Genetic relationships and evolution in Cucurbita pepo (pumpkin, squash, gourd) as revealed by simple sequence repeat polymorphisms [2012]

1 23

Genetic Resources and CropEvolutionAn International Journal ISSN 0925-9864Volume 60Number 4 Genet Resour Crop Evol (2013)60:1531-1546DOI 10.1007/s10722-012-9940-5

Genetic relationships and evolution inCucurbita as viewed with simple sequencerepeat polymorphisms: the centrality of C.okeechobeensis

Li Gong, Harry S. Paris, Gertraud Stift,Martin Pachner, Johann Vollmann &Tamas Lelley

1 23

Your article is protected by copyright and all

rights are held exclusively by Springer Science

+Business Media Dordrecht. This e-offprint

is for personal use only and shall not be self-

archived in electronic repositories. If you wish

to self-archive your article, please use the

accepted manuscript version for posting on

your own website. You may further deposit

the accepted manuscript version in any

repository, provided it is only made publicly

available 12 months after official publication

or later and provided acknowledgement is

given to the original source of publication

and a link is inserted to the published article

on Springer's website. The link must be

accompanied by the following text: "The final

publication is available at link.springer.com”.

RESEARCH ARTICLE

Genetic relationships and evolution in Cucurbita as viewedwith simple sequence repeat polymorphisms: the centralityof C. okeechobeensis

Li Gong • Harry S. Paris • Gertraud Stift •

Martin Pachner • Johann Vollmann •

Tamas Lelley

Received: 12 July 2012 / Accepted: 23 November 2012 / Published online: 7 December 2012

� Springer Science+Business Media Dordrecht 2012

Abstract Genetic relationships among 88 acces-

sions from nine of the dozen species of Cucurbita

(Cucurbitaceae) were assessed from polymorphisms at

74 SSR (simple sequence repeat) loci originating from

C. pepo and C. moschata, yielding a total of 315 alleles

distributed among 17 linkage groups, with an average

of 4.3 alleles per locus. Genetic distance (GD) values

were calculated, a principal coordinate analysis con-

ducted, and a dendrogram constructed. Average

within-species genetic distance values ranged from

0.07 for C. ecuadorensis and C. ficifolia to 0.46 for

C. pepo. Each species was clearly defined, as all mean

within-species GD values were lower than the respec-

tive mean between-species GD values. C. okeechobe-

ensis had the most central position in the genus

Cucurbita, with the lowest average GD to the other

species, 0.61. C. foetidissima, the only xerophytic

species examined, was the most distant, with a mean

GD of 0.73 to the other species. C. pepo and C. ficifolia

were the most outlying of the mesophytic species.

Mean across-species GDs generally corresponded

with crossability. However, there were some outstand-

ingly low GD values between particular accessions of

Cucurbita pepo, the economically most important

species, and disease-resistant wild species, particu-

larly C. okeechobeensis but also C. foetidissima. The

results suggest that more intensive search and collec-

tion of C. okeechobeensis populations would likely

yield genotypes that are more compatible with

C. pepo. Moreover, successful application of genetic

resources in the genus Cucurbita might be facilitated

by using GD values obtained from SSR polymor-

phisms as a guide in choosing parents for interspecific

crossing.

Keywords Cucurbita � Genetic resources �Genome analysis � Gourd � Pumpkin � Squash

Electronic supplementary material The online version ofthis article (doi:10.1007/s10722-012-9940-5) containssupplementary material, which is available to authorized users.

L. Gong � G. Stift � M. Pachner � T. Lelley

Department for Agrobiotechnology, University of Natural

Resources and Life Sciences, Vienna, Vienna, Austria

L. Gong � G. Stift � M. Pachner � T. Lelley

Institute for Biotechnology in Plant Production,

IFA-Tulln, Konrad Lorenz Str. 20, 3430 Tulln, Austria

Present Address:L. Gong

Department of Plant Pathology, North Dakota

State University, Fargo, ND 58108, USA

H. S. Paris (&)

Department of Vegetable Crops and Plant Genetics,

Agricultural Research Organization, Newe Ya‘ar

Research Center, P. O. Box 1021,

Ramat Yishay 30-095, Israel

e-mail: [email protected]

J. Vollmann

Plant Breeding Division, Department of Crop Sciences,

University of Natural Resources and Life Sciences

Vienna, Konrad Lorenz Str. 24, 3430 Tulln, Austria

123

Genet Resour Crop Evol (2013) 60:1531–1546

DOI 10.1007/s10722-012-9940-5

Author's personal copy

Introduction

The genus Cucurbita (2n = 2x = 40) is native to the

Americas, distributed in the wild from the USA to

Argentina. There is much diversity within the genus

for vegetative and flowering characteristics and fruit

size, shape, exterior and interior color, surface topog-

raphy, fruit flesh quality, and seed size and color

(Whitaker and Davis 1962; Robinson and Decker-

Walters 1997; Goldman 2004).

Five species of Cucurbita have been cultivated for

millennia in the Americas, mostly for their edible

fruits which are known as pumpkins and squash.

Although the genus was considered at one time to have

27 species (Whitaker and Bemis 1975), crossing

experiments have revealed free crossability and

unimpaired fertility in many cross-combinations. For

example, the wild plants named C. texana (Scheele)

Gray and C. fraterna Bailey hybridize freely with

C. pepo and therefore these three names are given for

the same biological species; C. texana and C. fraterna

are merely synonyms of C. pepo. For each of seven

other species, C. argyrosperma Huber, C. maxima

Duchesne, C. okeechobeensis (Small) Bailey, C. rad-

icans Naudin, C. pedatifolia Bailey, C. foetidissima

HBK, and C. digitata Gray, synonyms have been used

fairly commonly. The remaining species are C. mosch-

ata Duchesne, C. ficifolia Bouche, C. lundelliana

Bailey, C. ecuadorensis Cutler et Whitaker, and the

poorly known C. galeottii Cogniaux. Thus, the genus

consists of 12 or 13 species (Nee 1990).

Beginning with the European contact with the Amer-

icas a little over 500 years ago, fruits and seeds of

Cucurbita were carried to other continents, where they

have since been grown for centuries, from the cool

temperate zones to the tropics. The five cultivated species

derive from different eco-geographical zones in the

Americas, and this is reflected in their particular

adaptations to cultivation (Nee 1990). For example,

C. ficifolia is adapted to cool, short-day situations whilst

C. moschata is best adapted to the humid tropics. C. pepo

has the widest adaptation, reflected in its far greater

economic importance than the other species.

Antoine Nicolas Duchesne (1747–1827), working

in France, was the first to systematically attempt to

cross various cultigens of Cucurbita (Duchesne 1786).

While most of his germplasm collection was from a

single species, C. pepo, the lack of success that he

encountered in some of the crosses he attempted led

him to realize that his collection also included a few

accessions from two other species, which he named

C. moschata and C. maxima (Paris 2000, 2007).

Decades later, Charles Naudin studied crossability

within Cucurbita and confirmed Duchesne’s findings

(Naudin 1856). Since then, interspecific crosses have

been attempted repeatedly in the genus. One goal was

to test species barriers and obtain a better understand-

ing of phylogenetic and evolutionary relationships

among the species. Another, more practical goal was

the introgression of desirable traits from one species

into another (Whitaker and Davis 1962). C. pepo, the

species with the greatest monetary value, excels in

plant earliness and productivity but lacks genetic

resources for disease resistance. C. moschata, on the

other hand, carries resistance to various pathogens but

lacks earliness and productivity. Many attempts have

been made over the years to expedite interspecific

crossing within the genus in order to form a large pool

of genetic resources. Aids to these attempts have

employed various breeding strategies and biotechno-

logical approaches but the degrees of success in

introgressing desirable traits from one species to

another have, overall, been quite variable and incom-

plete (Lebeda et al. 2007).

For the past quarter century, Cucurbita has been

subjected to molecular genetic scrutiny (Lebeda et al.

2007). Potentially, molecular genetic approaches

could further elucidate genetic relationships among

the species of Cucurbita as well as indicate possible

interspecific cross combinations that would have a

greater probability of success. Most investigations,

though, have focused on variation within a single

species using nuclear DNA markers. Usually, the

studies encompassed accessions derived mostly from a

particular country or geographical region (Gwanama

et al. 2000; Ferriol et al. 2001, 2003a, b, 2004a, b; Wu

et al. 2011). Others, notably with C. pepo, have

focused on the worldwide genetic variation (Katzir

et al. 2000; Paris et al. 2003; Ferriol et al. 2003b; Gong

et al. 2012). A few investigations have focused on

interspecific variation within Cucurbita. The molec-

ular genetic approach first used for studying phyloge-

netic relationships among species of the genus

Cucurbita was allozyme variation of several enzyme

systems (Wilson 1989; Puchalski and Robinson 1990;

Decker-Walters et al. 1990). Torres Ruiz and Hemle-

ben (1991) were able to distinguish among 11

Cucurbita cultivars, nine of which were C. pepo, by

1532 Genet Resour Crop Evol (2013) 60:1531–1546

123

Author's personal copy

using ribosomal intergenic spacer probes. Subse-

quently, organellar DNA diversity was examined by

Wilson et al. (1992), who compared chloroplast DNA

in 30 accessions from 10 species of Cucurbita, and

Sanjur et al. (2002), who compared mitochondrial

gene DNA in 65 accessions from eight species.

