Molecular Breeding

, Volume 24, Issue 1, pp 77–91

Elucidation of origin of the present day hybrid banana cultivars using the 5′ETS rDNA sequence information

Authors

    • Biogenetics/Natural resourcesAustrian Research Centers GmbH—ARC
  • Silvia Fluch
    • Biogenetics/Natural resourcesAustrian Research Centers GmbH—ARC
  • Kornel Burg
    • Biogenetics/Natural resourcesAustrian Research Centers GmbH—ARC
Article

DOI: 10.1007/s11032-009-9273-z

Cite this article as:
Boonruangrod, R., Fluch, S. & Burg, K. Mol Breeding (2009) 24: 77. doi:10.1007/s11032-009-9273-z

Abstract

In the present study, our intention was to elucidate the genetic relation of M. acuminata subspecies and analyse the diversity of the M. balbisiana gene-pool using nuclear ribosomal gene loci based marker system. Additionally the obtained information allowed elucidating the structure and ancestry of the nuclear genomes of diploid and triploid cultivars. By establishing the nucleotide sequence of the rDNA locus for M. acuminata and partially for M. balbisiana and their comparative analysis revealed that the 5′ETS region was the most divergent between acuminata and balbisiana genomes. Based on the SNP sites identified in this region a PCR based system was built, which revealed four gene-pools among M. acuminata wild types, while M. balbisiana showed no sequence divergence. The developed markers proved to be a powerful tool in the identification of the acuminata component of diploid and triploid hybrid cultivars and discovery of unexpected genotypes.

Keywords

Musa acuminataM. balbisianaRibosomal DNAGene-pool specific markers

Introduction

Bananas (Musa ssp.) belong to the family Musaceae. They are native to Southeast Asia and the western Pacific, widely spread out in the subtropics where they have been propagated vegetatively. They constitute a major staple food for millions of people in these countries, it also ranks fourth among the most important global food crops where up to 90% of the production is for local consumption. The genus Musa represents four sections Eumusa, Rhodochlamys, Callimusa and Australimusa. The wild progenitors of the domesticated banana are Musa acuminata (A genome), Musa balbisiana (B genome), and to a much lesser extent, Musa schizocarpa (S genome) and Musa textilis/Musa maclayi (T genome) (Carreel et al. 2002; Daniells et al. 2001). Cultivated bananas are mostly seedless diploid, triploid or tetraploid plants resulting from combination of two diploid species, M. acuminata and M. balbisiana as it was classified by Simmonds and Shepherd (1955), who studied the taxonomy of triploid cultivars using 15 morphological characters for the first time. Based on phenotypic traits, currently the following subspecies are recognised in M. acuminata wild types: banksii, burmannica, burmannicoides,errans (Allen), malaccensis, microcarpa, siamea (Simmonds), truncata and zebrina. Recently by using DNA based markers like microsatellite and RFLP, M. acuminata genotypes were classified into four groups i.e. banksii, zebrina, malaccensis and burmannica/burmannicoides by Carreel et al. (1994) and Grapin et al. (1998). Latter, Ude et al. (2002) suggested three genetic subspecies in M. acuminata based on AFLP genotyping and on the predominance of microcarpa, malaccensis and burmannica genotypes in three identified groups. However, despite the existence of genetic variation, there is no record of subspecies classification in M. balbisiana.

Nuclear ribosomal RNA genes (rDNA) of plants are arranged as multiple copies of tandem repeats of the transcribed 18S–5.8S–26S rRNA genes. Each of the transcribed units consists of conserved coding regions separated by less conserved internal transcribed spacer (ITSs) and flanked by non-transcribed inter gene spacer (IGS). Due to the concerted evolution, which makes all repeated units expected to be similar within organism, but differ in closely related species, rDNA could be analysed as a single copy gene (Ganley and Kobayashi 2007). Ribosomal DNA regions have frequently been used for studying both intra- and inter-specific relations and evolutionary history. For example the 18S locus has been found useful for elucidating the ancient evolution of angiosperm families (Soltis et al. 1997). Additionally the sequence differences of the IGS (Cordesse et al. 1993; Polanco and Vega 1994; Da Rocha and Bertrand 1995) and the two ITSs along with the external transcribed spacer (ETS) regions (Baldwin and Markos 1998; Linder et al. 2000) were also frequently used for species determination, establishing evolutionary relations and population studies. Several publications reported that the sequences of 5′ETS region of the transcribed region had a great potential for investigating genetic diversity, since this region was shown to evolve fast and proved to be more informative compared to the ITS region. Therefore this region became a tool for studying phylogeny among closely related species (Volkov et al. 1996; Baldwin and Markos 1998; Linder et al. 2000; Sallares and Brown 2004).

