Molecular Analysis of the Genetic Diversity of Chinese

52

J. Japan. Soc. Hort. Sci. 80 (1): 52–65. 2011.

Available online at www.jstage.jst.go.jp/browse/jjshs1

JSHS © 2011

Molecular Analysis of the Genetic Diversity of Chinese Hami Melon and Its

Relationship to the Melon Germplasm from Central and South Asia

Yasheng Aierken1,2, Yukari Akashi1, Phan Thi Phuong Nhi1, Yikeremu Halidan1,

Katsunori Tanaka3, Bo Long4, Hidetaka Nishida1, Chunlin Long4, Min Zhu Wu2

and Kenji Kato1*

1Graduate School of Natural Science and Technology, Okayama University, Okayama 700-8530, Japan2Hami Melon Research Center, Xinjiang Academy of Agricultural Science, Urumuqi 830000, China3Research Institute for Humanity and Nature, Kyoto 603-8047, Japan4Kunming Institute of Botany, CAS, Heilongtan, Kunming, Yunnan 650204, China

Chinese Hami melon consists of the varieties cassaba, chandalak, ameri, and zard. To show their genetic diversity,

120 melon accessions, including 24 accessions of Hami melon, were analyzed using molecular markers of nuclear

and cytoplasmic genomes. All Hami melon accessions were classified as the large-seed type with seed length longer

than 9 mm, like US and Spanish Inodorus melon. Conomon accessions grown in east China were all the small-

seed type. Both large- and small-seed types were in landraces from Iran, Afghanistan, Pakistan, and Central Asia.

Analysis of an SNP in the PS-ID region (Rpl16-Rpl14) and size polymorphism of ccSSR7 showed that the melon

accessions consisted of three chloroplast genome types, that is, maternal lineages. Hami melon accessions were

T/338 bp type, which differed from Spanish melon and US Honey Dew (T/333 bp type), indicating a different

maternal lineage within group Inodorus. The gene diversity (D), calculated from random amplified polymorphic

DNA (RAPD) and simple sequence repeat (SSR) polymorphism, was 0.476 in 120 melon accessions; the largest

diversity was in Central Asian accessions (D = 0.377) but was low for Hami melon accessions (D = 0.243), even

though Hami melon has diverse morphological traits, earliness, and shelf life. Reflecting such small genetic

diversity, Hami melon accessions of vars. ameri and zard were grouped into cluster II, except for one accession,

by the unweighted pair group method and the arithmetic mean (UPGMA) cluster analysis. Variety chandalak

with distinct characters, such as early maturing and poor shelf life, was assigned to clusters IV and VI, indicating

inter-varietal genetic differentiation within Hami melon. Three accessions from Turkmenistan and Afghanistan,

with large seeds and T/338 bp type of chloroplast genome, were classified as cluster II with Hami melon accessions

of vars. ameri and zard. We therefore concluded that Hami melon may have been transmitted from the west. The

small-seed type melon of group Conomon grown in east China may have been introduced into China independently

of Hami melon, because it had the A/338 bp type of the chloroplast genome and was clustered distantly from Hami

melon according to nuclear genome analysis.

Key Words: chloroplast genome, Cucumis melo, genetic diversity, Hami melon, SSR.

Introduction

Melon (Cucumis melo L.) is one of the most importanthorticultural crops in China, and is cultivated in 353 kha

with a total production of 9.7 million tons a year. Chinesemelon is generally divided into two types based on thethickness of the fruit skin: thick-skinned melon isclassified as Group Inodorus or Cantalupensis amongseven horticultural groups defined by Munger andRobinson (1991), and thin-skinned melon is classifiedas Group Conomon. These two types also differ incultivation area. Thick-skinned melon is mostlyproduced in Xinjiang (Xinjiang Uyghur AutonomousRegion), and thin-skinned melon in eastern China. Bothtypes are cultivated between these two areas in Qinghai,Gansu, and Shaanxi provinces.

Received; July 5, 2010. Accepted; August 12, 2010.

This study was partly supported by a Grant-in-Aid for International

Scientific Research of Ministry of Education, Science, Culture and

Sports, Japan (No. 19255009), and JSPS Asian CORE Program. This

is Contribution number 23 from the Sato Project of Research Institute

for Humanity and Nature (RIHN), Japan.

* Corresponding author (E-mail: [email protected]).

J. Japan. Soc. Hort. Sci. 80 (1): 52–65. 2011. 53

Xinjiang is in north-west China, and is one-sixth ofChina’s national land. It has a dry continental climatewith great extremes of winter and summer temperature.Rainfall is scant, and the annual precipitation is less than500 mm in north Xinjiang and less than 100 mm in southXinjiang. It is below 40 mm in Tulufan and Hami, famousfor producing Hami melon, in the eastern part ofXinjiang. The daily fluctuation of air temperature islarge, and the average difference between day and nighttemperature is from 11.3°C to 14.8°C depending on thearea. Other important features of the climate are strongsunshine and long sunshine duration. These conditionsare advantageous for producing sweet melon. Thick-skinned melon called Hami melon (“Hami Gua” inChinese) is cultivated in 41 kha with a total productionof 1.1 million tons a year.

Local landraces of Hami melon are rich in diversity,and 101 local landraces of melon were collected inXinjiang to study melon germplasm in the 1950s and1980s (Wu, 1982). These landraces were classified asvar. cassaba Pang Greb (round fruits with three or fivecarpels, one accession), var. chandalak Pang Greb (extraearly Guadan melon, five accessions), var. ameri PangGreb (summer melon, 70 accessions), and var. zard PangGreb (winter melon, 25 accessions) (Lin, 1991). Figure 1shows photographs of typical fruits of these varieties.Var. ameri consists of early maturing and mid-seasonmaturing groups. Similarly, var. zard includes autumnmelon and late maturing winter melon groups. Thesowing time is from the end of March (south Xinjiang)to the middle of April (north Xinjiang). The harvest ofvar. chandalak starts from 10 June, and the latestmaturing winter melon (var. zard) is harvested inOctober; the fruits of var. zard are stored until May ofthe next year. Cultivation of these types facilitates theyear-round supply of Hami melon. Among these types,

var. chandalak is consumed in Xinjiang and is notexported because this type is unsuitable for long-distancetransportation because of its short shelf life. However,var. zard and the mid-season maturing group of var.ameri have a long (> six months) to moderately long(one month) shelf life and are exported as Hami melonto foreign countries, as well as to other provinces ofChina. Taking such diversification into account, Pitratet al. (2000) proposed to classify varieties cassaba andzard as var. inodorus (group Inodorus), and the varietieschandalak and ameri as independent varieties. However,the classification was primarily based on morphologicaland physiological characters, and the genetic diversitywithin each variety, and the genetic relationship betweenvarieties, has remained unclear.