Interspecific comparisons of nuclear DNA, employing

RAPDs (Baranek et al. 2000) and SSRs derived from

Cucumis (Pico et al. 2005–2006), have been narrow in

scope. Overall, the results of all of these investigations

were consistent with results obtained from crossing

attempts among the various species (Whitaker and

Davis 1962; Puchalski and Robinson 1990).

To our knowledge, there have been no attempts to

examine polymorphisms of nuclear DNA among the

majority of Cucurbita species. Genomic marker

systems of nuclear origin differ in type and frequency

of variation in comparison to those of organellar

origin. Organelle DNA is passed down to the next

generation through one parent, usually but not always

the female (Havey et al. 1998), while nuclear DNA is

passed down from both parents and is subject to

meiotic recombination. The resulting increased fre-

quency of nuclear polymorphisms leads to their

increased power of resolution. Moreover, some

nuclear DNA marker systems can provide more

reliable and informative results than others. In these

respects, simple sequence repeats (SSRs) are among

the preferred types of DNA polymorphisms, but they

have not been used on a wide survey of the genus

Cucurbita. Our objective was to observe SSR poly-

morphisms on a large collection of Cucurbita from

various geographical regions in order to obtain an

improved insight into relationships among most of the

species of this genus.

Materials and methods

Plant material

A collection comprised of 88 accessions derived from

nine species of Cucurbita was chosen for this study

(Table 1). Included were 17 accessions of C. argyro-

sperma, 18 of C. maxima, five each of C. ecuadorensis

and C. foetidissima, four each of C. ficifolia and

C. lundelliana, 20 of C. moschata, eight of C. okeech-

obeensis, and seven of C. pepo. Four of these species,

C. argyrosperma, C. maxima, C. okeechobeensis, and

C. pepo, have each been classified into subspecific taxa

(Nee 1990), and accessions representative of these taxa

were included in this study. Most of the accessions were

obtained from germplasm repositories in the USA.

Others are maintained at the Newe Ya‘ar Research

Center of Israel’s Agricultural Research Organization

(Paris 2001). Seed samples obtained from markets in

China are maintained at the Institute for Biotechnology

in Plant Production, Department of Agrobiotechnology,

Tulln, University of Natural Resources and Life

Sciences, Vienna, Austria. A sample of cucumber,

Cucumis sativus L., a commercial Austrian hybrid,

‘Sensation’, was included, as an outlier, for comparison.

DNA extraction, PCR amplification, and detection

of SSR loci

DNA isolation was conducted as described by Gong et al.

(2008a). A minimum of three seedlings per accession

were sampled in bulk. Out of the 500 Cucurbita SSR

markers developed from genomic DNA by Gong et al.

(2008a), 74 were used in this study. Of these, 60 % were

derived from C. pepo and the remainder from C.

moschata. Of the 74 SSR loci, 46 (62 %) have been

mapped (Gong et al. 2008b). Fifty-four of the same

markers were used in a recent study on relationships and

evolution within C. pepo (Gong et al. 2012).

The total volume of the PCR mixture was 10 ll,

and contained 27 ng DNA, 1 9 GoTaq� Green Mas-

ter Mix (Promega), 0.25 pmol forward primer with an

M13 tail added to its 50 end (50-CCCAGTCAC-

GACGTTG-30), 2.5 pmol reverse primer, and

2.25 pmol fluorescent labeled M13 tail (FAM,

HEX), synthesized by MWG, Ebensburg, D. A 2-step

PCR was performed as follows: initial denaturation at

95 �C for 2 min, followed by seven cycles of 45 s at

94 �C, 45 s at 68 �C (with each cycle the annealing

temp decreases 2 �C), and of 60 s at 72 �C. Products

were subsequently amplified in the second step for 30

cycles at 94 �C for 45 s, 54 �C for 45 s, and 72 �C for

60 s, with a final extension at 72 �C for 5 min. When

primers did not work in this PCR program, final

annealing temperature was reduced to 48 �C. PCR

products were separated using 10 % polyacrylamide

gels, 1 9 TBE buffer in a C.B.S. electrophoresis

chamber (C.B.S. Scientific Inc., Del Mar, CA, USA).

Electrophoresis conditions were set at constant 400 V

and 10 �C for 2 h. Gels were scanned by Typhoon (GE

Healthcare, Uppsala, Sweden) in fluorescent mode.

Genet Resour Crop Evol (2013) 60:1531–1546 1533

123

Author's personal copy

Table 1 Classification and sources of seeds of 88 Cucurbita and one Cucumis accessions

Cucurbita taxon Depository Accession designation Geographical origin

argyrosperma var. argyrosperma PGRC PI 163591 Guatemala

argyrosperma var. argyrosperma PGRC PI 438544 Belize, Corozal

argyrosperma var. argyrosperma PGRC PI 438703 Mexico

argyrosperma var. argyrosperma PGRC PI 442235 Mexico

argyrosperma var. argyrosperma PGRC PI 458724 Argentina

argyrosperma var. argyrosperma PGRC PI 512114 Nicaragua

argyrosperma subsp. sororia PGRC PI 318831 Mexico

argyrosperma subsp. sororia PGRC PI 438833 Mexico

argyrosperma subsp. sororia PGRC PI 438834 Mexico

argyrosperma subsp. sororia PGRC PI 512214 Mexico, Guerrero

argyrosperma subsp. sororia PGRC PI 532408 Mexico, Chiapas

argyrosperma var. palmeri PGRC PI 512199 Mexico, Sonora

argyrosperma var. palmeri PGRC PI 512227 Mexico, Sonora

argyrosperma var. palmeri PGRC PI 512233 Mexico, Sonora

argyrosperma var. stenosperma PGRC Grif 9453 Mexico, Veracruz

argyrosperma var. stenosperma PGRC PI 326184 Mexico

argyrosperma var. stenosperma PGRC PI 532339 Mexico, Oaxaca

ecuadorensis PGRC Grif 9446 Ecuador

ecuadorensis PGRC PI 390455 Ecuador

ecuadorensis PGRC PI 432441 Ecuador

ecuadorensis PGRC PI 432443 Ecuador

ecuadorensis PGRC PI 540895 USA, California

ficifolia PGRC Grif 9448 UK

ficifolia PGRC Grif 13898 Mexico, Chiapas

ficifolia PGRC PI 451832 Guatemala

ficifolia PGRC PI 512680 Spain

foetidissima NALPGRU PI 435094 USA, California

foetidissima NALPGRU PI 442200 Mexico, Guanajuato

foetidissima NALPGRU PI 512095 Mexico, Zacatecas

foetidissima NALPGRU PI 540902 USA, Texas

foetidissima NALPGRU W6 7115 USA, Arizona

lundelliana PGRC PI 438697 Mexico

lundelliana PGRC PI 489690 Guatemala

Lundelliana PGRC PI 540898 Honduras

lundelliana PGRC PI 636138 Belize

maxima subsp. andreana NERPIS G 5285 –

maxima subsp. andreana NERPIS G 29253 –

maxima subsp. andreana NERPIS PI 458653 Argentina

maxima subsp. maxima IFA CHI 0011 China (market)

maxima subsp. maxima IFA CHI 002 (flat, white fruit) China (market)

maxima subsp. maxima IFA CHI 003 (flat, gray-green fruit) China (market)

maxima subsp. maxima IFA CHI 004 (globular, white fruit) China (market)

maxima subsp. maxima IFA CHI 005 (elongate, white fruit) China (market)

maxima subsp. maxima IFA CHI 006 (globular, red fruit) China (market)

maxima subsp. maxima IFA CHI 007 (elongate, gray-green fruit) China (market)