Concerning Musa, relatively few data are available on the rDNA loci of M. acuminata and M. balbisiana. Valárik et al. (2002) isolated repetitive DNA sequences including sequences of 26S rDNA of M. acuminata. They observed that 26S rDNA were more abundant in M. acuminata whose genome is about 12% larger than M. balbisiana. The number of the 26S rDNA locus per haploid genome in M. acuminata was estimated as 1.6 × 104 copies, while in M. balbisiana it was 1.3 × 104 copies. The chromosomal localisation of rDNA revealed that in all accessions of Eumusa and Australimusa, the 45S (18S–5.8S–26S) loci are localised on one pair of nucleolar organising chromosomes, while the number of 5S rDNA loci in Musa vary from 4 to 8 per diploid cell and vary among species and subspecies. On the contrary, variable number of 45S rDNA sites was observed in the section Rhodochlamys (2–4 sites on one and two chromosome pair(s), respectively) and 6 sites (three chromosome pairs) were found in M. beccarii. (Bartoš et al. 2005; Dolezelova et al. 1998; Osuji et al. 1998).

As far as genetic markers based on the ribosomal loci are concerned, recently Nwakanma et al. (2003) developed a PCR-RFLP marker in the ITS region of banana, which could differentiate the acuminata and balbisiana type rDNA genes.

In the present study, our intention was to elucidate the genetic relation of M. acuminata subspecies and analyse the diversity of the M. balbisiana gene-pool using nuclear ribosomal gene loci based marker system and using the obtained information on elucidating the structure and ancestry of the nuclear genomes of diploid and triploid cultivars. Identification of gene-pools participating in the formation of the genomes of cultivars could greatly enhance our understanding on the genetic structure of cultivated diploid or polyploid accessions, thus providing valuable ammunition for present day banana breeding.

Materials and methods

Plant material

Purified DNA of 48 members of a ‘mini core’ collection of 51 accessions (http://www.musagenomics.org/index.php?id=137) were kindly provided by the Global Musa Genomics Consortium and by Isabelle Hippolyte (CIRAD), respectively Table 1). Additionally, purified DNA samples of three M. balbisiana (50, 51 and 52) and two M. acuminata (9 and 10) wild types kindly provided by J. Doležel (Laboratory of Molecular Cytogenetics and Cytometry. Institute of Experimental Botany, Olomouc, Czech Republic) were also included in the present study. Two duplicate DNA samples—Lal Velchi (BBwt; 48) and Kluai Tiparot (ABB; 36)—were additionally obtained from J. Doležel and from the collection of IAEA Laboratories Seibersdorf, respectively.
Table 1

List of analysed accessions

Number

Nuclear genotype

Species

Sub-species

Common name

Country of origin

INIBAP transit centre code

1

AA

acuminata

microcarpa

Borneo

Malaysia, S/E Borneo

ITC0253

2

AA

acuminata

burmannicoides

Calcutta 4

India, Calcutta

ITC0249

3

AA

acuminata

errans

Agutay

Philippines

ITC1028

4

AA

acuminata

siamea

Khae (Phrae)

Thailand

ITC0660

5

AA

acuminata

burmannica

Long Tavoy

 

ITC0283

6

AA

acuminata

banksii

Paliama

Papua New Guinea (PNG067)

ITC0766

7

AA

acuminata

banksii

Banksii

Papua New Guinea

ITC0623

8

AA

acuminata

zebrina

Zebrina

Indonesia

ITC1177

9

AA

acuminata

zebrina

Maia Oa

Hawaii

ITC0728

10

AA

acuminata

malaccensis

Malaccensis

Peninsular Malaysia

ITC0250

11

AA

AAcv (18)

