTCTGGTTCCGGCGAACGAA AS 57.5 4 HinfI No 1TrnS AACCACTCGGCCATCTCTCCTATrnT CATTACAAATGCGATGCTCT TF 57.5 2 TaqI No 2TrnF ATTTGAACTGGTGACACGAGTrnK1 GGGTTGCCCGGGACTCGAAC K1K2 62 2 TaqI, Yes 1TrnK2r CAACGGTAGAGTACTCGGCTTTTA CfoI, AluI NoRpoC1 GCACAAATTCCRCTTTTTATRGG RpoCC 47.5 4 HinfI No 3TrnCr CGACACCCRGATTTGAACTGGTrnS CGAGGGTTAGGAATCCCTCTC ST 57.5 2 HinfI No 1TrnT AGAGCATCGCATTTGTAATGTrnF CTCGTGTCACCAGTTCAAAT FV 57.5C 4 HinfI Yes 3TrnV1 CCGAGAAGGTCTACGGTTCGTrnV2 CGAACCGTAGACCTTCTCGG VL 57.5C 4 HinfI No 3RbcL GCTTTAGTCTCTGTTTGTGG

Mitochondrial primersnad1B GCATTACGATCTGCAGCTCA nad1/2–3 57.5C 4 TaqI, RsaI No 1nad1C GCAGCTCGATTAGTTTCTGC

*1, Demesure et al. 1995; 2, Taberlet et al. 1991; 3, Dumolin-Lapègue et al. 1997a.

pollen flowseed flow

------------------------------

1GSTb----------- 1–

2 1GSTc----------- 1–

–

1GSTc----------- 1–

---------------------------------------------------------------=


C H L O R O P L A S T D N A P O LY M O R P H I S M I N E U R O P E A N C H E S T N U T 1499

© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1495–1503

three polymorphic fragments (Table 2); in particular, apoint mutation was detected in one amplified fragment(K1K2), and length polymorphisms in the other twofragments (DT and FV), identifying a total of 13 variants(Table 3), which combined into 11 cpDNA haplotypes(Table 4). The distribution of the observed haplotypes ineach population is indicated in Table 1 and Fig. 1.

Haplotypes I and K are the most frequent ones (0.56and 0.12, respectively) (Table 1). The most commonhaplotype, I, occurs in all countries; it is absent only in thesouth-western part of Turkey (populations 34 and 35),in two populations from Spain (populations 2 and 6)and in six Italian populations (populations 20, 21, 22, 23,25 and 29).

The frequency of the other haplotypes is very low,ranging from 0.02 to 0.07 (Table 1). These haplotypesmay be divided into three classes: (i) rare but widelydistributed ones (haplotype B); (ii) haplotypes distributedin the eastern and central part of the area (A, K); and(iii) haplotypes distributed in the western and central

part of the area (D, E, F, G, H and J). One haplotype (C)was detected in a single population from Sicily (Italy).

Haplotype B is the only one represented in the southernTurkish population Oluk; it is also present in two otherpopulations of the same area (Kamalpasa and Sapanca),in two central Italian populations (Cavriglia and Allumiere),and in two Iberian populations (Hervas in Spain andBraganca in Portugal).

Haplotypes A and K occur in Turkey, thought to bethe area of origin of European chestnut, and in Italy.On the other hand, haplotypes D, E, F, G, H and J occur inthe Iberian peninsula, which might have been a refugialarea, and in Italy as well.

Twenty-five populations out of 38 are polymorphic.Among populations represented by only one haplotype,12 are monomorphic for the most common haplotype(haplotype I); two populations located in northern Italy(populations 21 and 22) are characterized by the presenceof haplotype K only, whereas one population from theTurkish Mediterranean coast (Oluk) is fixed for haplotype B.

Table 3 Size of the amplified fragments, number and type of variants detected

Primer codeApproximate size of PCR product (bp) Enzymes

Total number of RFLP bands Band number

Total number of variants Type of variants

K1K2 2500 TaqI 5 II 2 Point mutationFV 3500 HinfI 5 I 3 Indel

II 2 IndelIV 2 IndelV 2 Indel

DT 1200 AluI 6 II 2 Indel

Primer codes correspond to those indicated in Table 1.