In the last decade, the genetic diversity andphylogenetic relationship between various groups ofcultivated melon have been studied using analysis ofmolecular polymorphism, which can be easily detectedusing various types of DNA markers, such as randomamplified polymorphic DNA (RAPD: Dhillon et al.,2007; López-Sesé et al., 2003; Luan et al., 2008; Nakataet al., 2005; Soltani et al., 2010; Staub et al., 2000, 2004;Stepansky et al., 1999; Tanaka et al., 2007; Yi et al.,2009), simple sequence repeats (SSRs: Dhillon et al.,2007; Garcia-Mas et al., 2004; López-Sesé et al., 2002;Monforte et al., 2003; Nakata et al., 2005; Staub et al.,2000), and amplified fragment length polymorphism(AFLP: Garcia-Mas et al., 2000; Yashiro et al., 2005).These results showed distinct genetic differentiationbetween subsp. agrestis and subsp. melo, which weredefined by Pitrat et al. (2000) and Jeffrey (2001): subsp.agrestis has short hairs on the ovary and subsp. melo

has long hairs on the ovary. Seed size is also an importantcharacter to distinguish these two subspecies. Small-seedtype melon groups Conomon and Agrestis with seedlength shorter than 9 mm were classified as subsp.agrestis, and large-seed type melon groups Catalupensis,Inodorus and Flexuosus with seed length over 9 mmwere classified as subsp. melo (Akashi et al., 2002).However, the genetic differentiation between varietiesis not clear within each subspecies, and subsp. melo

accessions of groups Catalupensis, Inodorus andFlexuosus have been clustered together (López-Seséet al., 2003; Soltani et al., 2010; Stepansky et al., 1999).In contrast, the geographical variation in eachhorticultural group is often significant, as indicated byLópez-Sesé et al. (2003) who detected clear differencesbetween Spanish landraces and US-EU improvedcultivars of Inodorus, although most Inodorus accessionsanalyzed in these studies were selected from Europe andUSA, and little attention was paid to accessions fromCentral Asia and China. Although Monforte et al. (2003)analyzed one accession each of varieties chandalak andameri, these accessions were from Russia andUzbekistan, respectively. Therefore, little is known aboutthe genetic diversity and genetic structure of HamiFig. 1. Photographs of fruits of four varieties of Hami melon.

Y. Aierken, Y. Akashi, P.T.P. Nhi, Y. Halidan, K. Tanaka, B. Long, H. Nishida, C. Long, M.Z. Wu and K. Kato54

melon, irrespective of its importance as a novel geneticresource for a wide range of shelf life, flavor, high sugarcontent, large fruit, etc. (Lin, 1991; Liu et al., 2004).

In Central Asian countries located to the west ofXinjiang, group Inodorus is mainly grown, and groupConomon vars. conomon and makuwa are mainly grownin the eastern part of China. Based on their geographicaldistribution and their similar morphological traits, suchas fruit shape and seed length, speculating that Hamimelon was introduced from Central Asia seemsreasonable. Akashi et al. (2002) also supported this ideaby indicating independent origins of Hami melon andChinese Conomon using isozyme analysis, Tanaka et al.(2006) studied the sequence polymorphism in thechloroplast genome to classify cytoplasmic types, thatis, the maternal lineage, of cultivated melon, and showedthat both Hami melon and Inodorus from Central Asiahave the same PS-ID sequence (T-type), which differedfrom that (A-type) of group Conomon. However,analysis of ccSSR markers showed contradictory results(Tanaka et al., 2006): the amplified fragment size ofccSSR7 was 333 bp in Inodorus accessions from Spain,USA, and Central Asia, and was 338 bp in Hami melonand group Conomon. These results indicated theimportance of analyzing the genetic structure of Inodorusmelon of different geographical origins to know theorigin and diversification of Hami melon.

Therefore, in this study, the genetic diversity of melonlandraces cultivated in areas from Iran to Xinjiang wasstudied using RAPD and SSR markers and chloroplastgenome markers PS-ID and ccSSR7. The geneticrelationship between melon landraces was analyzed fromDNA polymorphism data, and the origin and diversifi-cation of Hami melon are discussed.

Materials and Methods

Plant materials

This study analyzed 120 accessions of melon(C. melo), including landraces from Iran to China(Table 1, Fig. 2): 23 accessions of Hami melon fromXinjiang Uyghur Autonomous Region of China, 19accessions of small-seed type melon from eastern China(group Conomon vars. conomon and makuwa) andYunnan, 10 accessions each from Iran and Afghanistan,11 accessions from Pakistan, 17 accessions from CentralAsia (Russia, Kazakhstan, Turkmenistan, Uzbekistanand Tajikistan), and 9 accessions from Spain. Asimproved cultivars, one accession of Hami melon (X011released in 1986, Fig. 1) from Xinjiang UyghurAutonomous Region and 16 accessions from USA (var.inodorus; 8, var. cantalupensis; 8) were also used.

Also analyzed as reference accessions (RA) were‘Earl’s Favourite’, ‘Melon Cantalupo di Charentais’(group Cantalupensis), ‘Kinpyo’ (group Conomon var.makuwa) and ‘Karimori’ (group Conomon var.conomon). Accessions, ‘Kokand’, ‘Tendral o Invernalea Buccia Verde’, ‘Honey Dew’, and ‘Homegarden’ were

included as Central Asian accessions, Spanish acces-sions, US Inodorus, and US Cantalupensis, respectively,in this study, even though Tanaka et al. (2007) used themas reference accessions.

For DNA analysis, total DNA was extracted from oneplant of each accession, as described below. The 120accessions were classified into large-seed type (≥9.0 mm) and small-seed type (< 9.0 mm) based on theirseed length according to Akashi et al. (2002).

Seeds of these accessions were provided by the NorthCentral Regional Plant Introduction Station, Iowa StateUniversity (USDA-ARS), USA; National Institute ofVegetable and Tea Science (NIVTS), Japan; Institute ofPlant Genetics and Crop Plant Research (IPK), Germany.These accessions were cultivated in the field or in aglasshouse at Okayama University, Japan.

DNA extraction

Seeds were sown on filter paper and were grown at26°C in a 16 h light-8 h dark cycle at light intensity46.5 μmol·m−2·s−1. Ten-day-old seedlings were individu-ally ground in liquid nitrogen, and total DNA wasextracted using the procedure of Murray and Thompson(1980) with minor modifications.

Analysis of PS-ID and ccSSR7

As molecular markers to determine the cytoplasmictype, two chloroplast markers, SNP in the PS-ID region(Rpl16-Rpl14) and size polymorphism of ccSSR7, wereanalyzed using the same method as Tanaka et al. (2006).

The SNP (A/T) in PS-ID sequences (Nakamura et al.,1997) was analyzed using dCAPS primers. Thenucleotide sequence of each primer was: Psid2F—5'AAAAAAAAACAATTGCAGATTRAATT 3' (R = A orG) and Psid1R—5' AGCATTTAAAAGGGTCTGAGGT3'. PCR amplification was performed in a 10 μL mixturecontaining 50 ng genomic DNA, 1 μL PCR buffer(Sigma, USA: 10 mM Tris-HCl (pH 8.3), 50 mM KCl),2.5 mM MgCl2, 0.1 mM dNTP, 0.25 μM of each primerand 0.25 U Taq polymerase (Sigma, USA). Amplificationreactions were performed by using an i-Cycler (Bio-Rad,USA). The PCR cycle was: an initial denaturing step at95°C for 3 min, 35 PCR cycles at 95°C for 1 min, 52°Cfor 2 min, and 72°C for 2 min. The final extension stepwas at 72°C for 5 min. The PCR products were digestedwith Apo I (New England BioLabs, USA) and thenelectrophoresed on 3% agarose gel (GenePure LE; BMBio, Japan) at constant voltage 100 V using a horizontalgel electrophoresis system (Mupid-2; Cosmo Bio,Japan). Gels were stained with ethidium bromide andvisualized by illumination with UV light.

Another marker of the chloroplast genome, theconsensus chloroplast SSR marker, ccSSR7 (Chung andStaub, 2003), was used in this study because Tanaka etal. (2006) reported that ccSSR7 was polymorphic bydeletion of 5 bp (ATATT). The nucleotide sequence ofeach primer was: ccSSR7-F—5' CGGGAAGGGCTCG


Table 1. List of melon accessions analyzed in this study.