Pisang jari buaya

Pisang Jari Buaya

Malaysia, kelatan, Thai border

ITC0312

12

AA

AAcv (2)

Sucrier

Pisang mas

Malaysia

ITC0653

13

AA

AAcv

Cooking AA

Tomolo

Papua New Guinea (PNG023)

ITC1187

14

AAA

AAA

Cavendish

Grande Naine

Guadeloupe

ITC0180

15

AAA

AAA

Cavendish

Petite Naine

 

ITC0654

16

AAA

AAA

Cavendish

Poyo

Nigeria

ITC0345

17

AAA

AAA

Orotava

Pisang Kayu

Indonesia (IDN098)

ITC0420

18

AAA

AAA

Ambon

Pisang bakar

Indonesia (IDN106)

ITC1064

19

AAA

AAA

Gros Michel

Gros Michel

Guadeloupe

ITC0484

20

AAA

AAA

Rio

Leite

 

ITC0277

21

AAA

AAA

Lujugira/Mutika

Mbwazirume

Burundi

ITC0084

22

AAA

AAA

Lujugira/Mutika

Intokatoke

Burundi

ITC0082

23

AAA

AAA

Ibota

Yangambi km5

DR Congo

ITC1123

24

AAB

AAB

Nadan

Lady Finger

India

ITC0582

25

AAB

AAB

Pome/Prata

Foconah

DR Congo

ITC0649

26

AAB

AAB

Pome/Prata

Prata Ana

Brazil

ITC0962

27

AAB

AAB

Plantain

Orishele

Nigeria

ITC1325

28

AAB

AAB

Plantain

Red Yade

Cameroon

ITC1140

29

AAB

AAB

Silk

Figue Pomme Géante

Guadeloupe

ITC0769

30

AAB

AAB

Popoulou/Maia Maoli

Popoulou

Cameroon

ITC0335

31

AAB

AAB

Pisang raja

Pisang Raja Bulu

Indonesia (IDN093)

ITC0843

32

AAB

AAB

Nendra padaththi

Pisang Rajah

Malaysia

ITC0243

33

AAB

AAB

Mysore

Pisang Ceylan

Malaysia

ITC1441

34

AB

ABcv

 

Safet Velchi

India

ITC0245

35

AB

ABcv

 

Kunnan

India, Kerala

ITC1034

36

ABB

ABB

Klue teparod

Kluai Tiparot

Thailand (THA020)

ITC0652

37

ABB

ABB

Pelipita

Pelipita

Philippines

ITC0472

38

ABB

ABB

Bluggoe

Dole

 

ITC0767

39

ABB

ABB

Saba

Saba

Philippines

ITC1138

40

ABB

ABB

Monthan

Monthan

India

ITC0046

41

ABB

ABB

Peyan

Simili Radjah

From India through DR Congo

ITC0123

42

ABB

ABB

Ney mannan

Ice Cream

 

ITC0020

43

ABB

ABB

Pisang Awak

Namwa Khom

Thailand (THA011)

ITC0659

44

ABBT

ABBT

ABBT Ssp/sgr 501

Yawa 2

Papua New Guinea (PNG072)

ITC1238

45

BB

balbisiana

Type 4

Pisang Klutuk Wulung

Indonesia (IDN056)

ITC1063

46

BB

balbisiana

Type 4

Pisang Batu

Indonesia (IDN080)

ITC1156

47

BB

balbisiana

Type 1

Honduras

Honduras (seeds)

ITC0247

48

BB

balbisiana

Type 3

Lal Velchi

India

NEU0051

49

BB

balbisiana

 

Tani

 

ITC1120

50

BB

balbisiana

 

Type Cameroon

Sri Lanka

ITC0246

51

BB

balbisiana

 

Singapurii

 

ITC0248

52

BB

balbisiana

 

Butuhan

Philippines

ITC0564

53

AS

AS

 

Wompa

Papua New Guinea (PNG063)

ITC1152

DNA sequencing

All sequencing reactions were made by the use of BigDye® Terminator v3.1 Ready Reaction Cycle Sequencing Kit (Applied Biosciences Cat. No: 4337456). The sequence reads were obtained on ABI Prism 3100 Genetic Analyzer (Applied Biosystems) using 50 cm capillary array.