1534 Genet Resour Crop Evol (2013) 60:1531–1546

123

Author's personal copy

Table 1 continued

Cucurbita taxon Depository Accession designation Geographical origin

maxima subsp. maxima IFA CHI 008 (elongate, dark green fruit) China (market)

maxima subsp. maxima IFA CHI 009 (flat, gray fruit) China (market)

maxima subsp. maxima NERPIS PI 244704 Brazil, Sao Paulo

maxima subsp. maxima NERPIS PI 458662 Argentina

maxima subsp. maxima NERPIS PI 458679 Argentina, Mendoza

maxima subsp. maxima NERPIS PI 470953 Peru

maxima subsp. maxima NERPIS PI 475749 Paraguay

maxima subsp. maxima NERPIS PI 518681 Bolivia

moschata IFA CHI 010 (globular, dark green fruit)1 China (market)

moschata IFA CHI 011 Huo-Teng China (market)

moschata IFA CHI 012 Qing No. 19 China (market)

moschata IFA CHI 013 China (market)

moschata IFA CHI 014 Qing No. 32 Henan Prov. China (market)

moschata IFA CHI 015 Qing No. 4 Taiwan China (market)

moschata IFA CHI 016 Shan Dong Prov. China (market)

moschata IFA CHI 017 Jin No. 3 Shan Dong Prov. China (market)

moschata IFA CHI 018 Qing No. 28 Shan Dong China (market)

moschata PGRC Grif 14225 Guatemala

moschata PGRC PI 162889 Paraguay

moschata PGRC PI 200736 El Salvador

moschata PGRC PI 209116 Puerto Rico

moschata PGRC PI 264551 Guatemala

moschata PGRC PI 369346 Costa Rica

moschata PGRC PI 406848 Honduras

moschata PGRC PI 438552 Belize

moschata PGRC PI 438732 Mexico

moschata PGRC PI 442271 Mexico

moschata PGRC PI 498429 Colombia

okeechobeensis subsp. martinezii PGRC PI 438698 Mexico

okeechobeensis subsp. martinezii PGRC PI 512099 Mexico, San Luis Potosi

okeechobeensis subsp. martinezii PGRC PI 512103 Mexico, Veracruz

okeechobeensis subsp. martinezii PGRC PI 512106 Mexico, Veracruz

okeechobeensis subsp. martinezii PGRC PI 532363 Mexico, San Luis Potosi

okeechobeensis subsp. okeechobeensis PGRC PI 561112 USA, Florida

okeechobeensis subsp. okeechobeensis PGRC PI 561117 USA, Florida

okeechobeensis subsp. okeechobeensis PGRC PI 561119 USA, Florida

pepo subsp. fraterna NCPIS PI 532356 Mexico, Tamaulipas

pepo subsp. pepo NYRC True French (Zucchini) UK

pepo subsp. pepo NYRC Wies 371 (Oil-pumpkin) Austria

pepo subsp. texana NYRC Gill’s Golden Pippin (Acorn) USA

pepo subsp. texana NYRC Golden Bush Scallop USA

pepo subsp. texana NYRC Jack-Be-Little (Acorn) USA

pepo subsp. texana NYRC Straightneck Early Yellow USA

Genet Resour Crop Evol (2013) 60:1531–1546 1535

123

Author's personal copy

Data analysis

Primers were selected by testing them on a set of 12

accessions belonging to four species, C. pepo,

C. moschata, C. maxima, and C. ecuadorensis. Only

those primers that rarely, if ever, produced difficult to

score, very weak bands, were selected. Amplified

fragments of each allele of each SSR locus were

scored as ‘‘1’’ (present) or ‘‘0’’ (absent). The resulting

binary matrix was converted to the required data input

format according to the manual of NTSYSpc version

2.11S (Exeter Software, Setauket, NY, USA, Septem-

ber 2000) program.

Genetic similarity among accessions, based on the

presence or absence of alleles, was calculated using the

Dice (Nei and Li 1979) coefficient of similarity/

dissimilarity. Genetic distance (GD) values were cal-

culated as 1- (one minus) the coefficient of similarity.

Principal coordinate analysis (PCOORDA) was carried

out employing the Dice genetic distance matrix and the

DCENTER and EIGEN program options of NTSYSpc.

The distribution of the 88 accessions of Cucurbita was

depicted in a three-dimensional scatter plot using the

first three principal coordinates. From the similarity

matrix, a dendrogram was constructed with the UP-

GMA clustering method. The robustness of the phe-

netic tree was evaluated by bootstrap analysis with

1,000 replicates using the bootstrap function of the

FreeTree program (Hampl et al. 2001).

Results

The 74 SSR loci yielded a total of 315 alleles, with an

average of 4.3 alleles per locus, ranging from one

allele (monomorphic) at three loci to nine at one locus.

Over 80 % of the loci had three to six alleles. The loci

were distributed in 17 of the 20 linkage groups,

ranging from 1 to 7 per linkage group with an average

of 2.9 (Table 2).

The number of polymorphic SSR loci differed

considerably among the species, from 64 loci in

C. pepo to only nine in C. ficifolia (Table 3). Of the 31

null alleles, the plurality, 10, occurred in C. ecuador-

ensis; there were no null alleles in C. pepo and

C. argyrosperma. Fifteen of the Cucurbita SSRs

(20.3 %), produced a band in the single Cucumis

sativus accession, and three of these loci had two

alleles.

Genetic distance values

The mean genetic distance values (GDs) among

Cucurbita species had a broad range, from 0.37 to

0.78 (Table 4). The xerophytic C. foetidissima had the

highest average GD to the other Cucurbita species,

0.73. C. okeechobeensis had the lowest average GD to

the eight other species, 0.61. The species that were

closest to one another were C. okeechobeensis

and C. lundelliana (GD = 0.37), C. ecuadorensis and

C. maxima (GD = 0.50), and C. argyrosperma and

C. moschata (GD = 0.56).

The mean GDs among accessions within the same

species were markedly lower than among species

(Table 4). Within species, mean GDs were particu-

larly low, 0.07, for C. ecuadorensis and C. ficifolia.

The accessions of C. pepo had by far the highest mean

GD, 0.46; most outstanding was the oil-seed pumpkin,

‘Wies 371’, its mean GD to conspecific accessions

being 0.62, the extreme of 0.66 being with the acorn

squash, ‘Gill’s Golden Pippin’. These values are

higher than those observed for the means of some

interspecific comparisons. In C. argyrosperma, acces-

sions derived from the same subspecies had lower

Table 1 continued

Cucurbita taxon Depository Accession designation Geographical origin

Cucumis sativus IFA Tulln Sensation F1 Austria

PGRC Plant Genetic Resources Conservation Unit, Griffin Georgia, USA, NALPGRU National Arid Land Plant Genetic Resources

Unit, Parlier, California, USA, NERPIS Northeast Regional Plant Introduction Station, Geneva, New York, USA, IFA Institute for

Biotechnology in Plant Production, Tulln, Austria; NCPIS North Central Plant Introduction Station, Ames, Iowa, USA, NYRC Newe

Ya‘ar Research Center, Agricultural Research Organization, Department of Vegetable Crops and Plant Genetics, Ramat Yishay,

Israel1 The IFA C. maxima accessions designated CHI 001through CHI 009 and the C. moschata accessions CHI 010 to CHI 018 were

purchased in China at local seed vendors. Information on fruit morphology of the C. maxima accessions and varietal names of the

C. moschata accessions were as given on the seed packets

1536 Genet Resour Crop Evol (2013) 60:1531–1546

123

Author's personal copy

mean GDs than accessions from different subspecies.

C. maxima subsp. andreana had little diversity, mean

GD = 0.08. In C. moschata, the mean GDs for the 11

accessions from the Americas and for the nine

accessions from China were 0.20 and 0.16, respec-

tively, but the mean GD across geographic origins was

double, 0.33.

Some particular accessions of one species had

rather low mean GDs to accessions of other species

(Online Resource 1). Most notable of all was PI

532356, a gourd of Cucurbita pepo subsp. fraterna

(Bailey) Lira, Andres et Nee, which had mean GDs

with four other species that were nearly identical to the

value it had (0.58) with the conspecific accession of

oil-seed pumpkin, C. pepo subsp. pepo ‘Wies 371’.

Two accessions had a markedly lower mean GD than

all others to the distant C. foetidissima, 0.63, one of

these being the oil-seed pumpkin.

Principal coordinate analysis and dendrogram

Principal coordinate analysis (Fig. 1) shows clear

separation of all Cucurbita species. C. okeechobeensis

has a central position, with all other species radially

arranged around it. The taxon nearest to C. okeech-

obeensis is C. lundelliana.

The dendrogram (Fig. 2) shows cucumber, Cuc-

umis sativus, as the most outlying accession. All of the

accessions of Cucurbita are clearly divided into

clusters according to species. The accessions of

Cucurbita foetidissima form a stark cluster outlying

in respect to all other Cucurbita. The cluster of

C. ficifolia is also outlying and has a position showing

no close relationship to any other species and the same

is apparent for the cluster formed by the seven

accessions of Cucurbita pepo. The clusters of the six

remaining species form three pairs, C. maxima with

C. ecuadorensis, C. okeechobeensis with C. lundelli-

ana, and C. moschata with C. argyrosperma.