Accession

No.Country of origin Province/Cultivar Accession No.x Group/Varietyw Seed sizev PS-ID

ccSSR7

(bp)Clusteru

US 84 Iran Mazandaran PI 140666 — L T 338 IV

US 86 Iran Khorasan PI 140814 — L T 333 III

US 88 Iran West Azerbaijan PI 143231 — L T 333 III

US 90 Iran Esfahan PI 230185 — L T 338 Xa

US 415 Iran Kerman PI 137834 — L T 333 IV

US 417 Iran Mazandaran PI 140675 — L T 338 IV

US 422 Iran Tehran PI 211922 — L A 338 IX

US 423 Iran East Azerbaijan PI 211923 — S A 338 IX

US 425 Iran Fars PI 211942 — S A 338 IX

US 431 Iran — PI 351132 — L T 338 Ia

IPK 4 Afghanistan — CUM 254 Dudaim S, L A 338 IX

US 10 Afghanistan Balkh PI 125942 — L T 333 VIIa

US 14 Afghanistan Badakhshan PI 126050 — L T 333 IIa

US 19 Afghanistan Samangan PI 127534 — L T 338 IIa

US 21 Afghanistan Herat PI 212089 — L T 338 III

US 386 Afghanistan Takhar PI 126090 — S A 338 Xa

US 387 Afghanistan Jowzjan PI 126105 — S, L T 338 III

US 389 Afghanistan Ghazni PI 127550 — L T 338 III

US 390 Afghanistan Kabul PI 207478 — L T 338 IIa

US 392 Afghanistan Helmand PI 220515 — S T 338 VIIa

US 457 Pakistan Punjab PI 116824 — L T 338 III

US 458 Pakistan Punjab PI 123188 — S T 338 VIII

US 459 Pakistan Sind PI 124552 — L T 338 V

US 460 Pakistan Sind PI 124553 — L T 338 V

US 461 Pakistan Punjab PI 163211 — S T 338 V

US 462 Pakistan Punjab PI 217525 — S T 338 III

US 463 Pakistan Mingora PI 217945 — S T 338 VIII

US 464 Pakistan Punjab PI 218070 — S T 338 V

US 465 Pakistan Mingora PI 218071 — S T 338 V

US 466 Pakistan — PI 426629 Flexuosus S A 338 VI

US 467 Pakistan Skardu PI 532929 — L T 333 IIa

P 173 Russia, Altai region Altajskaja skorospelaja — — L T 338 VIIa

P 174 Russia, Altai region Barnaulka — — S T 338 VIIa

P 175 Russia, Ukrine Kolkhoznitsa — — L T 338 IV

US 95 Kazakhstan Imljskaja VIR 6809 PI 476342 — L T 338 Ib

US 531 Kazakhstan Alma Ata Ames 19036 — L T 338 Ib

US 119 Turkmenistan Zaami 672 VIR 40689 PI 476331 — L T 338 IIb

IPK 64 Turkmenistan — CUM 209 Agrestis S A 338 Xa

P 250 Turkmenistan — — — L A 338 VIIa

P 251 Turkmenistan — — — S A 338 VIIa

US 121 Uzbekistan Kokca 588 VIR 5149 PI 476333 — L T 333 IIc

US 122 Uzbekistan Sakor-polak 554 VIR 5883 PI 476337 — L T 333 III

P 118z Uzbekistan Kokand 930172 Inodorus L T 333 IIb

P 119 Uzbekistan Mirzuchulskaja 930177 Inodorus L T 333 IIb

P 120 Uzbekistan Ak-Urug 930190 var. ameri L T 333 IIb

IPK 61 Tajikistan Dushanbe CUM 333 — L T 333 IIc

IPK 63 Tajikistan Dushanbe CUM 334 var. chandalak L T 338 IIc

IPK 62 Tajikistan Dushanbe CUM 389 — L T 338 III

X 001 China Laoguniang — var. ameri (early maturing) L T 338 III

X 002 China Mizigua — var. ameri (early maturing) L T 338 IIa

X 003 China Kukeqi — var. ameri L T 338 IIb

X 005 China Paotaihong — var. zard L T 338 IIb

X 006 China Hongxincui — var. ameri L T 338 IIb

X 007 China Wanshudonggua — var. zard L T 338 IIb

X 008 China Kakeqie — var. ameri L T 338 IIb

X 009 China Bixiekexin — var. cassaba L T 338 IIb

X 010 China Kuche — var. ameri L T 338 IIb

X 011 China Huanghou — var. ameri L T 333 IIb

X 012 China Bawudong — var. ameri (early maturing) L T 338 IIb

X 013 China Xiekesu — var. ameri (early maturing) L T 338 IIb

X 014 China Kashi — var. ameri L T 338 IIb

X 021 China Baipicui — var. ameri (early maturing) L T 338 IIb

X 023 China Kalakusai — var. zard L T 338 IIb

X 024 China Kaernaishi — var. ameri L T 338 IIb


Table 1. Continued.

z Reference accessions used by Tanaka et al. (2007).y Tendral; Tendral o Invernale a Buccia Verde, Charentais; Melon Cantalupo di Charentais.x Accession No. at USDA, IPK, and NIVTS.w —; Local landraces whose horticultural group/variety was not specified.v S; Small-seed type, L; Large-seed type.u Cluster number shown in Figure 6.

Accession

No.Country of origin Province/Cultivar Accession No.x Group/Varietyw Seed sizev PS-ID