Small fragment DNA libraries

Small fragment libraries were constructed using DNA from Calcutta4 (M. acuminata), Tani (M. balbisiana) and Yangambi KM5 (AAA) after HpaII/MspI digestion of the genomic DNAs. 192 clones from each library (Calcutta4, Tani and Yangambi KM5) were sequenced. Annotation of the sequences by using the NCBI sequence database revealed that about 23% of the clones were found to be representing ribosomal RNA gene locus. Sequence data of 19 clones have been deposited in NCBI sequence database under the accession numbers: EU418628–EU418634 (Calcutta4), EU433921–EU433927 (YangambiKM5) and EU433928–EU433932 (Tani).

Sequence comparisons and phylogenetic analysis of M. acuminata subspecies

For elucidating sequence differences between acuminata and balbisiana rDNA loci the consensus sequence of each species was obtained from two repeated sequencing analysis of the clones submitted to the NCBI database.

For the comparative analysis of the 5′ETS region of the rDNA loci of acuminata subspecies, it was amplified with the primer combination ETS(+280)/18S(+801) (Table 2). Four to six repeats of direct sequencing reactions were performed using the amplification primers. Nucleotide sequence alignment was made by CLUSTAL W version 2 (http://www.ebi.ac.uk/Tools/clustalw2/index.html) software and the alignments were subsequently improved manually in GeneDoc version 2.6.001 (Nicholas and Nicholas 1997) environment.
Table 2

List of PCR primers used for detection of gene-pool specific SNPs and for amplification of the analysed 5′ETS region

Primer

SNP site (bp)

Primer sequence

Tan (°C)a

Forward primers

ETS(+280)b

5′-GACTGTGTCGGCATGTGAG-3′

55

SNP310B

310

5′-CATGTGAGTGGTGATGCATC-3′

62

SNP352A1

352

5′-CTGCTATTGTGCAGTTGATCAT-3′

63

SNP420A2

420

5′-GTGTGTTGCCTGTACAATATAT-3′

63

SNP469A3

469

5′-CCTTGCTTATCCACTTGTCT-3′

63

SNP527A4a

527

5′-CTTGTGTGTCCTCGAACCATT-3′

63

SNP615A4b

615

5′-TAAGAATGTACCATGGGTGTTC- 3′

63

SNP669B

669

5′-CGCCCTTCGACGGATCACC-3′

62

Reverse primer

18S(+801)

5′-CATGCATGGCTTAATCTTTGA-3′

aAnnealing temperature applied in the PCR reactions for the SNP forward primer/18S(+801) reverse primer combination

bForward primer used for the amplification of the analysed 5′ETS region in combination with 18S(+801) reverse primer

The amplified fragment of M. acuminata ssp. malaccensis was cloned by TOPO TA Cloning kit (Invitrogen Cat. No: K4575-40) using manufacturer’s protocol. Eight clones were picked and eventually sequenced twice using M13 sequencing primers.

Genetic relationships of the M. acuminata subspecies were analysed by principal component analysis (PCA) based on correlation using the PAST software package (Hammer et al. 2001) using genotypes based on the SNP differences listed in Table 5.

SCAR markers

Seven primer pairs using one common reverse primer were planned for detecting the SNP sites characterising the four acuminata rDNA gene-pools and the presence of the balbisiana rDNA locus (Table 2). The primers were planned using the Amplification Refractory Mutation system (ARMS) (Newton et al. 1989), which is a useful technique for increasing efficacy of specificity of the PCR primer for the specific SNP allele. The technique requires an allele/SNP specific terminal 3′-nucleotide of the PCR primer with an additional mismatch near to the 3′-end. In this study we introduced this additional mismatch at the fourth position upstream to the 3′-terminal nucleotide of PCR primers. This way primers were designed based either on M. acuminata gene-pool specific SNPs of the 5′ETS region (SNP352A1, SNP420A2, SNP469A3, SNP527A4a and SNP615A4b) or SNPs detecting the presence of the balbisiana rDNA locus (SNP310B and SNP669B).