Table 4 cpDNA haplotypes detected for Castanea sativa

K1K2 FV FV FV FV DT TaqI HinfI HinfI HinfI HinfI AluI

Haplotype Band II Band I Band II Band IV Band V Band II Population

A 1 1 2 1 1 1 20, 24, 34B 1 1 2 1 1 2 10, 13, 20, 21, 34, 35, 37C 1 1 2 1 2 2 25D 1 2 1 2 2 1 3, 5, 6, 20E 1 2 1 2 2 2 2, 6, 10, 27F 1 2 2 1 1 2 2, 21G 0 1 2 1 1 1 1, 18, 27H 0 2 1 1 2 1 8, 14, 15, 24I 0 2 1 2 2 1 1, 3, 4, 5, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18,

19, 24, 26, 27, 28, 30, 31, 32, 33, 36, 37, 38J 0 2 1 2 2 2 6, 20, 24, 25, 27, 29K 0 3 1 2 2 1 21, 22, 23, 28, 29, 30, 31, 36, 37, 38

Mutations are ordered according to decreasing molecular weight. A zero represents the absence of a restriction fragment from its expected position on a gel. The last column indicates the population number in which the different haplotypes were found.


1500 S . F I N E S C H I E T A L .


Within-population haplotypic diversity for the Italianpopulations was higher than in the other part of the distri-bution range (hs = 0.49, 0.32, 036 for Italy, Western Europe(Portugal, Spain and France) and Turkey, respectively).

The level of population subdivision was relativelylow, GSTc = 0.43 (hs = 0.38; ht = 0.68) compared with otherangiosperms studied (review by Petit 1999). The level ofpopulation subdivision for ordered alleles was higher thanthe level of subdivision for unordered alleles: NSTc = 0.52(vs = 0.28; vt = 0.60). The difference between GSTc and NSTcwas highly significant (U test = 3.21, P < 0.01, accordingto Pons & Petit 1996). This results was confirmed bypermutation analysis of haplotype identities.

Results of the pollen flow/seed flow ratio gave a verylow value, close to 1, indicating that gene flow via pollenis less or at least not more efficient than by seed.

Discussion

A general picture of the history of chestnut after thelast glacial period can be deduced from fossil pollen data.In Europe, two periods have to be considered: the first

comprising the first pollen records in Spain and Greece,the second characterized by the cultivation of the speciesoutside its natural range (Huntley & Birks 1983). Thissecond period coincides with the Roman and post-Romanage, in which many human activities, including tree cul-tivation, were spread towards central and northern Europe.According to pollen maps provided by Huntley & Birks(1983), at least three glacial refugia, corresponding to thethree Mediterranean peninsulas, are indicated for chestnut.

Further information on the history of chestnut in thewhole Mediterranean basin is given by Zohary & Hopf(1988), who point out the important role played by theTurkish populations for the survival of the species duringice periods. Palynological data support the hypothesisthat European chestnut originated in Turkey. During theWürm glaciation, chestnut apparently survived only insouth-western Asia and disappeared from southern Europe.Fossil records became copious in Anatolia and Greecearound 3500 years ago, and in Italy and western Medi-terranean around 2500 years ago: this strongly suggeststhat chestnut ‘did not arrive in Greece and westernMediterranean countries as a wild element but was

Fig. 1 Distribution and frequency of cpDNA haplotypes within European chestnut populations. The size of the pie charts indicates thenumber of individuals for each population (from 3 to 5).




introduced by man’ (Zohary & Hopf 1988). Data on nuclearmarkers also support the hypothesis of Turkey as a zoneof possible origin and expansion for chestnut towardswestern European countries (Villani et al. 1991a, 1994).More recently, another possible refugium in northern Spainwas suggested (Villani et al. 1997). Our data do not provideadditional useful information about the possible migrationpathways during the post-glacial period, but the actualdistribution of the haplotypic diversity does seem to haveexperienced the influence of human activities, typicalof the second period of chestnut history. However, thesedata must be interpreted with caution considering thatpopulations from the Balkan peninsula are not includedin this survey.