ccSSR7

(bp)Clusteru

X 030 China Wangwenxiang — var. ameri (early maturing) L T 338 IIb

X 031 China Heimeimaomijigan — var. zard L T 338 IIb

X 032 China Kutuerkukeqi — var. ameri L T 338 IIb

X 043 China Huangdanzi — var. chandalak L T 338 IV

X 044 China Qingpiqingrouguadanzi — var. chandalak L T 338 VI

X 045 China Reguadan — var. chandalak L T 338 VI

X 051 China Qingpiqingroukekouqi — var. zard L T 338 IIb

X 053 China Qingpihongroukekouqi — var. zard L T 338 IIb

P 117z Spain Tendraly 950076 Inodorus L T 333 Ic

US 102 Spain Zaragoza PI 512411 — L T 333 Ic

US 103 Spain Zaragoza PI 512413 — L T 333 Ic

US 104 Spain Cadiz PI 512462 — L T 333 Ic

US 105 Spain Lerida PI 512489 — L T 333 Ic

US 106 Spain Caceres PI 512501 — S, L T 333 Ic

US 107 Spain Badajoz PI 512510 — L T 333 Ib

US 108 Spain Valencia PI 512564 — L T 333 Ic

US 109 Spain Castellon de Plana PI 512581 — S, L T 333 Ic

P 73z USA Honey Dew 940325 Inodorus L T 333 Ia

P 74 USA Honey Dew 600011 Inodorus L T 333 Ia



US 7 USA Floridew NSL 20616 Inodorus (casaba type) L T 333 Ia

US 217 USA Golden Crenshaw NSL 5647 Inodorus (casaba type) L T 338 Ib

US 218 USA Golden Beauty Casaba NSL 5659 Inodorus (casaba type) L T 338 Ib

US 221 USA Sungold Casaba NSL 5709 Inodorus (casaba type) L A 338 Ia

P 67 USA Rocky Ford 920049 Cantalupensis L T 338 VIIb

P 68z USA Homegarden 600063 Cantalupensis L T 333 VIIb

P 69 USA Georgia 47 920047 Cantalupensis L T 338 VIIb

P 72 USA #58-21 600068 Cantalupensis L T 333 VIIb

P 93 USA SC108 590034 Cantalupensis L T 338 VIIb

P 107 USA Rio Gold 600015 Cantalupensis L T 338 VIIa

P 108 USA Hales Best 600021 Cantalupensis L T 338 VIIb

P 109 USA Spicy 940326 Cantalupensis L T 333 VIIb

C 28 China Xingtangmiangua — Conomon var. makuwa S A 338 Xc

P 83 China Mi-tang-tin 910008 Conomon var. makuwa S A 338 Xb

P 142 China Damiangua 760007 Conomon var. makuwa S A 338 Xb

P 143 China Shidaodaqinggua 760008 Conomon var. makuwa S A 338 Xb

P 144 China Shilinghuangjingua 780143 Conomon var. makuwa S A 338 Xb

P 147 China Wengua 910055 Conomon var. makuwa S A 338 Xb

P 153 China Chi-86-56 940178 Conomon var. makuwa S A 338 Xb

P 155 China Qianzhong-5 940184 Conomon var. makuwa S A 338 Xb

P 208 China Chi-87-12 2000121 Conomon var. makuwa S A 338 Xc

C 32 China Heipilengzisudigua — Conomon var. conomon S A 338 Xc

P 154 China Chi-86-61 940182 Conomon var. conomon S A 338 Xb

P 158 China Qingpilürouxianggua 940307 Conomon var. conomon S A 338 Xb

P 169 China Caigua — Conomon var. conomon S A 338 Xb

P 171 China Qingpicaigua — Conomon var. conomon S A 338 Xb

CYW 37 China Yunnan — Conomon S A 338 Xd





P 62z RA UK Earl’s Favourite 900074 Cantalupensis L T 338 IIa

P 94z

RA Italy Charentaisy

910049 Cantalupensis L T 338 VIIb

P 90z RA Japan Kinpyo 920007 Conomon var. makuwa S A 338 Xc

P 130z RA Japan Karimori 940333 Conomon var. conomon S A 338 Xc


KGCAG 3' (K = T or G) and ccSSR7-R—5' GTTCGAATCCCTCTCTCTCCTTTT 3'. The PCR mixture and PCRcycle were the same as for PS-ID analysis, but annealingwas performed at 56°C for 1 min. PCR products wereelectrophoresed on 10% nondenatured polyacrylamidegel at a constant voltage of 260 V. Gels were stained asabove.

RAPD and SSR analysis

Eighteen random primers (12-mer; Bex, Japan), whichwere selected for their ability to detect polymorphismby Tanaka et al. (2007), were used in this study (Table 2).The PCR mixture was the same as for PS-ID analysis,but the concentration of primers was 0.5 μM. The PCRcycle was: an initial denaturing step at 95°C for 3 min,40 PCR cycles at 93°C for 1 min, 40°C for 2 min, and

72°C for 2 min. The final extension step was at 72°Cfor 5 min. After amplification, electrophoresis and gelstaining were performed as for PS-ID analysis, but theconcentration of agarose gel was 1.5%.

For SSR analysis, as a preliminary experiment, sizepolymorhism was examined for large-seed type melonaccessions Hami 2 (China, group Inodorus), Kokand(Uzbekistan, group Inodorus), and Melon Cantalupo diCharentais (Italy, group Cantalupensis), and a small-seedtype melon accession SUD-4 (Sudan, group Agrestis).Sixteen SSR markers showing distinct stable polymor-phism were selected among 177 SSR markers developedby Danin-Poleg et al. (2000), Akashi et al. (2001), Chibaet al. (2003), Ritschel et al. (2004), and Fukino et al.(2007). Table 3 shows details of the SSR markers used.The method for ccSSR7 analysis was used for SSRanalysis.

Data analysis

Marker bands of RAPD were scored as 1 for presentand 0 for absent. For SSR, marker fragments were scoredbased on their size from smallest (1) to largest. Fromthese data, the polymorphic index content (PIC) wascalculated according to Anderson et al. (1993). Thegenetic similarity (GS) between accessions wascalculated as described by Apostol et al. (1993), and thegene diversity (D) within each group and genetic distance(GD) among groups were calculated as described byWeir (1996) and Nei (1972), respectively. A dendrogramwas constructed using the Phylip program with theunweighted pair group method and the arithmetic mean

Table 2. Eighteen random primers used in this study, with the size of polymorphic fragments and their polymorphic

index content (PIC).

*; No polymorphism among 24 Hami melon landraces.

PIC was indicated from the larger fragment, when two or three markers were produced.

Primer number Sequence

(5' → 3')

Size of polymorphic

fragments (bp)

Polymorphic index

content (PIC)

A07 GATGGATTTGGG 970* 0.358

A20 TTGCCGGGACCA 1100*, 800* 0.219, 0.255

A22 TCCAAGCTACCA 1520* 0.255

A23 AAGTGGTGGTAT 1200 0.489

A26 GGTGAGGATTCA 1400 0.439

A31 GGTGGTGGTATC 800 0.391

A39 CCTGAGGTAACT 2027 0.499

A41 TGGTAGGTAACT 1353, 1020, 930 0.460, 0.495, 0.231

A57 ATCATTGGCGAA 800 0.460

B15 CCTTGGCATCGG 600 0.391

B32 ATCATCGTACGT 900, 700 0.193, 0.080

B68 CACACTCGTCAT 1078 0.444

B71 GGACCTCCATCG 1220 0.483

B84 CTTATGGATCCG 700, 600, 550 0.499, 0.499, 0.455

B86 ATCGAGCGAACG 1500, 1350* 0.499, 0.049

B96 CTGAAGACTATG 850, 750 0.439, 0.489

B99 TTCTGCTCGAAA 1400* 0.375

C00 GAGTTGTATGCG 1350* 0.320

Fig. 2. Map of China and neighboring countries. Countries are named

whose melon landraces were examined in this study.


(UPGMA) method. Principal coordinate analysis (PCO:Gower, 1966) based on the genetic similarity matrix wasperformed to show multiple dimensions of each groupand the accessions in a scatter plot.

Results

Seed size

Seed size was from 4.8 mm to 16.0 mm amongaccessions, and 120 melon accessions were grouped intothe large-seed type (81 accessions) and small-seed type(35 accessions) (Table 1). The original seed samples offour accessions from Afghanistan and Spain were amixture of small and large seeds, which might reflectthe polymorphism in a farmer’s field or derive fromspontaneous hybridization in a farmer’s fields, and thus

they could not be grouped into the large- or small-seedtype. As expected, accessions of Hami melon, USInodorus, and US Cantalupensis were all large-seed type,and all accessions of Conomon from China and Japanwere small-seed type. However, both large- and small-seed types were in landraces from Iran, Afghanistan,Pakistan, and Central Asia.

PS-ID and ccSSR7 of chloroplast genome

Based on the SNP in the PS-ID region (Rpl16-Rpl14),120 melon accessions were grouped as A-type (31accessions) and T-type (89 accessions) (Tables 1 and 4).T-type accessions were further divided into two groupsby the size polymorphism of ccSSR7: 338 bp type (59accessions) and 333 bp type (30 accessions). All A-type

Table 3. Primer sequences of 16 SSR markers used in this study, with the expected size of PCR products and their polymorphic index content (PIC).

*; No polymorphism among 24 Hami melon landraces.

Marker F primer (5' → 3') R primer (5' → 3')Expected fragment

size (bp)

Polymorphic index

content (PIC)