Well-defined sub-clustering within species, with

bootstrap values of 100 %, is evident for C. maxima,

C. ecuadorensis, C. okeechobeensis, C. moschata,

C. argyrosperma, and C. foetidissima (Fig. 2). Some

sub-clusterings are in accordance with the differing

subspecific classifications, as in C. argyrosperma and

C. maxima, or geographical origins (Table 1). Notably,

in C. okeechobeensis, the accessions from Mexico and

Florida sub-cluster separately. In C. moschata, the

accessions of Chinese origin sub-cluster separately

from the ones originating in the Americas. Sub-

clustering is also evident in C. pepo, with the oil-seed

pumpkin being an outlier and the accessions of

C. pepo subsp. texana forming a sub-cluster.

Discussion

The SSR markers used in this investigation were

isolated from genomic DNA of two species, Cucurbita

pepo and C. moschata. These markers have been

shown to be approximately 90 % transferrable to other

Cucurbita species (Stift et al. 2004; Gong et al. 2008a)

even though only about 20 % amplified a band in

Cucumis sativus. There was no readily apparent

decline in primer efficiency associated with increased

genetic distance within the genus Cucurbita, likely

attributable to the derivation of the SSR markers,

C. moschata being the third most distant species to

C. pepo (GD = 0.69), after C. foetidissima (0.73) and

C. ficifolia (0.70) (Table 4). Indeed, among the five

accessions C. foetidissima, which is the most distant

species, 22 of the SSR alleles were observed to be

polymorphic (Table 3).

Among the 88 Cucurbita accessions examined, we

found only 31 null alleles, that is, primer pairs not

amplifying a fragment in any given accession

(Table 3). The highest number of null alleles, 10,

occurred in C. ecuadorensis, which is not one of the

most remote nor thought to be one of the most ancient

species of the genus (Robinson and Decker-Walters

1997). Of the 74 SSR loci, all but three were

polymorphic. The average of 4.3 alleles per locus

(Table 2) exceeds the average of 3.0 observed among

104 accessions of Cucurbita pepo (Gong et al. 2012).

This relatively low level of genus-wide polymorphism

for SSR loci is consistent with our previous observa-

tions within C. pepo and C. moschata (Gong et al.

2008a).

Our results are derived from 88 accessions of nine

species (Table 1) and thus any interpretations are

necessarily limited to this practically sized but none-

theless relatively small sample of the hundreds or

perhaps thousands of Cucurbita accessions in exis-

tence. Moreover, measure of genetic distance within a

population using SSR markers is positively influenced

by the size of that population. Therefore, genetic

variation within C. ficifolia, for example, of which

there were only four samples available, may be

Genet Resour Crop Evol (2013) 60:1531–1546 1537

123

Author's personal copy

underestimated in comparison to C. moschata, of

which there were 20 samples. Due to the nature of the

agglomeration procedure used in cluster analysis,

the dendrogram depicts GD-based relations between

the most similar individual accessions and merges

them into distinct branches (clades, species) of the

tree. Relations among branches, though, are shown

with less precision. In contrast, principal coordinate

analysis minimizes variance within clades (species)

while maximizing variance among them. Thus, the

dendrogram and principal coordinate analysis provide

differing perspectives of infrageneric relationships

within Cucurbita.

Cucurbita foetidissima, the only xerophytic species

included in this investigation, is the most distant and

outlying to the other Cucurbita (Table 4; Fig. 1). This

species does not produce fertile progeny when crossed

with the mesophytic species (Whitaker and Bemis

1964; Robinson and Decker-Walters 1997). Polymor-

phisms in allozymes (Puchalski and Robinson 1990)

and organellar DNA (Wilson et al. 1992; Sanjur et al.

2002) have shown that this xerophyte is most distant to

the mesophytes. Our results are consistent with the

understanding of Whitaker and Bemis (1964) that

C. foetidissima is a specialized derivative of a

mesophytic ancestral species.

Cucurbita pepo was the most outlying of the

mesophytic species (Fig. 2). Wilson et al. (1992)

obtained similar results but the general consensus had

been that C. pepo is loosely allied with C. argyrosperma

and C. moschata, based on morphology and crossability

(Whitaker and Davis 1962; Robinson and Decker-

Walters 1997) as well as polymorphisms observed in

allozymes and mitochondrial DNA (Decker-Walters

et al. 1990; Sanjur et al. 2002). C. pepo contains

hundreds of cultivated and wild accessions. The seven

accessions chosen for the present investigation were

known to be quite divergent (Gong et al. 2012), but this

alone should not have contributed to the high average

GD of C. pepo to the other species, 0.67, or its outlying

position to them (Table 4; Fig. 2). GDs within C. pepo

are larger than in all other species, even though three

others were represented by more than double the

number of accessions (Table 1). The high within-

species GD for C. pepo, 0.46, is likely a reflection of

its 10,000 year-old history as a domesticate, the oldest

of the genus, and the most intensive and divergent

selection to which it has been subjected, for use of

immature fruits and for use of mature fruits or exclu-

sively for seed oil (Paris 2000).

Table 2 The 74 SSR markers used in this study and their map

location (Gong et al. 2008a, b)

Linkage

group

Markers

1 –

2 CMTm48

3 CMTp55, CMTm102, CMTp131, CMTm42

4 –

5 CMTp88, CMTp129, CMTp235, CMTp175,

CMTp254, CMTp46, CPGA78*

6 CMTm130, CMTp224

7 CMTmC11, CMTp248

8 CMTm66, CMTm77, CMTp142, CMTp53

9 CMTm261

10 CMTm54, CMTp145, CMTp66, CMTm183

11 CMTm29, CMTp243, CMTp245, CMTm111

12 CMTp141

13 CMTp68

14 CMTm65, CMTp176, CMTm6

15 CMTp169, CMTm259, CMTm87

16 CPGA67*, CMTp125, CMTp231, CMTp107

17 CMTp210, CMTp209

18 CMTm49, CMTmC21

19 CMTp132

20 –

Not

mapped

CMTm14, CMTm21, CMTm110, CMTm112,

CMTm113, CMTm127, CMTm128,

CMTm132, CMTmC17, CMTmC20,

CMTmC32, CMTmC60, CPGA8*k, CMTp37,

CPGA51*, CMTp49, CMTp80, CMTp83,

CPGA101*, CMTp109, CMTp126, CMTp127,

CPGA160*, CMTp202, CMTp223,

CPAAC18*, CMTp252, CPCCA9*

Table 3 Allele distribution among Cucurbita species

Species No.

accessions

No. alleles

Polymorphic Monomorphic Null

Argyrosperma 17 49 25 0

Ecuadorensis 5 12 52 10

Ficifolia 4 9 59 6

Foetidissima 5 22 45 7

Lundelliana 4 19 63 2

Maxima 18 28 43 3

Moschata 20 45 28 1

Okeechobeensis 8 23 49 2

Pepo 7 64 10 0

1538 Genet Resour Crop Evol (2013) 60:1531–1546

123

Author's personal copy

Cucurbita ficifolia is another outlying cultivated

species (Fig. 2). Like C. foetidissima, C. ficifolia has a

specialized ecological adaptation. C. ficifolia is grown

at high altitudes, in the Americas from northern

Mexico to northern Argentina and Chile. It has not

been found growing wild (Hernandez and Leon 1994).

The extent of its traditional cultivated distribution in

the Americas exceeds that of all other Cucurbita

species (Nee 1990), but it exhibits surprisingly little

variation in fruit shape, size, and color (Andres 1990),

and the least SSR polymorphism, having a within-

species GD of 0.07 (Tables 1, 4). From its distinct

morphology and lack of cross-compatibility with other

species, C. ficifolia had been considered to be the most

isolated of the five domesticates (Andres 1990).