Results by Villani et al. (1991a, 1994, 1997, 1999) demon-strated the peculiar importance of Turkish populations. Inthis region, three ecologically different areas have beenstudied using genetic markers: the north-eastern one,located along the Black Sea coasts, which is characterizedby humid climate; the Mediterranean region, which ismore arid; and a third zone, on the north-western part ofAsian Turkey, the Bythynian region, which was identifiedon the basis of genetic parameters and classified as ahybrid zone (Villani et al. 1999). In our study, populationsfrom the three zones have been analysed: Hopa, Sinopand Akcakoca are populations from the Black Sea; Bursa,Kemalpasa and Oluk are Mediterranean populations;finally Golcuk, Sapanca and Karadere represent the hybridzone. Chloroplast DNA suggests a division betweentwo Mediterranean populations (Oluk and Kemalpasa),characterized by the haplotypes A and B, and all otherTurkish populations, in which only haplotypes I and Kare present, with the only exception of Sapanca; in thispopulation, which is located in the hybrid zone, haplotypescommon to the two groups of populations (haplotypes B,I, and K) are present.

Haplotypes A and B might have migrated westwards,but their presence is extremely sporadic; on the otherhand, haplotype I is the most common one and spreads inall analysed regions. For this reason, at this stage theidentification of possible migration pathways is verydifficult. However, the absence of several haplotypes inTurkey and their presence in the remaining part of thedistribution area suggests the role of the Iberian peninsulaas a possible refugial area. Haplotypes D, E, F, G, H and Jmight have migrated from the Iberian region eastwards,colonizing France and Italy.

A high within-population haplotypic diversity was foundin Italy. The mixed distribution of ubiquitous haplotypesdetected in Italy does not seem to support the hypothesisof an Italian refugium. The analysis of cpDNA poly-morphism confirmed the presence of Italian refugia forother species belonging to the family Fagaceae (Fagussylvatica, Demesure et al. 1996; Quercus robur, Q. petraea

and Q. pubescens, Dumolin-Lapègue et al. 1997a). It shouldbe stressed once more that chestnut is, within the membersof the Fagaceae, the species that has experienced thestrongest human impact. The important role played byhuman populations, not only by Romans, in moving andtransferring propagation material throughout Europeancountries has already been pointed out. The presence inItaly of both groups of haplotypes, those detected in theeast (haplotypes A, B, I and K) and those present in thewest (haplotypes D, E, F, G, H and J) might also indicateItaly as a possible meeting point of two colonization routesfrom east and west, or might be the result of introductionof material from both areas. As already commented,according to Huntley & Birks (1983), the first pollen recordsin Italy appeared 5500 years ago, more recently than inSpain and Greece. An evident increase of pollen wasdetected about 2000 years ago, a period which coincideswith the most active phase of Roman expansion. This maysuggest that the present geographical distribution of thespecies and therefore the distribution of genetic diversitymight have been influenced more by human activitiesthan by the natural migration and colonization.

European chestnut showed a geographically structureddistribution of the diversity using nuclear isozyme markers,and in particular a strong differentiation between easternTurkish populations and all others. Villani et al. (1999)estimated a high value of population subdivision whencomparing the two groups of populations (GSTb = 0.184).

The level of cpDNA differentiation among populationsin chestnut is relatively low: the most common haplotype(haplotype I) is widely diffused in all distribution areasand other haplotypes are widespread as well; however,the higher value of NSTc (0.522) than GSTc (0.433), and thehighly significant difference between them, suggests thatsome geographical structure of the diversity is still retained.The level of population subdivision was higher for the otherEuropean Fagaceae investigated so far: GSTc = 0.902 forFagus sylvatica (Demesure et al. 1996); GSTc = 0.829 forQuercus petraea, GSTc = 0.907 for Q. pubescens, and GSTc =0.782 for Q. robur (Dumolin-Lapègue et al. 1997a). Theonly related Fagaceae of the southern hemisphere studiedso far, Nothofagus nervosa, showed a very clear geographicalstructure of cpDNA polymorphism, revealing a geo-graphical divide across its distribution area in Argentina(Marchelli et al. 1998).