ACS2-ms1 TCTTTTGTTCTTGGTTGTGAGT GATTGCTTTATTTTGAATCTTTTG 200 0.892

CMBR 2 TGCAAATATTGTGAAGGCGA ATCCCCACTTGTTGGTTTG 114 0.839

CMBR 12 ACAAACATGGAAATAGCTTTCA GCCTTTTGTGATGCTCCAAT 134 0.651

CMBR 22 TCCAAAACGACCAAATGTTCC ATACAGACACGCCTTCCACC 177 0.724

CMBR 53 GCCTTTTGTGATGCTCCAAT AAACAAACATGGAAATAGCTTTCA 134 0.607

CMBR 83 CGGACAAATCCCTCTCTGAA GAACAAGCAGCCAAAGACG 142 0.595

CMBR 105 TGGTAAGCATTTTGAAATCACTTTT TTCCAGACATCTAAAGGCATTG 139 0.673

CMBR 120 CTGGCCCCCTCCTAAACTAA CAAAAAGCATCAAAATGGTTG 167 0.582

CMN 04-03* ATCACAGAGACCGCCAAAAC GGTTGAAGATTGCGCTTGAT 218 0.575

CMN 04-07 GAAAGCATTAAATATGGCATTGG AAGCTTAACAGCTTCCAGGG 286 0.771

CMN 04-40 CACCTGACGATAGGGGTGTT AGTATTCGGGTTGGCAAAAA 212 0.552

CMN 08-22* CATCCTCCTCATCCTCCTCA ACGGATGAATCGGAACTTCA 223 0.406

CMN 08-90* CCACGCCCTCTATACCCATA GGGACTGTTGGGTTTTCTGA 210 0.341

CMN 21-41 GAGGAAATTTTGGAGTTTTTCAA TTCCAGACATCTAAAGGCATTG 281 0.527

CMN 22-16 CAGAGGAGGTGGAACTAACCA CCATTTTCAACCTCCCAAGA 233 0.609

CMN 61-44 TGTTGGAGTTTAATGAGGAAGGA AGAGAAGATGAATGGGGCAC 233 0.901

Table 4. Variation of chloroplast genome type among melon accessions with different geographical origin.

z Reference accessions (RA).y Chloroplast genome type was indicated by the combination of PS-ID SNP and ccSSR7, like A/338.x For accessions with large and small seeds mixed in original seed sample, the count was divided into large- and small-seed types.

Area/Group No. of

accessions

Large-seed type Small-seed type

A/338y T/338y T/333y Total A/338y T/338y T/333y Total

Iran 10 1 4 3 8 2 — — 2

Afghanistan 10 0.5x 4.5x 2 7 1.5x 1.5x — 3

Pakistan 11 — 3 1 4 1 6 — 7

Central Asia 17 1 7 6 14 2 1 — 3

Hami melon 24 — 23 1 24 — — — 0

Spain 9 — — 8 8 — — 1 1

US Inodorus 8 1 2 5 8 — — — 0

US Cantalupensis 8 — 5 3 8 — — — 0

Chinese Conomon 19 — — — 0 19 — — 19

RAz 4 — 2 — 2 2 — — 2

Total 120 3.5 50.5 29 83 27.5 8.5 1 37


accessions were 338 bp type. As a result, the 120 melonaccessions were classified into three cytoplasmic types(Table 4) by analysis of PS-ID and ccSSR7.

The frequency of the three cytoplasmic types differedbetween large- and small-seed types (χ2

= 65.97, df = 2,P < 0.01). Large-seed type accessions consisted mostlyof cytoplasmic types T/338 bp type and T/333 bp type.All accessions of Spanish melon and US Honey Dewwere T/333 bp type. Hami melon accessions wereT/338 bp type, except X011, showing a difference inmaternal lineage within group Inodorus, as well asbetween Hami melon and Chinese Conomon.

RAPD analysis

Eighteen primers provided 26 polymorphic markerbands of approximately 550–2027 bp in size, and theaverage number of marker bands produced by eachprimer was 1.44 (Table 2). Most polymorphic bands wereproduced by primers A39 and B84 (Fig. 3A), and theirmarker bands of 2027 bp and 700 bp (A39-2027 andB84-700, respectively) were amplified in 63 accessionsamong 120 melon accessions. They were followed bymarkers B84-600 and B86-1500, which were amplifiedin 58 and 62 accessions, respectively. In contrast, theleast polymorphic band was produced by primer B86,whose marker band of 1350 bp was amplified in 117accessions. Hami melon accessions were monomorphicfor seven markers (Table 2, asterisk).

The gene diversity was 0.376 for 120 melonaccessions, of which the largest diversity was for Iranaccessions (D = 0.288), except RA, followed by acces-sions of neighboring countries, such as Afghanistan,Pakistan, and Central Asia (Table 5). However, D waslow for Hami melon accessions (D = 0.194), which wassimilar to accessions of US Cantalupensis and ChineseConomon, indicating rather small genetic variations in

Hami melon, even though its morphological andphysiological traits are diverse.

SSR analysis

A total of 114 polymorphic marker bands wereamplified from 120 melon accessions using 16 primersets. The average number of polymorphic bands of eachprimer set, which was considered to represent the numberof alleles, was 7.13, and ranged from two (CMN 08-22)to 17 (CMN 61-44 and ACS2-ms1). The mostpolymorphic marker was CMN 61-44 (PIC = 0.901),followed by ACS2-ms1 (PIC = 0.892) (Fig. 3B). The leastpolymorphic marker was CMN 08-90 (PIC = 0.341),followed by CMN 08-22 (PIC = 0.406). Hami melonaccessions were monomorphic for three markers(Table 3, asterisk).

The gene diversity was 0.640 for 120 melonaccessions, and the largest diversity was for CentralAsian accessions (D = 0.522), followed by accessionsfrom neighboring countries, such as Iran, Afghanistan,and Pakistan (Table 5). However, D was low for Hamimelon accessions (D = 0.323), which was similar toSpanish and US Inodorus accessions. This result alsoindicated rather small genetic variations in Hami melon.

Analysis by combined RAPD and SSR

The gene diversity calculated by combining RAPDand SSR data was 0.476 for 120 melon accessions, andthe largest diversity was observed for Central Asianaccessions (D = 0.377), except RA, followed byaccessions from neighboring countries, such as Iran,Afghanistan, and Pakistan (Table 5). However, it waslow for Hami melon accessions (D = 0.243).

The GD between nine melon populations wascalculated from the frequency of RAPD and SSR markerbands in each population. The GD from Hami melonwas 0.089 (Central Asia) to 0.997 (Chinese Conomon)(Table 6). The genetic relationship between populations

Fig. 3. Polymorphism among melon landraces studied. (A) Random

amplified polymorphic DNA (RAPD) with B84 primer on 1.5%

agarose gel. (B) Simple sequence repeat (SSR) with primer

ACS2-ms1 on 10% nondenatured polyacrylamide gel. Arrows

show polymorphic bands. Lane M, 100 bp DNA ladder marker.

Table 5. Gene diversity within each population, calculated from

RAPD and SSR data, singly or combined.

z Reference accessions (RA).

Area/GroupNo. of

accessions

Gene diversity

RAPD SSR SSR + RAPD

Iran 10 0.288 0.518 0.376

Afghanistan 10 0.278 0.491 0.360

Pakistan 11 0.275 0.487 0.355

Central Asia 17 0.287 0.522 0.377

Hami melon 24 0.194 0.323 0.243

Spain 9 0.156 0.330 0.222

US Inodorus 8 0.147 0.338 0.219

US Cantalupensis 8 0.197 0.371 0.263

Chinese Conomon 19 0.182 0.373 0.255

RAz 4 0.293 0.516 0.378

All accessions 120 0.376 0.640 0.476


was visualized using UPGMA cluster analysis and PCOanalysis. Cluster I consisted of Chinese Conomon, whichis related distantly to other populations (Fig. 4). ClusterII was US Cantalupensis. Hami melon formed clusterIII together with landraces from Pakistan, Iran,Afghanistan, and Central Asia. Spanish accessions andUS Inodorus were classified as cluster IV. The geneticrelationship indicated by cluster analysis was clearlyreproduced on a PCO plot of the 1st and 3rd principalcoordinates that explained 57.3% and 9.8% of totalvariance, respectively (Fig. 5).

Genetic relationship among melon accessions

The GD between the 120 melon accessions wascalculated from the RAPD and SSR data. The averageGD was 0.480 and ranged from 0.024 to 0.881 (data notshown). The largest GD was recorded between US121(Uzbekistan) and P158 (Chinese Conomon). Thesmallest GD was between accession pairs X010 andX024 (Hami melon), CYW60 and CYW61 (Yunnan),and P74 and P76 (US Honey Dew). The GD, calculatedby combining the RAPD and SSR data, related to thosecalculated singly from RAPD data (r = 0.950, P < 0.01)and from SSR data (r = 0.895, P < 0.01). The correlationcoefficient between the GDs calculated singly fromRAPD and SSR data was 0.711 (P < 0.01).