Similar to the results obtained using organellar DNA

(Wilson et al. 1992; Sanjur et al. 2002), SSR

polymorphisms also place C. ficifolia in an isolated

position to the other mesophytic Cucurbita species

(Fig. 2). C. ficifolia is far closer, though, to the South

Table 4 Genetic distances within (diagonal) and among Cucurbita species, and Cucumis sativus, based on SSR markers

Cucurbita species Number of accessions Foe Arg Ecu Fic Lun Oke Max Mos Pep

Foe 5 0.17

Arg 17 0.70 0.32

Ecu 5 0.69 0.67 0.07

Fic 4 0.71 0.65 0.65 0.07

Lun 4 0.77 0.64 0.63 0.70 0.21

Oke 8 0.74 0.61 0.61 0.68 0.37 0.14

Max 18 0.73 0.65 0.50 0.60 0.61 0.60 0.15

Mos 20 0.78 0.56 0.70 0.70 0.64 0.60 0.63 0.26

Pep 7 0.73 0.62 0.66 0.70 0.67 0.63 0.67 0.69 0.46

Averagea (Total 88) 0.73 0.64 0.64 0.67 0.63 0.61 0.62 0.66 0.67

Cucumis sativus 0.93 0.89 0.84 0.83 0.88 0.89 0.85 0.89 0.83

Foe Cucurbita foetidissima, Arg Cucurbita argyrosperma, Ecu Cucurbita ecuadorensis, Fic Cucurbita ficifolia, Lun Cucurbitalundelliana, Oke Cucurbita okeechobeensis, Max Cucurbita maxima, Mos Cucurbita moschata, Pep Cucurbita pepoa Average genetic distance of the accessions of one species to all other species of Cucurbita

Fig. 1 Principal coordinate

analysis of the 88 accessions

of Cucurbita, with

groupings labeled by species

Genet Resour Crop Evol (2013) 60:1531–1546 1539

123

Author's personal copy

1540 Genet Resour Crop Evol (2013) 60:1531–1546

123

Author's personal copy

American C. maxima, GD = 0.60, than it is to any

other species (Table 4), consistent with the Andean

origin proposed by Andres (1990) and Nee (1990) for

C. ficifolia.

Cucurbita maxima and C. ecuadorensis, two South

American species, are closely related, having a mean

GD of 0.50 (Table 4) and forming sister clusters

(Fig. 2). The close relationship seen here is consistent

with results obtained using mitochondrial DNA

(Sanjur et al. 2002). Their next nearest relative is

C. okeechobeensis (GD = 0.60 and 0.61, respec-

tively) and, in the case of C. maxima, C. ficifolia

(GD = 0.60). Diversity among all 18 accessions of

C. maxima, GD = 0.15, was unexpectedly low, espe-

cially given that nine of the C. maxima accessions

were local cultivars collected in China, six were from

several countries in the Americas, and three were of

subsp. andreana. The C. maxima cluster consisted of

two sub-clusters, in accordance with subspecific

designations. Within the subsp. maxima sub-cluster,

the cultivars from China and South America are well-

integrated, suggesting that the C. maxima accessions

in Chinese horticulture are neither derived from a

single introduction nor has this species undergone

intensive selection in China. The three accessions of

subsp. andreana, the wild ancestor (Whitaker 1951;

Nee 1990), formed a sub-cluster separate and outlying

to the 15 accessions of the cultivated subsp. maxima

(Fig. 2). The C. maxima subsp. andreana accessions

did not show a special affinity to C. ecuadorensis, as

the mean GD between these two taxa was 0.51 (Online

Resource 1).

Another closely related pair of Cucurbita species is

C. argyrosperma and C. moschata, their mean GD

being 0.56 (Table 4) and they form sister clusters

(Fig. 2). Their close relationship has also been viewed

through variations in organellar DNA (Wilson et al.

1992; Sanjur et al. 2002). These two species have great

morphological similarity such that it can be difficult to

distinguish between them (Nee 1990). Moreover,

under field conditions, they may cross with each other

(Wessel-Beaver 2000b; Montes-Hernandez and Egui-

arte 2002). Whitaker and Bemis (1975) considered

C. moschata to be the ‘‘nearest to the common

ancestor of the genus’’, but its average GD to all other

species, 0.66, is not low and its average GD to the

other cultivated species, 0.65, is about the same as that

of C. argyrosperma (0.62) and C. maxima (0.64)

(Table 4). Both, C. moschata and C. argyrosperma,

have a considerable amount of polymorphism for

SSRs, with mean GDs of 0.32 and 0.26, respectively,

about double that possessed by C. maxima but

markedly less than that of C. pepo (Table 4). Lira

and Montes (1994), Wessel-Beaver (2000a), and

Andres (2004a, b) have indicated that C. moschata is

far more polymorphic than generally appreciated and

Merrick (1995) has discussed the wide infraspecific

variation of C. argyrosperma. In C. moschata, the

accessions from China formed a sub-cluster separate

from those of the Americas (Fig. 2). Evidently, the

samples from China are all derived from a single or a

few closely related accessions taken there centuries

ago and not represented in the American samples of

C. moschata that were included in this experiment. In

C. argyrosperma, sub-clustering occurred according

to the subspecific taxa designated by Merrick and

Bates (1989), but var. stenosperma showed a closer

relationship to var. palmeri than either showed to var.

arygrosperma and subsp. sororia (Fig. 2). As subsp.

sororia, the wild ancestor, is widely distributed

(Merrick and Bates 1989), the clustering suggests that

subsp. argyrosperma var. palmeri and var. stenosper-

ma are early offshoots of subsp. sororia, and var.

argyrosperma is a separate, later domesticate. The

nearest relative of C. moschata and C. argyrosperma is

C. okeechobeensis, with which they have mean GDs of

0.61 and 0.60, respectively (Table 4).

The two most centrally placed species in our

investigation are Cucurbita okeechobeensis and

C. lundelliana (Figs. 1, 2). C. okeechobeensis had

the lowest average GD to all other species of the genus

that we have included in this study, 0.61, lower than

that of C. lundelliana, 0.63 (Table 4). Based on its

crossability with other species, C. lundelliana was

considered by Whitaker (1956), Whitaker and Davis

(1962), and Whitaker and Bemis (1964) to be the most

centrally placed and most nearly ancestral to the

cultivated species. However, crossing experiments

involving C. okeechobeensis have not been nearly as

systematic or numerous as those involving C. lundel-

liana (Whitaker and Cutler 1965; Robinson and

Decker-Walters 1997). Moreover, C. okeechobeensis

Fig. 2 Dendrogram of 88 accessions of Cucurbita and one of

Cucumis sativus. The labeling of the accessions corresponds

with the designations listed in Table 1. The designations are

addended with abbreviations of the subspecific nomenclature

and/or geographical origin of the respective accessions, as listed

in Table 1

b

Genet Resour Crop Evol (2013) 60:1531–1546 1541

123

Author's personal copy

subsp. martinezii is distributed in a long belt in eastern

Mexico and from the sub-tropical coastal plain to

temperate foothills, 1500 m elevation, at forest edges

of the Sierra Madre Oriental; C. lundelliana, however,

is known ‘‘only from the limestone plains of the

Yucatan peninsula at elevations near sea level’’ (Nee

1990).

Cucurbita okeechobeensis and C. lundelliana were

the closest pair of all species, with a GD of only 0.37

(Table 4). Chloroplast DNA variation has also indi-

cated a central position within the genus and close

relationship between C. okeechobeensis and C. lun-

delliana (Wilson et al. 1992). The GD between the two

species, nonetheless, is markedly higher than their

infraspecific GDs of 0.14 and 0.21, respectively

(Table 4). There have been attempts to cross

C. okeechobeensis and C. lundelliana. Whitaker and

Bemis (1964), Singh (1990), and Puchalski and

Robinson (1990) reported that they produced fertile

hybrids but gave no botanical description of the

hybrids and the two species have not been formally

united. Both taxa have gray-green seeds but they differ

in corolla color and leaf shape. Further attempts at

crossing the two species and scrutiny of their filial and

backcross offspring, particularly in regard to their

fertility, could resolve whether they are indeed the

same or different biological species.

The Okeechobee gourd, C. okeechobeensis, was

first described from Florida in the early twentieth

century by J. K. Small (Walters and Decker-Walters

1993). A gourd from Veracruz, Mexico was first

described and named C. martinezii by L. H. Bailey

later, in 1943. The synonymy of the two was

discovered by Robinson and Puchalski (1980) and

confirmed by Andres and Nabhan (1988). The Florid-

ian C. okeechobeensis subsp. okeechobeensis is

endangered (US Fish and Wildlife Service 2009)

and, in an isozyme survey, most of its alleles were

observed to be a subset of those found in the Mexican

C. okeechobeensis subsp. martinezii (Walters and

Decker-Walters 1993). Consistent with this observa-

tion, we found less SSR variation among the subsp.

okeechobeensis accessions (mean GD = 0.02) than

among the subsp. martinezii accessions (mean

GD = 0.15; Online Resource 1). Although the former

has to the present day been documented only from the

vicinity of Lake Okeechobee and along the St. Johns

River in Florida, the latter has been found to have a

considerably wider geographical range than that

known to Bailey, from northeastern Mexico to Oaxaca

in the south, usually in the vicinity of streams or in

orchards (Andres and Nabhan 1988; Nee 1990). Thus,

C. okeechobeensis has a disjunct geographic distribu-

tion, in Florida and along the east coast of Mexico.

C. okeechobeensis may have reached Florida by way

of oceanic dispersal (Schaefer et al. 2009). Alterna-

tively, it may have had a continuous distribution along

the east coast of North America that was split by post-

Ice Age glacial melting or it may have been diffused to

Florida by people (Andres and Nabhan 1988). Thus,

the lower mean GD within subsp. okeechobeensis and

its somewhat lower GD to the other species can be

interpreted as resulting from genetic drift and founder

effects.