PCRs were carried out in a total volume of 25 μl consisting of 5–10 ng of template DNA, 160 nM of each forward and reverse primers, 250 μM each of dATP, dCTP, dGTP and dTTP, 0.625 Unit KlearTaq polymerase (KBiosciences) and 1.2× PCR buffer. The PCR program applied was: 95°C for 15 min, followed by 28 cycles of 95°C for 30 s, Tan for 30 s (annealing temperature for each primer listed in Table 2), 72°C for 30 s, and completing the reaction by 72°C for 5 min. The fragments were separated by agarose gel (1%) electrophoresis in 1× TBE buffer. The gels were stained with ethidium-bromide solution and the PCR fragments were visualised under UV light.

Results

Sequence of the Musa ribosomal locus

Based on the sequence information of 19 clones originating from our small fragment libraries representing either the ‘A’ (M. acuminata ssp. burmannicoides; Calcutta4 and triploid cultivar Ibota; Yangambi KM5) or the ‘B’ (M. balbisiana Tani) genomes supplemented by already published (AF069226, AF399947.1 for M. acuminata and AF331972.1 for AAB cultivar) sequences, a contig spanning 6,460 bp region of the Musa rDNA locus could be assembled (Fig. 1). This contig provides nearly full sequence information for the acuminata rDNA locus (6,264 bp, −27 to +6,237), while partial sequence information for the balbisiana rDNA locus spanning up to the 3′ end of the ITS2 region could also be obtained (3,407 bp; −247 to +3,160). The sequence information of the acumunata contig revealed the presence of a full copy of 18S, 5.8S and 26S ribosomal genes of 1,745, 158 and 3,086 bp, respectively. The two internal transcribed spacers (ITS1 and 2) were 256 and 224 bp, respectively, while the 5′ETS spanned 768 bp. Based on conserved sequence motif in the RNA polymerase I transcription initiation site (TIS) of higher plants, the putative TIS could also be identified (TATAGCAGGGA), which proved to be highly similar to that in rice (TATAGTAGGGG) published by Fujisawa et al. (2006). This sequence motif was used for orientation/basepair-counting by assigning position +1 to its 5′ first base. Additionally the highly conserved sequence motif described by Bena et al. (1998) was also present in the Musa 5′ETS (CATGTGAGTGGTGATGGA; +291 to +308 bp). The transcribed region was preceded by one sub repeat of the inter gene spacer (IGS) region identified in a balbisiana clone (EU433928). This IGS sequence showed high homology to the clones submitted by Teo and Schwarzacher to NCBI sequence data bank (AM905874–AM905898), which form tandem repeat series spanning about 1.7 kb and also including several TIS sites and conserved regions.
https://static-content.springer.com/image/art%3A10.1007%2Fs11032-009-9273-z/MediaObjects/11032_2009_9273_Fig1_HTML.gif
Fig. 1

aM. acuminata (EU418628–EU418634) and triploid acuminata cultivar Yangambi KM5 (EU433921–EU433927) clones used for contig assembly. b Clones representing the M. balbisiana rDNA locus (EU433928–EU433932). c Clones available in sequence databases—Musa acuminata (AF069226 and AF399947.1), M. beccarii (AF434900) and M. paradisiaca (AF331972.1). d Frequency of sequence differences between M. acuminata versus M. balbisiana rDNA regions