To determine the genetic relationship between melonlandraces, a dendrogram was constructed based on GDvalues calculated by combining RAPD and SSR data(Fig. 6). The 120 accessions were grouped into 10 major

clusters. Four of these clusters were further divided into12 subclusters. Table 7 summarizes the number of melonaccessions classified into each cluster and subcluster.

Cluster I was divided into three subclusters, mostlyof US Inodorus and Spanish accessions. Cluster II wasalso divided into three subclusters; 20 of 24 Hami melonaccessions were classified in this cluster, together withlandraces from Afghanistan, Pakistan, and Central Asia.Nineteen accessions of Hami melon, of var. ameri, var.zard, or var. cassaba, were grouped in subcluster IIb,which also included one accession of Turkmenistan(T/338 bp type) and three accessions of Uzbekistan(T/333 bp type). Interestingly, reference accession ‘Earl’sFavourite’ was grouped into subcluster IIa, even thoughthis is a famous Japanese cultivar of Cantalupensis, theremaining accessions of which were grouped in clusterVII. Clusters III to VI included melon landraces fromIran, Afghanistan, Pakistan, and Central Asia. Fouraccessions of Hami melon were also classified in theseclusters; three of the four were var. chandalak. ClusterVII was also divided into two subclusters. AllCantalupensis accessions, except ‘Earl’s Favorite’, weregrouped in this cluster together with landraces fromCentral Asia and Afghanistan. Clusters VIII waslandraces from Pakistan and cluster IX was landraces

Table 6. Genetic distance between nine populations of melon of different geographical origin, calculated from RAPD and SSR analysis.

Area/Group Iran Afghanistan Pakistan Central Asia Hami melon SpainUS

Inodorus

US

Cantalupensis

Afghanistan 0.086 —

Pakistan 0.130 0.110 —

Central Asia 0.124 0.086 0.130 —

Hami melon 0.195 0.153 0.182 0.089 —

Spain 0.203 0.196 0.236 0.146 0.155 —

US Inodorus 0.182 0.206 0.250 0.142 0.183 0.146 —

US Cantalupensis 0.252 0.261 0.243 0.252 0.389 0.421 0.313 —

Chinese Conomon 0.606 0.744 0.669 0.815 0.997 1.036 0.996 0.581

Fig. 4. Genetic relationship between nine melon populations, shown

by UPGMA cluster analysis based on the GD.

Fig. 5. Distribution on the first and third principal coordinates of

nine melon groups.


Fig. 6. Genetic relationship between 120 melon accessions, including 24 accessions of Hami melon, shown by UPGMA cluster analysis based

on genetic distance calculated from RAPD and SSR. Table 1 explains the accession numbers within each cluster and subcluster.


from Iran and one accession of group Dudaim fromAfghanistan. Cluster X related distantly to other clusterswas divided into four subclusters, among which twosubclusters (Xb and Xc) comprised Chinese and Japaneseaccessions of Conomon vars. conomon and makuwa.Subcluster Xd consisted of Conomon accessions fromYunnan. Subcluster Xa included landraces from Iran,Turkmenistan, and Afghanistan.

Discussion

To show the genetic diversity of Hami melon, 16 SSRmarkers and 26 RAPD markers were used in this study.The gene diversity detected by SSR and RAPD analyseswas 0.640 and 0.376, respectively, for 120 melonaccessions (Table 5). The GD between melon accessionsalso differed depending on the marker system used. TheGD averaged 0.646 (SSR) and 0.379 (RAPD), and thevariation was 0.037 for SSR and 0.029 for RAPD (datanot shown). Such differences depending on the markersystem were also reported by Staub et al. (2000) andNakata et al. (2005), indicating that the ability of SSRmarkers to discriminate was better than that of RAPDmarkers. Polymorphism was successfully detected in allnine melon populations by five markers (CMN 04-07,CMN 61-44, CMBR 2, CMN 04-40, CMN 21-41) amongthe 16 SSR markers used (data not shown). For the firstthree markers especially, PIC was high and over 0.771(Table 3). These five SSR markers are useful for diversityanalysis of melon. Although SSR markers, CMBR 22,CMBR 83, CMBR 105, ACS2-ms1, and CMBR 53, werealso highly polymorphic, the first three markers weremonomorphic in US Cantalupensis, and the last twomarkers were monomorphic in Chinese Conomon.ACS2-ms1 was developed to detect the repeat numberpolymorphism of (TA) in the 5' flanking region of CMe-

ACS2 (Akashi et al., 2001). The amplicon size of ChineseConomon seemed unique and the same size of fragmentwas amplified only from two accessions from Iran andone accession from Afghanistan; thus, this could be used

as a unique marker for small-seed type melon from EastAsia.

Although the ability of SSR to discriminate was higherthan that of RAPD, RAPD markers A41-1353, A41-1020, B68-1078, and B71-1220 were also polymorphicin all nine populations. RAPD markers B84-700 andB84-600 were polymorphic in eight populations, exceptChinese Conomon, and in B86-1500, except USCantalupensis. For these seven RAPD markers, the PICwas from 0.444 to 0.499 (Table 2). The 18 RAPD primersused in this study were strictly selected from 176 primers,based on their ability to detect polymorphism and thestability of PCR amplification by Tanaka et al. (2007).The strictness of their selection can be seen in this studyfrom the small number of marker bands of each primer(1.44), similar to 1.69 of Mliki et al. (2001) and muchsmaller than 11.22 of Dhillon et al. (2007).

Among the melon accessions studied, ChineseConomon should be classified as C. melo ssp. agrestis

and the others as C. melo ssp. melo according to theclassification of Pitrat et al. (2000) and Jeffrey (2001).Figures 4 and 5 clearly show the genetic differentiationbetween these subspecies, as already reported byStepansky et al. (1999) and Monforte et al. (2003). Thesetwo figures also show a close relationship betweenSpanish accessions and US Inodorus, which was ratherdistantly related to US Cantalupensis. AlthoughCantalupensis (Galia type) and Inodorus (Cassaba, Pielde Sapo, and Tendral types) are grown in Spain (López-Sesé et al., 2002), Galia type was not included in theSpanish accessions examined here and thus should beclassified as group Inodorus. Therefore, Figures 4 and 5also successfully show a close relationship within groupInodorus of members of different geographical origin.These results show the usefulness of the RAPD and SSRmarker sets used in this study for diversity andphylogenetic analyses of melon, as well as for cultivaridentification.

We believe this is the first study of the genetic diversity

Table 7. Number of melon accessions classified into 10 clusters, in each melon population.

z Reference accessions (RA).

Area/GroupNo. of

accessions

Cluster

Ia Ib Ic IIa IIb IIc III IV V VI VIIa VIIb VIII IX Xa Xb Xc Xd

Iran 10 1 – – – – – 2 3 – – – – – 3 1 – – –

Afghanistan 10 – – – 3 – – 3 – – – 2 – – 1 1 – – –

Pakistan 11 – – – 1 – – 2 – 5 1 – – 2 – – – – –

Central Asia 17 – 2 – – 4 3 2 1 – – 4 – – – 1 – – –

Hami melon 24 – – – 1 19 – 1 1 – 2 – – – – – – – –

Spain 9 – 1 8 – – – – – – – – – – – – – – –

US Inodorus 8 6 2 – – – – – – – – – – – – – – – –

US Cantalupensis 8 – – – – – – – – – 1 7 – – – – – –

Chinese Conomon 19 – – – – – – – – – – – – – – – 11 3 5

RAz 4 – – – 1 – – – – – – – 1 – – – – 2 –

Total 120 7 5 8 6 23 3 10 5 5 3 7 8 2 4 3 11 5 5


of Hami melon analyzed using RAPD and SSR markers.The gene diversity of Hami melon (D = 0.243) wassimilar to the D of Spanish accessions, US Cantalupensis,and Chinese Conomon, and was smaller than the D oflandraces from Iran to Central Asia (Table 5). The genediversity in each variety of Hami melon was even smallerat 0.186 in var. ameri, 0.160 in var. zard, and 0.201 invar. chandalak (data not shown). Hami melon landracesof these varieties shared an identical chloroplast genometype (T/338 bp), while an improved cultivar of Hamimelon (X011, Fig. 1), bred by crossing with foreigngermplasm, had a different chloroplast genome type(T/333 bp) (Table 4). Therefore, the genetic variation ofHami melon was rather small, even though Hami melonconsisted of different varieties (chandalak, ameri, zard,cassaba) with diverse morphological traits, earliness,and fruit shelf life (Lin, 1991).