In a comparative study on cytoplasmic diversity (Petit1999), the mean value of population subdivision meas-ured for 97 plant species was GSTc = 0.70; in particular theGSTc value was 0.73 for angiosperm tree species, higherthan the value estimated for chestnut. The history of thisspecies and the long period of cultivation experiencedduring the last thousand years could explain the distribu-tion of cpDNA diversity, and the peculiarity of chestnutamong European Fagaceae.


1502 S . F I N E S C H I E T A L .


The pollen/seed flow value for chestnut is close to 1.This low value may be explained by the insect contribu-tion to the pollination and the active role of humans inmoving and transferring not only fruits but also propaga-tion material. It is interesting to note that this parameteragain indicates a difference between chestnut and theother European genera of Fagaceae, in fact Petit (1999)reported much higher values for oaks (Quercus robur = 286,Q. petraea = 500) and beech (84). It is worth rememberingthat the entomophilous syndrome does not belong tothe flowering feature of the other genera Quercus andFagus, which are described as completely anemophilous(Kubitzki et al. 1993). On the other hand, the pollen flow/seed flow ratio is low when calculated for tree speciescharacterized either by insect pollination or by veryefficient seed dispersal mechanisms; for example Petit(1999) reports values similar to that of chestnut (< 5) forEucalyptus nitens, Sorbus aucuparia and Argania spinosa.

The results obtained in this study show that under-standing of the history of a tree species, as known fromhuman tradition and culture and documented by fossilpollen analysis, can be complemented by a molecularapproach. Further information about migration pathwaysand the history of chestnut after the last ice age canbe acquired by intensive sampling and analysis of othergeographical regions.

Acknowledgements

The authors wish to thank Maria Fernanda Sanchez Goñi(Bordeaux, France), Rémy Petit (INRA-Pierroton, France) andSalvatore Cozzolino (Naples, Italy) for helpful suggestions andcomments. The co-operation of Marcello Cherubini, AntonioSansotta from Porano and Osman Taskin, together with the staffof the Turkish Forest Research Institutes, is warmly acknowledged.

References

Bennet KD, Tzedakis PC, Willesk KJ (1991) Quaternary refugia ofthe North Europeans trees. Journal of Biogeography, 18, 103–115.

Burban C, Petit RJ, Carcreff E, Jactel H (1999) Rangewide vari-ation of the maritime pine bast scale Matsucoccus feytaudi Duc.(Homoptera: Matsucoccidea) in relation to the genetic structureof its host. Molecular Ecology, 8, 1593–1602.

Demesure B, Sodzi N, Petit RJ (1995) A set of universal primersfor amplification of polymorphic non-coding regions of mito-chondrial and chloroplast DNA in plants. Molecular Ecology, 4,129–131.

Demesure B, Comps B, Petit RJ (1996) Chloroplast DNA phylo-geography of the common beech (Fagus sylvatica L.) in Europe.Evolution, 50, 2515–2510.

Doyle JJ, Doyle JH (1990) Isolation of plant DNA from freshtissue. Focus, 12, 13–15.

Dumolin S, Demesure B, Petit RJ (1995) Inheritance of chloroplastand mitochondrial genomes in pedunculate oak investigatedwith an efficient PCR-based method. Theoretical and AppliedGenetics, 91, 1253–1256.

Dumolin-Lapègue S, Demesure B, Le Corre V, Fineschi S, Petit RJ(1997a) Phylogeographic structure of white oaks throughoutthe European continent. Genetics, 146, 1475–1487.