Sequence divergence of M. acuminata and M. balbisiana rDNA locus

Comparative sequence analysis of the acuminata and balbisiana rDNA loci including the 5′ETS, 18S, ITS1 5.8S and ITS2 revealed several sequence differences, which could be categorised as SNPs, DNPs, TNPs (Single, Double or Triple Nucleotide Polymorphisms, respectively) and indels. The M. acuminata type rDNA sequence was based on sequence information obtained in Calcutta4 and Yangambi KM5 diploid wild type and triploid hybrid genomes, respectively, while diploid wild type accession Tani served as source for M. balbisiana genome. In the analysed region (about 3,100 bp) the frequency of sequence divergence was high in the non-coding transcribed regions (ETS and ITS), being the highest in the 5′ETS region (7.55%), while the 18S gene yielded the lowest level of divergence (0.23%) (Table 3; Fig. 1). Within this region we could identify 66 SNPs, 8 DNPs, 4 TNPs and 6 indels. Two out of the six indels were large gaps spanning 29 (position 39–68) and 20 (108–128) basepairs in the balbisiana 5′ETS region (Table 3). About 12 SNP sites out of the 64 resulted in restriction site mutations differentiating the acuminata and balbisiana genomes (Table 4).
Table 3

Type and frequency of sequence differences in the regions of M. acuminata and M. balbisiana rDNA loci based on sequence information of Calcutta4, Tani and Yangambi KM5 representing M. acuminata, M. balbisiana and Triploid acuminata, respectively

Type

Position

Length (bp)

SNP

DNP

TNP

InDel

Total (site)

Frequency (%)

IGS

(<)−247 to +1

(>)247

N/A

N/A

N/A

N/A

N/A

N/A

5′-ETS

+1 to +768

768

47

6

3

2a

58

7.55

18S

+769 to +2513

1,745

3

1

4

0.23

ITS1

+2514 to +2769

256

7

1

2

10

4.0

5.8S

+2770 to +2927

158

1

1

0.63

ITS2

+2928 to +3151

224

8

1

2

11

4.9

26S

+3152 to (>)+6237

(>)3,086

N/A

N/A

N/A

N/A

N/A

N/A

N/A not analysed

a30 and 21 bp deletion in Tani

Table 4

List of restriction enzyme sites differentiating M. acuminata and M. balbisiana rDNA loci

SNP

REa

M. acuminata

M. balbisiana

5′-ETS

310

Sau3AI

+

320

SacII

+

345

AluI

+

388

HpyCH4 N

+

561

HpaII

+

603

RsaI

+

630

CviAII

+

669

HaeIII

+

18S

978

HaeIII

+

1292

Sau3AI

+

ITS1

2704b

RsaI

+

ITS2

3025

Sau3AI

+

aAffected restriction enzyme site

bOriginally published by Nwakanma et al. (2003)

Comparative sequence analysis of M. acuminata subspecies

In our study the 5′ETS region proved to be the most unstable concerning the accumulation of mutations by comparing the two species M. acuminata and M. balbisiana. Since the sequence of the 5′ end of the 5′ETS region (<+280 bp) showed high similarity to that of IGS sub repeat, this region was neglected for PCR primer nesting. Consequently to analyse the phylogenetic relations of already established M. acuminata subspecies and discover genetic structure of M. balbisiana wild type gene-pool, the region spanning +280 to +801 bases was amplified. This fragment mostly contained sequences of the 5′ETS region along with 33 bp of the 18S rRNA gene. Direct sequencing of the amplified 522 bp long fragment of ten M. acuminata wild types revealed 29 SNP sites within the analysed subspecies (Table 5). On the contrary, the analysis of the same region in M. balbisiana did not yield any sequence polymorphisms in the eight analysed genotypes. Three SNP sites (376, 536 and 596) represented putative mutational hot spots since at these positions three different bases could be detected by including both species in the evaluation (Table 5). Direct sequencing of M. acuminata ssp. malaccensis revealed several SNP sites of approximately equal strength suggesting the presence of two alleles. Therefore, this amplificate was subsequently cloned and the presence of two alleles (10a and 10b) could be confirmed (Table 5).
Table 5

List of SNP sites identified by sequence comparison of the +280/+801 bp region of ten accessions of M. acuminata wild types and two additional sites differentiating M. acuminata and M. balbisiana wild type (310 and 669)

https://static-content.springer.com/image/art%3A10.1007%2Fs11032-009-9273-z/MediaObjects/11032_2009_9273_Tab5_HTML.gif

 