Overall, the present results, taken in consideration

with previous knowledge gained from plant morphol-

ogy, crossability, and molecular genetic polymor-

phisms, favor an ultimate tropical, mesophytic

ancestor for the genus Cucurbita that is most similar

to C. okeechobeensis. One offshoot, which either

occurred earlier than all others or developed more

rapidly due to strong selection pressure, led to the

xerophytic C. foetidissima. Another early offshoot, for

adaptation to cooler ecogeographical regions, was

C. pepo. Yet another, also to more temperate climates

but southward, became C. maxima subsp. andreana, of

which an early offshoot led to the high-altitude

C. ficifolia and a later offshoot to C. ecuadorensis.

C. argyrosperma was a more recent offshoot from the

ancestral tropical mesophytes, with adaptation to

tropical as well as semi-arid environments, and

C. moschata is the cultivated species that is least

changed, retaining the ecological adaptation of the

wild progenitors. C. lundelliana is perhaps a newly

diverged ecotype from C. okeechobeensis.

The across-species GDs do, overall, correspond

with crossability and phenotype. Mean GDs are higher

in all cases across biological species, as summarized

by Nee (1990), than within them (Table 4). The five

domesticated species are reproductively isolated from

one another to different degrees (Whitaker and Davis

1962). Although fruits often develop from interspe-

cific crosses, usually few viable seeds can be obtained,

and in most of these cases, the progeny has impaired

development and fertility (Merrick 1995). The differ-

ent species are seen as separate entities in numerical

taxonomy based on phenotypic characteristics (Bemis

et al. 1970; Whitaker and Bemis 1975) and in

1542 Genet Resour Crop Evol (2013) 60:1531–1546

123

Author's personal copy

preferences of squash and gourd bees, pollinators

which co-evolved with Cucurbita (Hurd et al. 1971).

Relatively high success and fertility of progeny

have been observed for the cross combination of

C. argyrosperma and C. moschata, both in artificial

crosses and under field conditions, and there is

evidence for introgression under natural conditions

(Decker-Walters et al. 1990; Merrick 1995; Wessel-

Beaver 2000b; Montes-Hernandez and Eguiarte 2002;

Wessel-Beaver et al. 2004; Ortiz-Alamillo et al. 2007).

This interspecific combination has the lowest mean

GD, 0.56, across domesticated species (Table 4).

C. maxima and C. moschata, which have a mean GD

of 0.63, can be crossed to produce hybrids having

reduced fertility, the most widespread use of which

have been as rootstocks for other cucurbits (Edelstein

et al. 2004). There have not been many attempts to

cross C. maxima with C. ficifolia, but as these two

species have a mean GD of 0.60, they might meet with

a greater degree of success than the cross of C. maxima

with C. moschata. Crossability between species,

however, is not necessarily a polygenic character that

would be revealed by SSR polymorphisms. Wheat and

rye, for example, belong to different genera, yet their

crossability is determined by only two genes (Lein

1943). The underlying causes of interspecific incom-

patibility in Cucurbita are not adequately understood.

Whitaker (1954) observed in the meiotic metaphase

of the hybrid C. ficifolia 9 C. maxima subsp. andreana

a maximum of only 10 bivalents. In the same interspe-

cific cross, Weiling (1959) found a maximum of only

14 paired chromosomes. Weiling (1959), who studied

chromosome pairing in a number of interspecific

Cucurbita hybrids, suggested that Cucurbita (2n =

40) is actually an allo-tetraploid, having a basic

chromosome number of 10 and the genome constitution

of AABB. However, C. ficifolia and possibly

C. lundelliana possess a different second genome pair,

AACC and AAWW respectively. He also observed

chromosomal rearrangements such as translocations

and inversions, which possibly are also in part respon-

sible for reduced hybrid fertility. Weiling’s proposal is

consistent with the earlier finding of Hayase (1954),

who studied meiosis of haploid twin plants of

C. maxima, which he obtained from a cross of

C. maxima 9 C. moschata. In metaphase I, most

chromosomes remained unpaired, univalents, rarely a

rod bivalent or a trivalent occurred. Further support to

Weiling’s proposal came by cytogenetic studies on

Cucurbita by Groff (1966), cited by Singh (1979). Also

consistent with the polyploid nature of Cucurbita are

results of allozyme studies. While gene duplication in

diploid species leads to an increased number of

isozymes in one enzyme system, a substantial increase

in a number of enzyme systems suggests a polyploid

ancestry (Gottlieb 1982). Weeden (1984) studied the

number of expressed loci of eight isozyme systems in

Cucumis and Cucurbita. He observed that, while in

Cucumis sativus (2n = 14) 17 loci were expressed,

in five Cucurbita species at least 28 loci were con-

tributing isozymic forms. Subsequently, Weeden and

Robinson (1986) found, in the hybrid of C. maxima 9

C. ecuadorensis, six apparent duplications of allozyme

loci. Such duplications had also been reported by

Kirkpatrick et al. (1985). Like Cucurbita, Sicana Naud.

and possibly all other genera of the tribe Cucurbiteae

have a diploid chromosome number of 40 (2n = 40).

Thus, the polyploidy may be linked with the origin of

the tribe Cucurbiteae rather than with the origin of the

genus Cucurbita (Schaefer and Renner 2011).

The within-species GDs corresponded well with

known phenotypic, geographical, and molecular rela-

tionships. The cultivated species having by far the

least phenotypic variation, C. ficifolia (Andres 1990),

had the lowest within-species mean GD, 0.07, and the

cultivated species having the greatest phenotypic

variation, C. pepo (Paris 2000), had the highest, 0.46

(Table 4). Accessions of the same subspecific taxa had

lower GD values than accessions derived from differ-

ent subspecies and sub-clustered together, as is readily

apparent for C. argyrosperma, C. maxima, C. okeech-

obeensis and C. pepo (Online Resource 1; Fig. 2).

There are, however, some surprisingly low GD values

between particular accessions belonging to different

species (Online Resource 1). Some outstanding cases

involved C. pepo, the economically most important

species. The oil-seed pumpkin of C. pepo subsp. pepo

had a lower GD to C. foetidissima PI 540902, 0.60,

than it did to three of the other six C. pepo accessions;

this is astounding since C. foetidissima has the least

crossability with all other species, domestic and wild

(Whitaker and Davis 1962), has the most distinctive

plant morphology (DeVeaux and Shultz 1985), and is

seen to be overall the most distant to all other species,

with a mean GD of 0.73 (Table 4). The acorn squash,

C. pepo subsp. texana ‘Jack-Be-Little’, was unexpect-

edly close, with a GD of 0.54, to PI 209116 of

C. moschata. Strangest of all, the accession of C. pepo

Genet Resour Crop Evol (2013) 60:1531–1546 1543

123

Author's personal copy

subsp. fraterna, PI 532356, had mean GDs of less than

0.60 to four other species (not including C. moschata),

similar to its GD of 0.58 with the conspecific oil-seed

pumpkin. C. pepo subsp. fraterna has been crossed

successfully with C. argyrosperma experimentally

and perhaps these two taxa can genetically interact

with one another under natural conditions (Wilson

et al. 1994).

There are numerous reports describing combina-

tions of particular genotypes from different Cucurbita

species that result in considerably greater success in

interspecific crossing than others (Whitaker and Davis

1962; Merrick 1995; Robinson and Decker-Walters

1997). It remains to be determined whether or not the

low interspecific GD values obtained using SSR

marker polymorphisms can predict the potential

success that could be obtained through particular

interspecific-cross combinations (Online Resource 1),

but if they do, the implications for more efficiently

tapping the pool of genetic resources that is the dozen

species of Cucurbita are significant.

Cucurbita foetidissima has been identified as resis-

tant to a host of pathogens and pests affecting

Cucurbita, but is genetically isolated (Whitaker and

Davis 1962; Robinson and Decker-Walters 1997). Oil-

seed pumpkins of C. pepo are peculiar, given their

outlying positions in the species (Gong et al. 2012) and

the high GD values of ‘Wies 371’ to all present

conspecific representatives (Online Resource 1). The

oil-seed pumpkins are a specialized gene pool formed

in the 1870s when a recessive mutation resulting in

non-lignification of the seed coat was selected by

growers in the southeastern part of the then Austro-

Hungarian Empire, where pumpkin-seed oil had been

highly valued for centuries as a salad oil and for its

believed medicinal effects in the prevention of bladder

and prostate problems (Teppner 2000, 2004). A most

appropriate test, then, of the predicative value of GDs

derived from SSR polymorphisms in Cucurbita, both

from theoretical and practical considerations, would

be an attempt to cross the oil-seed pumpkin ‘Wies 371’

with C. foetidissima PI 540902, a combination which

had the unexpectedly low GD of only 0.60 (Online

Resource 1).