Dumolin-Lapègue S, Pemonge MH, Petit RJ (1997b) An enlargedset of consensus primers for the study of organelle DNA inplants. Molecular Ecology, 6, 393–397.

Ennos RA (1994) Estimating the relative rates of pollen and seedmigration among plant populations. Heredity, 80, 584–593.

Ferris C, Oliver RP, Davy AJ, Hewitt GM (1993) Native oakchloroplasts reveal an ancient divide across Europe. MolecularEcology, 2, 337–344.

Ferris C, Oliver RP, Davy AJ, Hewitt GM (1995) Using chloroplastDNA to trace postglacial migration of oaks into Britain. MolecularEcology, 4, 731–738.

Ferris C, King RA, Vainola R, Hewitt GM (1998) ChloroplastDNA recognises three refugial sources of European oaks andshows independent eastern and western immigrations toFinland. Heredity, 80, 584–593.

Fineschi S (1988) Genetics of Chestnut (Castanea sativa Mill.) II.Uniformity of enzyme phenotypes in grafted tree orchards. SilvaeGenetica, 37, 82–83.

Fineschi S, Malvolti ME (1991) Genetic conservation and geneticresources in Chestnut (Castanea sativa Mill.). In: Genetic Vari-ation in European Populations of Forest Trees (eds Müller-Starck G,Ziehe M), pp. 181–189. Sauerlaender’s Verlag, Frankfurt am Main.

Fineschi S, Malvolti ME, Morgante M, Vendramin GG (1994a)Allozyme variation within and among cultivated varieties ofsweet Chestnut (Castanea sativa Mill.). Canadian Journal of ForestResearch, 24, 1160–1165.

Fineschi S, Taurchini D, Müller-Starck G, Conedera M (1994b)Genetic characterisation of cultivated varieties of Europeanchestnut (Castanea sativa Mill.) in southern Switzerland. III.Analysis of RAPD molecular markers. In: Proceedings of anInternational Congress on Chestnut, Spoleto, Italy (ed. Antognozzi E),pp. 309–313. Instituto di Coltivazioni Arboree, Universitá diPerugia, Perugia, Italy.

Gellini R, Grossoni P (1997) Botanica Forestale, volume II, CEDAM,Padova.

Harlow WM, Harrar ES, Hardin JW, White FM (1996) Textbook ofDendrology. McGraw-Hill Inc., New York.

Hewitt GM (1996) Some genetic consequences of ice ages andtheir role in divergence and speciation. Biological Journal of theLinnean Society, 58, 247–276.

Huntley B, Birks HJB (1983) An Atlas of Past and Present PollenMaps for Europe: 0–13 000 Years Ago. Cambridge UniversityPress, Cambridge.

Kremer A, Petit RJ (1993) Gene diversity in natural populations ofoak species. Annales Sciences Forestiers, 50 (Suppl. 1), 186–202.

Kubitzki K, Rohwer JG, Bittrich V (1993) The Families and Generaof Vascular Plants. II Flowering Plants, Dicotyledons. Springer Verlag,Berlin.

Marchelli P, Gallo L, Scholz F, Ziegenhagen B (1998) ChloroplastDNA markers reveal a geographical divide across Argentineansouthern beech Nothofagus nervosa (Phil.) Dim. et Mil. distribu-tion area. Theoretical and Applied Genetics, 97, 642–646.

Müller-Starck G, Conedera M, Fineschi S (1994) Genetic character-isation of cultivated varieties of European chestnut (Castaneasativa Mill.) in southern Switzerland. II. Genetic inventorybased on enzyme gene markers. In: Proceedings of an Inter-national Congress on Chestnut, Spoleto, Italy (ed. Antognozzi E),pp. 303–307. Instituto di Coltivazioni Arboree, Universitá diPerugia, Perugia, Italy.




Petit RJ (1999) Diversité Génétique et Histoire des Populations d’ArbresForestiers. Dossier d’habilitation à diriger des recherches, Universitéde Paris-Sud, Université Formation de Recherche Scientifiqued’Orsay, Paris.