The SNP sites used for gene-pool discrimination are shown in bold italics

Principal Component Analysis (PCA) of the obtained sequence information revealed four gene-pools within the acuminata subspecies (Fig. 2). Gene-pool A1 was represented by ssp. zebrina (8 and 9), while one allele found in ssp. malaccensis (10a) identified gene-pool A2. The third gene-pool (A3) was formed by ssp. burmannicoides (2), siamea (4) and burmannica (5), while A4 consisted of ssp. microcarpa (1), banksii (6 and 7) and errans (3) along with the second allele identified in ssp. malaccensis (10b).
https://static-content.springer.com/image/art%3A10.1007%2Fs11032-009-9273-z/MediaObjects/11032_2009_9273_Fig2_HTML.gif
Fig. 2

Principal component analysis of ten Musa acuminata wild type accessions including eight subspecies based on the sequence differences of the +280/+801 bp amplified region of rDNA. The numbering of accessions refers to Table 1

Development of gene-pool specific SCAR marker system

Based on gene-pool specific SNP sites, Sequence Characterised Amplified Region (SCAR) marker system could be established by developing subspecies/gene-pool specific PCR system. These gene-pool representative SNP sites are shown in Table 5. This way position SNP352 was used for gene-pool A1, SNP420 for gene-pool A2 and SNP469 for gene-pool A3. Two SNPs were used for gene-pool A4. SNP527 represented a subgroup within A4 by identifying ssp. microcarpa, errans and banksii only (subgroup A4a), while SNP615 represented the whole gene-pool including the allele found in ssp. malaccensis, thus identifying a second subgroup (A4b) represented by the latter allele. For identifying the presence of the M. balbisiana genome, SNP310 and SNP669 were used. The SNP specific primers and the parameters of the obtained amplificates are listed in Table 2 and the patterns for wild types and some selected hybrid genotypes are shown in Fig. 3.
https://static-content.springer.com/image/art%3A10.1007%2Fs11032-009-9273-z/MediaObjects/11032_2009_9273_Fig3_HTML.gif
Fig. 3

Agarose gel electrophoresis of gene-pool specific PCRs on 1% agarose gels in 1× TBE buffer. aM. acuminata wild type subspecies representing the different gene-pools. b Genotype of selected cultivars including the AAA genotypes Gros Michel (19), Cavendish (16) and Rio (20). c Wild types with unexpected pattern: M. acuminata ssp. banksii (6) showing balbisiana, and M. balbisiana Lal Velchi (48) showing acuminata component

The application of the marker system on the available sample set revealed that the M. acuminata ssp. banksii (ITC0766; 6) apart of the acuminata genome unexpectedly contains also balbisiana rDNA component, while M. balbisiana (Neu0051) Lal Velchi (48) contains acuminata rDNA component beside the balbisiana genome (Fig. 3).

Contribution of the gene-pools to the genotypes of the cultivars

Since the balbisiana genome did not yield any diversity in the analysed 5′ETS of the ribosomal locus, it was possible to analyse the acuminata part of the hybrid accessions only. Since the balbisiana genome was always present in the acuminata/balbisiana hybrid accessions, we exclusively dealt with the acuminata genome part of the hybrids.

According to the number of acuminata gene-pools participating in the formation of the genome of the cultivars, three groups containing one, two or three different gene-pools could be identified. In the analysed cultivars, the presence of A1, A2 and A4 gene-pools could be observed, however, that of gene-pool A3 representing subspecies burmannica, burmannicoides and siamea could not be verified.

Single type of acuminata gene-pool was identified in 17 accessions (Table 6). The presence of A4 type acuminata genome could be detected in ten out of the seventeen cultivars namely in five ABB (37, 38, 39, 40 and 42), in four AAB (27, 28, 30 and 33) and in Rio (20), this latter is a triploid acuminata (AAA) genotype. In all these cases the A4 gene-pool was of A4a type represented by the subspecies banksii, errans and microcarpa. A2 type gene-pool could be identified in one ABB triploid hybrid (43) in two diploid AB hybrids (34 and 35) in one AA diploid cultivar (13) and in the tetraploid genotype (44). A single AA cultivar (11) and the AS (53) interspecies hybrid represented A1 gene-pool. Axiomatically all ABB genotypes should also belong to this group (accessions 36–43), however, 36 and 41 (Kluai Tiparot and Peyan) did not show the presence of any acuminata type ribosomal locus but showed only the presence of the balbisiana genome with the applied marker system (Table 6).
Table 6