Figure 6 clearly shows the genetic relationship anddifferentiation between Hami melon varieties. Hamimelon accessions of vars. ameri and zard were mostlygrouped as subcluster IIb, but those of var. chandalak

as clusters IV and VI. The GD calculated between eachpair of ameri accessions and zard accessions averaged0.210, but it was 0.409 between three accessions of var.chandalak and 20 accessions of vars. ameri and zard

(data not shown). These results clearly indicated thatvar. chandalak was rather distantly related to vars. ameri

and zard, which were closely related to each other. Var.chandalak is early maturing and is grown mainly forlocal consumption because of its poor shelf life (oneweek). Var. ameri consists of early and mid-seasonmaturing groups (Lin, 1991). The fruit shelf life alsodiffers, at approximately two weeks and one month,respectively. The mid-season maturing group of var.ameri, as well as of var. zard, whose shelf life is overone month, are commonly grown for export to easternChina and other countries as Hami melon. Theseaccessions with a long shelf life were grouped intosubcluster IIb (Table 1, Fig. 6), indicating their closerelationship. Therefore, we suggest that not only var.zard but also the mid-season maturing group of var.ameri should be classified as Inodorus, even thoughPitrat et al. (2000) did not include var. ameri as Inodorus.The fruit shelf life of the early maturing group of var.ameri is 1–2 weeks, and two of six accessions weregrouped into clusters IIa and III, not into IIb (Table 1,Fig. 6). Similar variation of shelf life, from two weeksto three months, is also known in var. ameri of Uzbekmelon (Mavlyanova et al., 2005). Therefore, theclassification of var. ameri should be reconsidered.

The genetic diversity between Chinese melonlandraces has been studied in the last decade. Based onisozyme variation, Akashi et al. (2002) indicated distinctgenetic differentiation between the large-seed type,including Hami melon, and the small-seed typecultivated in eastern China. RAPD analysis has indicatedgenetic differentiation between thick-skinned and thin-

skinned Chinese melon (Luan et al., 2008). Thick-skinned melon accessions correspond to either Inodorusor Cantalupensis, and thin-skinned accessions corre-spond to Conomon. However, only one accession ofthick-skinned melon from Xinjiang was studied by Luanet al. (2008), and thus the genetic relationship betweenHami melon and melon accessions from other areasremains unknown. In this study, genetic differentiationbetween Hami melon and Conomon of eastern Chinawas clearly indicated (Table 6, Fig. 4). Analysis of thechloroplast genome type showed that Hami melon(T/338 bp type) and Conomon (A/338 bp type) belongto different maternal lineages (Table 4).

In the case of melon accessions of Central Asia,varieties chandalak, ameri, zard, cassaba, bucharica,and gurvak are grown in Uzbekistan (Mavlyanova et al.,2005). Figures 4 and 5, based on the genetic distanceshown in Table 4, show that Hami melon is relatedclosely to melon landraces of areas from Iran to CentralAsia. Large-seed type melon accessions sharing the samematernal lineages as Hami melon (T/338 bp type) werein Central Asia (Table 4), even though Tanaka et al.(2006) reported only the T/333 bp type. The T/338 bptype was distributed frequently from Iran to Central Asia(Table 4), and several accessions, such as US 119(Turkmenistan, subcluster IIb), US 19, US 390 (bothAfghanistan, subcluster IIa), were closely related toHami melon (vars. ameri and zard) by analysis of thenuclear genome (Table 1, Fig. 6). We therefore concludedthat Hami melon may have been transmitted from thewest along the so-called “Silk Road.”

Another interesting conclusion was deduced about theorigin of ‘Earl’s Favourite’, which is the leading breedof Japanese Cantalupensis melon. This cultivar wasimported from the UK in 1925. Among Cantalupensisaccessions imported from France and the UK, only‘Earl’s Favourite’ adapted to the Japanese growingconditions and it was selected as one of the sweetest.Although the mature fruit is less fragrant, this cultivarhas a good shelf life and has been intensively used as across parent for melon breeding. In this study, ‘Earl’sFavourite’ was used as a reference accession of groupCantalupensis, and was expected to show a closerelationship to accessions of US Cantalupensis. ‘Earl’sFavourite’ had an identical chloroplast genome type toUS Cantalupensis (T/338 bp type) (Tables 1 and 4).However, they were classified as distant subclusters, IIa(‘Earl’s Favourite’) and VIIb (US Cantalupensis), byanalysis of nuclear genome markers (Table 1, Fig. 6).Cluster II consisted of vars. ameri and zard of Hamimelon and landraces from Iran to Central Asia. Wetherefore suggest that ‘Earl’s Favourite’ might beselected from the hybrid of Inodorus melon from Iranto Xinjiang and European Cantalupensis melon, eventhough no information is available about the origin ofthis cultivar.


Acknowledgments

We would like to thank Dr. Y. Sakata, NationalInstitute of Vegetable and Tea Science (NIVTS), Japan,Dr. K. R. Reitsma, Iowa State University, USA, andDr. A. Graner, Leibniz Institute of Plant Genetics andCrop Plant Research (IPK), Germany, for kindlysupplying the seeds, and Dr. N. Fujishita for his kindencouragement throughout this study.

Literature Cited

Akashi, Y., N. Fukuda, T. Wako, M. Masuda and K. Kato. 2002.

Genetic variation and phylogenetic relationships in East and

South Asian melon, Cucumis melo L., based on the analysis

of five isozymes. Euphytica 125: 385–396.

Akashi, Y., S. Shiomi, Y. Kubo, M. Masuda and K. Kato. 2001.

Microsatellite and CAPS markers for ethylene-related genes,

1-Aminocyclopropane-1-Carboxylic Acid (ACC) synthase

and ACC Oxidase genes, and their variation in melon

(Cucumis melo L.). Breed. Sci. 51: 107–112.

Anderson, J. A., G. A. Churchill, J. E. Autrique, S. D. Tanksley

and M. E. Sorrells. 1993. Optimizing parental selection for

genetic-linkage maps. Genome 36: 181–186.

Apostol, B. L., W. C. I. V. Black, B. R. Miller, P. Reiter and B. J.

Beaty. 1993. Estimation of the number of full sibling families

at an oviposition site using RAPD-PCR markers: applications

to the mosquito Aedes aegypti. Theor. Appl. Genet. 86: 991–

1000.

Chiba, N., K. Suwabe, T. Nunome and M. Hirai. 2003.

Development of microsatellite markers in melon (Cucumis

melo L.) and their application to major Cucurbit crops. Breed.

Sci. 53: 21–27.

Chung, S. M. and J. E. Staub. 2003. The development and

evaluation of consensus chloroplast primer pairs that possess

highly variable sequence regions in a diverse array of plant

taxa. Theor. Appl. Genet. 107: 757–767.

Danin-Poleg, Y., N. Reis, S. Baudracco-Arnas, M. Pitrat, J. E.