Powdery mildew is a devastating disease of

Cucurbita pepo and resistance to this disease was

introgressed from C. okeechobeensis subsp. martinezii

by using C. moschata as a genetic bridge (Contin and

Munger 1977; Jahn et al. 2002); this resistance has

been widely deployed commercially (Jahn et al. 2002;

Paris and Cohen 2002). DeVaulx and Pitrat (1979)

reported modest success in attempting to cross C. pepo

directly with C. okeechobeensis subsp. martinezii;

specifically, they obtained eight F1 plants with

impaired fertility and some F2 and BC1 progeny. The

present results show that C. okeechobeensis has a

central position in the genus Cucurbita (Figs. 1, 2),

suggesting that crossing this species directly with a

wide range of C. pepo accessions might encounter

a markedly greater degree of success. In addition, a

more intensive search and collection of the widely

distributed C. okeechobeensis subsp. martinezii pop-

ulations in Mexico may yield genotypes that are

particularly compatible with C. pepo and, as was

proposed by Munger (1976), widen the spectrum of

available resistances to various pathogens.

Acknowledgments We thank Kathleen Reitsma of the North

Central Plant Introduction Station, Ames, Iowa, USA; Robert

Jarret of the Plant Genetic Resources Conservation Unit, Griffin

Georgia, USA.; Allan Brown of the National Arid Land Plant

Genetic Resources Unit, Parlier, California, USA.; and Larry

Robertson of the Northeast Regional Plant Introduction Station,

Geneva, New York, USA., for providing seed samples used in

this investigation. This research was financially supported by the

Austrian Science Fund (FWF project No. P19662-B16) and by

the State of Lower Austria.

References

Andres TC (1990) Biosystematics, theories on the origin, and

breeding potential of Cucurbita ficifolia. In: Bates DM,

Robinson RW, Jeffrey C (eds) Biology and utilization of

the Cucurbitaceae. Cornell University Press, Ithaca,

pp 102–119

Andres TC (2004a) Diversity in tropical pumpkin (Cucurbitamoschata): a review of infraspecific classifications. In:

Lebeda A, Paris HS (eds) Progress in cucurbit genetics and

breeding research, Proceedings of Cucurbitaceae 2004, the

8th EUCARPIA meeting on cucurbit genetics and breed-

ing. Palacky University in Olomouc, Czech Republic,

pp 107–112

Andres TC (2004b) Diversity in tropical pumpkin (Cucurbitamoschata): cultivar origin and history. In: Lebeda A, Paris

HS (eds) Progress in cucurbit genetics and breeding

research, Proceedings of Cucurbitaceae 2004, the 8th

EUCARPIA meeting on cucurbit genetics and breeding.

Palacky University in Olomouc, Czech Republic,

pp 113–118

Andres TC, Nabhan GP (1988) Taxonomic rank and rarity of

Cucurbita okeechobeensis. Cucurbit Genet Coop Rep

11:83–85

Baranek M, Stift G, Vollmann J, Lelley T (2000) Genetic

diversity within and between the species Cucurbita pepo,

1544 Genet Resour Crop Evol (2013) 60:1531–1546

123

Author's personal copy

C. moschata and C. maxima as revealed by RAPD markers.

Cucurbit Genet Coop Rep 23:73–77

Bemis WP, Rhodes AM, Whitaker TW, Carmer SC (1970)

Numerical taxonomy applied to Cucurbita relationships.

Am J Bot 57:404–412

Contin M, Munger HM (1977) Inheritance of powdery mildew

resistance in interspecific crosses with Cucurbita martin-ezii. HortScience 12:397 (abstract)

Decker-Walters DS, Walters TW, Posluszny U, Kevan PG

(1990) Genealogy and gene flow among annual domesti-

cated species of Cucurbita. Can J Bot 68:782–789

DeVaulx RD, Pitrat M (1979) Interspecific cross between Cu-curbita pepo and C. martinezii. Cucurbit Genet Coop Rep

2:35

DeVeaux JS, Shultz EB Jr (1985) Development of buffalo gourd

(Cucurbita foetidissima) as a semiaridland starch and oil

crop. Econ Bot 39:454–472

Duchesne AN (1786) Essai sur l’histoire naturelle des courges.

Panckoucke, Paris

Edelstein M, Burger Y, Horev C, Porat A, Meir A, Cohen R

(2004) Assessing the effect of genetic and anatomic vari-

ation of Cucurbita rootstocks on vigour, survival and yield

of grafted melons. J Hort Sci Biotech 79:370–374

Ferriol M, Pico B, Nuez F (2001) Genetic variability in pumpkin

(Cucurbita maxima) using RAPD markers. Cucurbit Genet

Coop Rep 24:94–96

Ferriol M, Pico B, Nuez F (2003a) Genetic diversity of some

accessions of Cucurbita maxima from Spain using RAPD

and SRAP markers. Genet Resourc Crop Evol 50:227–238

Ferriol M, Pico B, Nuez F (2003b) Genetic diversity of a

germplasm collection of Cucurbita pepo using SRAP and

AFLP markers. Theor Appl Genet 107:271–282

Ferriol M, Pico B, Nuez F (2004a) Morphological and molecular

diversity of a collection of Cucurbita maxima landraces.

J Am Soc Hort Sci 129:60–69

Ferriol M, Pico B, Fernandez de Cordova P, Nuez F (2004b)

Molecular diversity of a germplasm collection of squash

(Cucurbita moschata) determined by SRAP and AFLP

markers. Crop Sci 44:653–664

Goldman A (2004) The compleat squash. Artisan, New York

Gong L, Stift G, Kofler R, Pachner M, Lelley T (2008a)

Microsatellites for the genus Cucurbita and an SSR-based

genetic linkage map of Cucurbita pepo L. Theor Appl

Genet 117:37–48

Gong L, Pachner M, Kalai K, Lelley T (2008b) SSR-based

genetic linkage map of Cucurbita moschata and its synteny

with Cucurbita pepo. Genome 51:878–887

Gong L, Paris HS, Nee MH, Stift G, Pachner M, Vollmann J,

Lelley T (2012) Genetic relationships and evolution in

Cucurbita pepo (pumpkin, squash, gourd) as revealed by

simple sequence repeat polymorphisms. Theor Appl Genet

124:875–891

Gottlieb LD (1982) Conservation and duplication of isozymes in

plants. Science 216:373–380

Groff DW (1966) A cytogenetic study of microsporogenesis in

selected Cucurbita species. Ph.D. dissertation, University

of Arizona, Tucson

Gwanama C, Labuschagne MT, Botha AM (2000) Analysis of

genetic variation in Cucurbita moschata by random

amplified polymorphic DNA (RAPD) markers. Euphytica

113:19–24

Hampl V, Pavlıcek A, Flegr J (2001) Construction and bootstrap

analysis of DNA fingerprinting-based phylogenetic trees

with a freeware program FreeTree: application to tricho-

monad parasites. Int J Syst Evol Microbiol 51:731–735

Havey MJ, McCreight JD, Rhodes B, Taurick G (1998) Dif-

ferential transmission of the Cucumis organellar genomes.

Theor Appl Genet 97:122–128

Hayase H (1954) Cucurbita crosses. V. Occurrence of a haploid

twin pair from a F1 progeny of C. maxima 9 C. moschata.

Jap J Breed 4:115–121

Hernandez JE, Leon J (1994) Neglected crops, 1492 from a

different perspective. FAO Plant Production and Protection

Series, no. 26, Rome. http://www.fao.org/docrep/T0646E/

T0646E00.htm

Hurd PD Jr, Linsley EG, Whitaker TW (1971) Squash and gourd

bees (Peponapsis, Xenoglossa) and the origin of the culti-

vated Cucurbita. Evolution 25:218–234

Jahn M, Munger HM, McCreight JD (2002) Breeding cucurbit

crops for powdery mildew resistance. In: Belanger RR,

Bushnell WR, Dik AJ, Carver TLW (eds) The powdery

mildews: a comprehensive treatise. APS Press, St. Paul,

pp 239–248

Katzir N, Mozes-Daube N, Danin-Poleg Y, Paris HS (2000)

Potential usefulness of SSR markers for studying infra-

specific variability in Cucurbita pepo. Cucurbit Genet

Coop Rep 23:71–72

Kirkpatrick KJ, Decker DS, Wilson HD (1985) Allozyme dif-

ferentiation in the Cucurbita pepo complex: C. pepo var.

medullosa vs C. texana. Econ Bot 38:289–299

Lebeda A, Widrlechner MP, Staub J, Ezura H, Zalapa J,

Kristkova E (2007) Cucurbits (Cucurbitaceae; Cucumisspp., Cucurbita spp., Citrullus spp.). In: Singh RJ (ed)