Petit R, Kremer A, Wagner DB (1993) Geographic structure ofchloroplast DNA polymorphisms in European oaks. Theoreticaland Applied Genetics, 87, 122–128.

Pigliucci M, Villani F, Benedettelli S (1990) Geographic and climaticfactors associated with the spatial structure of gene frequenciesin Castanea sativa Mill. from Turkey. Journal of Genetics, 69, 141–149.

Pons O, Petit RJ (1995) Estimation, variance and optimal samplingof gene diversity. 1. Haploid locus. Theoretical and Applied Genetics,90, 462–470.

Pons O, Petit RJ (1996) Measuring and testing genetic differentiationwith ordered versus unordered alleles. Genetics, 144, 1237–1245.

Sánchez Goñi MF (1993) De la taphonomie pollinique à la reconsti-tution de l’environnement. L’exemple de la région cantabrique.Tempus Reparatum. British Archeological Reports 586, 44–45.

Soltis DE, Soltis PS, Milligan BG (1992) Intraspecific chloroplastDNA variation: systematics and phylogenetic implications. In:Molecular Plant Systematics (eds Soltis PS, Soltis DE, Doyle JJ),pp. 117–150. Chapman & Hall, New York.

Taberlet P, Fumagalli L, Wust-Saucy AG, Cosson JF (1998) Com-parative phylogeography and postglacial colonization routesin Europe. Molecular Ecology, 7, 453–464.

Taberlet P, Gielly L, Patou G, Bouvet J (1991) Universal primersfor amplification of three non-coding regions of chloroplastDNA. Plant Molecular Biology, 17, 1105–1109.

Villani F, Pigliucci M, Benedettelli S, Cherubini M (1991a)Genetic differentiation among Turkish chestnut (Castanea sativaMill.) populations. Heredity, 66, 131–136.

Villani F, Benedettelli S, Paciucci M, Cherubini M, Pigliucci M(1991b) Genetic variation and differentiation between naturalpopulations of chestnut (Castanea sativa Mill.) from Italy. In:

Biochemical Markers in the Population Genetics of Forest Trees (edsFineschi S, Malvolti ME, Cannata F, Hattemer HH), pp. 91–103. SPB Academic Publishing, The Hague, The Netherlands.

Villani F, Pigliucci M, Cherubini M (1994) Evolution of Castaneasativa Mill. in Turkey and Europe. Genetic Research Cambridge,63, 109–116.

Villani F, Cherubini M, Sansotta A, Taurchini D, Fornari B (1997)Inter- and intra-populations genetic diversity of Mediterraneanchestnut (Castanea sativa Mill.) in relation to various environ-mental factors. I. In: Dynamics and Function of Chestnut ForestEcosystems in Mediterranean Europe (eds Romane F, GrossmannA), pp. 119–128. Commission of the European Communities,DG XII, Brussels.

Villani F, Sansotta A, Cherubini M, Cesaroni D, Sbordoni V(1999) Genetic structure of natural populations of Castaneasativa in Turkey: evidence of a hybrid zone. Journal of Evolu-tionary Biology, 12, 233–244.

Zohary D, Hopf M (1988) Domestication of Plants in the Old World.Clarendon Press, Oxford.

Silvia Fineschi is a forest geneticist interested in the phylogeo-graphy and ecological genetics of angiosperm tree species andinvolved in the application of molecular markers for assessinggenetic variation in forest tree populations. Daniela Taurchiniis a biologist with experience in molecular laboratory techniques.Fiorella Villani is a population geneticist with competence in geneconservation, evolutionary and ecological genetics, and in theuse of integrated approaches (biochemical, morphological andphysiological) in systematic and evolutionary studies. Giovanni G.Vendramin is a population geneticist involved in the applicationof molecular markers to population genetics of forest tree species,with particular emphasis on the study of the importance ofhistorical factors in shaping genetic variation.


Chloroplast DNA polymorphism reveals little geographical structure in Castanea sativa Mill. (Fagaceae) throughout southern European countries

Documents