Staub, M. Oliver, P. Arus, C. M. deVicente and N. Katzir.

2000. Simple sequence repeats in Cucumis mapping and map

merging. Genome 43: 963–974.

Dhillon, N. P. S., R. Ranjana, K. Singh, I. Eduardo, A. J. Monforte,

M. Pitrat, N. K. Dhillon and P. P. Singh. 2007. Diversity

among landraces of Indian snapmelon (Cucumis melo var.

momordica). Genet. Resour. Crop. Evol. 54: 1267–1283.

Fukino, N., Y. Sakata, M. Kunihisa and S. Matsumoto. 2007.

Characterisation of novel simple sequence repeat (SSR)

markers for melon (Cucumis melo L.) and their use for

genotype identification. J. Hort. Sci. Biotech. 82: 330–334.

Garcia-Mas, J., A. J. Monforte and P. Arus. 2004. Phylogenetic

relationships among Cucumis species based on ribosomal

internal transcribed spacer region. Plant Sys. Evol. 248: 191–

203.

Garcia-Mas, J., M. Oliver, H. Gomez-Paniagua and M. C. de

Vicente. 2000. Comparing AFLP, RAPD and RFLP markers

for measuring genetic diversity in melon. Theor. Appl. Genet.

101: 860–864.

Gower, J. C. 1966. Some distance properties of latent root and

vector methods used in multivariate analysis. Biometrika 53:

325–338.

Jeffrey, C. 2001. Cucurbitaceae. p. 1510–1557. In: P. Hanelt and

Institute of Plant Genetics and Crop Plant Research (eds.).

Mansfeld’s Encyclopedia on Agricultural and Horticultural

Crops. Springer, Berlin-Heidelberg-New York.

Lin, D. P. 1991. Classification of melon, p. 9–21. In: S. K. Qi,

D. K. Wu and D. P. Lin (eds.). Chinese Melon. Popular

Science Press, Beijing (In Chinese).

Liu, L., F. Kakihara and M. Kato. 2004. Characterization of six

varieties of Cucumis melo L. based on morphological and

physiological characters, including shelf-life of fruit.

Euphytica 135: 305–313.

López-Sesé, A. I., J. E. Staub and M. L. Gómez-Guillamón. 2003.

Genetic analysis of Spanish melon (Cucumis melo L.)

germplasm using a standardized molecular-marker array and

geographically diverse reference accessions. Theor. Appl.

Genet. 108: 41–52.

López-Sesé, A. I., J. E. Staub, N. Katzir and M. L. Gómez-

Guillamón. 2002. Estimation of between and within accession

variation in selected Spanish melon germplasm using RAPD

and SSR markers to assess strategies for large collection

evaluation. Euphytica 127: 41–51.

Luan, F., I. Delannay and J. E. Staub. 2008. Chinese melon

(Cucumis melo L.) diversity analyses provide strategies for

germplasm curation, genetic improvement, and evidentiary

support of domestication patterns. Euphytica 164: 445–461.

Mavlyanova, R., A. Rustamov, R. Khakimov, A. Khakimov, M.

Turdieva and S. Padulosi. 2005. Melons of Uzbekistan.

IPGRI’s Sub-regional office for Central Asia, Tashkent.

Mliki, A., J. E. Staub, Z. Y. Sun and A. Ghorbel. 2001. Genetic

diversity in melon (Cucumis melo L.): An evaluation of

African germplasm. Genet. Resour. Crop Evol. 48: 587–597.

Monforte, A. J., J. Garcia-Mas and P. Arús. 2003. Genetic

variability in melon based on microsatellite variation. Plant

Breed. 122: 153–157.

Munger, H. M. and R. W. Robinson. 1991. Nomenclature of

Cucumis melo L. Cucurbit Genet. Coop. Rep. 14: 43–44.

Murray, G. C. and W. F. Thompson. 1980. Rapid isolation of high

molecular weight DNA. Nucl. Acids Res. 8: 4321–4325.

Nakamura, I., N. Kameya, Y. Kato, S. I. Yamanaka, H. Jomori

and Y. Sato. 1997. A proposal for identifying the short ID

sequence which addresses the plastid subtype of higher plants.

Breed. Sci. 47: 385–388.

Nakata, E., J. E. Staub, A. I. López-Sesé and N. Katzir. 2005.

Genetic diversity of Japanese melon cultivars as assessed by

random amplified polymorphic DNA and simple sequence

repeat markers. Genet. Resour. Crop Evol. 52: 405–419.

Nei, M. 1972. Genetic distance between populations. Am. Nat.

106: 283–292.

Pitrat, M., P. Hanelt and K. Hammer. 2000. Some comments on

intraspecific classification of cultivars of melon. Proc.

Cucurbitaceae: 29–36.

Ritschel, P. S., T. C. L. Lins, R. L. Tristan, G. S. C. Buso, J. A.

Buso and M. E. Ferreira. 2004. Development of microsatellite

markers from an enriched genomic library for genetic analysis

of melon (Cucumis melo L.). BMC Plant Biol. 4: 9.

Soltani, F., Y. Akashi, A. Kashi, Z. Zamani, Y. Mostofi and K.

Kato. 2010. Characterization of Iranian melon landraces of

Cucumis melo L. Groups Flexuosus and Dudaim by analysis

of morphological characters and random amplified polymor-

phic DNA. Breed. Sci. 60: 34–45.

Staub, J. E., Y. Danin-Poleg, G. Fazio, T. Horejsi, N. Reis and N.

Katzir. 2000. Comparative analysis of cultivated melon

groups (Cucumis melo L.) using random amplified

polymorphic DNA and simple sequence repeat markers.

Euphytica 115: 225–241.

Staub, J. E., A. I. López-Sesé and N. Fanourakis. 2004. Diversity

among melon landraces (Cucumis melo L.) from Greece and

their genetic relationship with other melon germplasm of

diverse origins. Euphytica 136: 151–166.

Stepansky, A., I. Kovalski and R. Perl-Treves. 1999. Intraspecific


classification of melons (Cucumis melo L.) in view of their

phenotypic and molecular variation. Plant Sys. Evol. 217:

313–332.

Tanaka, K., K. Fukunaga, Y. Akashi, H. Nishida, K. Kato and

M. T. Khaing. 2006. Polyphyletic origin of cultivated

melon inferred from analysis of its chloroplast genome.

Cucurbitaceae 2006: 372–379.

Tanaka, K., A. Nishitani, Y. Akashi, H. Nishida, H. Yoshino and

K. Kato. 2007. Molecular characterization of South and East

Asian melon, Cucumis melo L., and the origin of group

Conomon var. makuwa and var. conomon revealed by RAPD

analysis. Euphytica 153: 233–247.

Weir, B. S. 1996. Genetic data analysis II. Sinauer Associate Inc.

Publisher, Massachusetts.

Wu, M. Z. 1982. Genetic resources of thick-skinned melon

collected from Turpan Depression and their utilization. China

Fruits 1982: 28–33 (In Chinese).

Yashiro, K., H. Iwata, Y. Akashi, K. Tomita, M. Kuzuya, Y.

Tsumura and K. Kato. 2005. Genetic relationship among East

and South Asian melon (Cucumis melo L.) revealed by AFLP

analysis. Breed. Sci. 55: 197–206.

Yi, S. S., Y. Akashi, K. Tanaka, T. T. Cho, M. T. Khaing, H.

Yoshino, H. Nishida, T. Yamamoto, K. Win and K. Kato.

2009. Molecular analysis of genetic diversity in melon

landraces (Cucumis melo L.) from Myanmar and their

relationship with melon germplasm from East and South Asia.

Genet. Resour. Crop Evol. 56: 1149–1161.

Documents

Molecular Analysis of the Genetic Diversity of Chinese