Genetic resources, chromosome engineering, and crop

improvement: vegetable crops. CRC Press, Boca Raton,

pp 271–376

Lein A (1943) Die genetische Grundlage der Kreuzbarkeit

zwischen Weizen und Roggen. Z Indukt Abstamm Verer-

bungslehre 81:28–61

Lira R, Montes S (1994) Cucurbits (Cucurbita spp.). In: Her-

nandez JE, Leon J (eds) Neglected crops: 1492 from a

different perspective. Rome, FAO, pp 63–77

Merrick LC (1995) Squashes, pumpkins and gourds, Cucurbita(Cucurbitaceae). In: Smartt J, Simmonds NW (eds) Evo-

lution of crop plants, 2nd edn. Longman Scientific &

Technical, London, pp 97–105

Merrick L, Bates DM (1989) Classification and nomenclature of

Cucurbita argyrosperma. Baileya 23:94–102

Montes-Hernandez S, Eguiarte LE (2002) Genetic structure and

indirect estimates of gene flow in three taxa of Cucurbita(Cucurbitaceae) in western Mexico. Am J Bot 89:

1156–1163

Munger HM (1976) Cucurbita martinezii as a source of disease

resistance. Veg Imprt Newsl 18:4

Naudin C (1856) Nouvelles recherches sur les caracteres spe-

cifiques et les varietes des plantes du genre Cucurbita. Ann

Sci Nat Bot IV 6:5–73, 3 pl

Nee M (1990) The domestication of Cucurbita. Econ Bot

44(3 suppl.):56–68

Nei M, Li WH (1979) Mathematical model for studying genetic

variation in terms of restriction endonucleases. Proc Natl

Acad Sci USA 76:5269–5273

Genet Resour Crop Evol (2013) 60:1531–1546 1545

123

Author's personal copy

Ortiz-Alamillo O, Garza-Ortega S, Sanchez-Estrada A, Tronc-

oso-Rojas R (2007) Yield and quality of the interspecific

cross Cucurbita argyrosperma 9 C. moschata. Cucurbit

Genet Coop Rep 30:56–59

Paris HS (2000) First two publications by Duchesne of Cucur-bita moschata (Cucurbitaceae). Taxon 49:305–319

Paris HS (2001) Characterization of the Cucurbita pepo col-

lection at the Newe Ya‘ar Research Center, Israel. Plant

Genet Resourc Newslett 126: Cover, 41–45

Paris HS (2007) The drawings of Antoine Nicolas Duchesne for

his natural history of the gourds. C. Erard (ed). Museum

National d’Histoire Naturelle, Paris

Paris HS, Cohen R (2002) Powdery mildew-resistant summer

squash hybrids having higher yields than their susceptible,

commercial counterparts. Euphytica 124:121–128

Paris HS, Yonash N, Portnoy V, Mozes-Daube N, Tzuri G,

Katzir N (2003) Assessment of genetic relationships in

Cucurbita pepo (Cucurbitaceae) using AFLP, ISSR, and

SSR markers. Theor Appl Genet 106:971–978

Pico B, Sifres A, Esteras C, Nuez F (2005–2006) Cucumis SSR

markers applied to the study of genetic diversity in the

Cucurbita genus. Cucurbit Genet Coop Rep 28–29:70–72

Puchalski JT, Robinson RW (1990) Electrophoretic analysis of

isozymes in Cucurbita and Cucumis and its application for

phylogenetic studies. In: Bates DM, Robinson RW, Jeffrey

C (eds) Biology and utilization of the Cucurbitaceae.

Comstock Publishing Associates, Ithaca, pp 60–76

Robinson RW, Decker-Walters DS (1997) Cucurbits. CAB

International, Wallingford

Robinson RW, Puchalski JT (1980) Synonymy of Cucurbitamartinezii and Cucurbita okeechobeensis. Cucurbit Genet

Coop Rep 3:45–46

Sanjur OI, Piperno DR, Andres TC, Wessel-Beaver L (2002)

Phylogenetic relationships among domesticated and wild

species of Cucurbita (Cucurbitaceae) inferred from a

mitochondrial gene: implications for crop plant evolu-

tion and areas of origin. Proc Natl Acad Sci USA

99:535–540

Schaefer H, Renner SS (2011) Cucurbitaceae. In: Kubitzki K

(ed) The families and genera of vascular plants, vol 10,

Eudicots. Springer, New York, pp 112–174

Schaefer H, Heibl C, Renner SS (2009) Gourds afloat: a dated

phylogeny reveals an Asian origin of the gourd family

(Cucurbitaceae) and numerous oversea dispersal events.

Proc R Soc B 276:843–851

Singh AK (1979) Cucurbitaceae and polyploidy. Cytologia

44:897–905

Singh AK (1990) Cytogenetics and evolution in the Cucurbit-

aceae. In: Bates DM, Robinson RW, Jeffrey C (eds) Biol-

ogy and utilization of the Cucurbitaceae. Comstock

Publishing Associates, Ithaca, pp 10–28

Stift G, Zraidi A, Lelley T (2004) Development and character-

isation of microsatellite markers (SSR) in Cucurbita spe-

cies. Cucurbit Genet Coop Rep 27:61–65

Teppner H (2000) Cucurbita pepo (Cucurbitaceae)—history,

seed coat types, thin coated seeds and their genetics. Phy-

ton (Horn) 40:1–42

Teppner H (2004) Notes on Lagenaria and Cucurbita (Cucur-

bitaceae)—review and new contributions. Phyton (Horn)

44:245–308

Torres Ruiz RA, Hemleben V (1991) Use of ribosomal DNA

spacer probes to distinguish cultivars of Cucurbita pepo L.

and other Cucurbitaceae. Euphytica 53:11–17

US Fish and Wildlife Service (2009) Okeechobee gourd

(Cucurbita okeechobeensis ssp. okeechobeensis) 5-Year

review: summary and evaluation. US Fish and Wildlife

Service, Vero Beach, Florida

Walters TW, Decker-Walters DS (1993) Systematics of the

endangered Okeechobee gourd (Cucurbita okeechobeen-sis: Cucurbitaceae). Syst Bot 18:175–187

Weeden NF (1984) Isozyme studies indicate that the genus

Cucurbita is an ancient tetraploid. Cucurbit Genet Coop

Rep 7:84–85

Weeden NF, Robinson RW (1986) Allozyme segregation ratios

in the interspecific cross Cucurbita maxima 9 C. ecuador-ensis suggest that hybrid breakdown is not caused by minor

alterations in chromosome structure. Genetics 114:593–609

Weiling F (1959) Genomanalytische Untersuchungen bei Kur-

bis (Cucurbita L.). Der Zuchter 29:161–179

Wessel-Beaver L (2000a) Evidence for the center of diversity of

Cucurbita moschata in Colombia. Cucurbit Genet Coop

Rep 23:54–55

Wessel-Beaver L (2000b) Cucurbita argyrosperma sets fruit in

fields where C. moschata is the only pollen source.

Cucurbit Genet Coop Rep 23:62–63

Wessel-Beaver L, Cuevas EH, Andres TC, Piperno DR (2004)

Genetic compatibility between Cucurbita moschata and

C. argyrosperma. In: Lebeda A, Paris HS (eds) Progress in

cucurbit genetics and breeding research, Proceedings of

Cucurbitaceae 2004, the 8th EUCARPIA meeting on

cucurbit genetics and breeding. Palacky University in

Olomouc, Czech Republic, pp 393–400

Whitaker TW (1951) A species cross in Cucurbita. J Hered

42:65–69

Whitaker TW (1954) A cross between an annual species and a

perennial species of Cucurbita. Madrono 12:213–217

Whitaker TW (1956) The origin of the cultivated Cucurbita. Am

Naturalist 90:171–176

Whitaker TW, Bemis WP (1964) Evolution in the genus

Cucurbita. Evolution 18:553–559

Whitaker TW, Bemis WP (1975) Origin and evolution of the

cultivated Cucurbita. Bull Torrey Bot Club 102:362–368

Whitaker TW, Cutler HC (1965) Cucurbits and cultures in the

Americas. Econ Bot 19:344–349

Whitaker TW, Davis GN (1962) Cucurbits. Interscience,

New York

Wilson HD (1989) Discordant patterns of allozyme and

morphological variation in Mexican Cucurbita. Syst Bot

14:612–623

Wilson HD, Doebley J, Duvall M (1992) Chloroplast diversity

among wild and cultivated members of Cucurbita(Cucurbitaceae). Theor Appl Genet 84:859–865

Wilson HD, Lira R, Rodriguez I (1994) Crop/weed gene flow:

Cucurbita argyrosperma Huber and C. fraterna L. H. Bai-

ley (Cucurbitaceae). Econ Bot 48:293–300

Wu J, Chang Z, Wu Q, Zhan H, Xie S (2011) Molecular

diversity of Chinese Cucurbita moschata germplasm col-

lections detected by AFLP markers. Sci Hort 128:7–13

1546 Genet Resour Crop Evol (2013) 60:1531–1546

123

Author's personal copy