Contribution of the different acuminata gene pools to the cultivar genomes

Number

Nuclear genotype

Sub-species

Common name

INIBAP Transit Centre code

acuminata

B (balbisiana)

Identified gene-pool

A1

A2

A3

A4a

A4b

36

ABB

Klue teparod

Kluai Tiparot

ITC0652

0

0

0

0

0

1

-/B

41

ABB

Peyan

Simili Radjah

ITC0123

0

0

0

0

0

1

-/B

20

AAA

Rio

Leite

ITC0277

0

0

0

1

1

0

A4a

27

AAB

Plantain

Orishele

ITC1325

0

0

0

1

1

1

A4a/B

28

AAB

Plantain

Red Yade

ITC1140

0

0

0

1

1

1

A4a/B

30

AAB

Popoulou/Maia Maoli

Popoulou

ITC0335

0

0

0

1

1

1

A4a/B

33

AAB

Mysore

Pisang Ceylan

ITC1441

0

0

0

1

1

1

A4a/B

37

ABB

Pelipita

Pelipita

ITC0472

0

0

0

1

1

1

A4a/B

38

ABB

Bluggoe

Dole

ITC0767

0

0

0

1

1

1

A4a/B

39

ABB

Saba

Saba

ITC1138

0

0

0

1

1

1

A4a/B

40

ABB

Monthan

Monthan

ITC0046

0

0

0

1

1

1

A4a/B

42

ABB

Ney mannan

Ice Cream

ITC0020

0

0

0

1

1

1

A4a/B

13

AA

Cooking AA

Tomolo

ITC1187

0

1

0

0

0

0

A2

34

AB

 

Safet Velchi

ITC0245

0

1

0

0

0

1

A2/B

35

AB

 

Kunnan

ITC1034

0

1

0

0

0

1

A2/B

43

ABB

Pisang Awak

Namwa Khom

ITC0659

0

1

0

0

0

1

A2/B

44

ABBT

ABBT Ssp/sgr 501

Yawa 2

ITC1238

0

1

0

0

0

1

A2/B

11

AA

Pisang jari buaya

Pisang Jari Buaya

ITC0312

1

0

0

0

0

0

A1

53

AS

 

Wompa

ITC1152

1

0

0

0

0

0

A1

29

AAB

Silk

Figue Pomme Géante

ITC0769

0

1

0

1

1

1

A2/A4a/B

12

AA

Sucrier

Pisang mas

ITC0653

1

0

0

1

1

0

A1/A4a

17

AAA

Orotava

Pisang Kayu,

ITC0420

1

0

0

1

1

0

A1/A4a

18

AAA

Ambon

Pisang bakar,

ITC1064

1

0

0

1

1

0

A1/A4a

21

AAA

Lujugira/Mutika

Mbwazirume

ITC0084

1

0

0

1

1

0

A1/A4a

22

AAA

Lujugira/Mutika

Intokatoke

ITC0082

1

0

0

1

1

0

A1/A4a

31

AAB

Pisang raja

Pisang Raja Bulu

ITC0843

1

0

0

1

1

1

A1/A4a/B

24

AAB

Nadan

Lady Finger

ITC0582

1

1

0

0

0

1

A1/A2/B

25

AAB

Pome/Prata

Foconah

ITC0649

1

1

0

0

0

1

A1/A2/B

26

AAB

Pome/Prata

Prata Ana

ITC0962

1

1

0

0

0

1

A1/A2/B

14

AAA

Cavendish

Grande Naine

ITC0180

1

1

0

0

1

0

A1/A2/A4b

15

AAA

Cavendish

Petite Naine

ITC0654

1

1

0

0

1

0

A1/A2/A4b

16

AAA

Cavendish

Poyo

ITC0345

1

1

0

0

1

0

A1/A2/A4b

32

AAB

Nendra padaththi

Pisang Rajah

ITC0243

1

1

0

0

1

1

A1/A2/A4b/B

19

AAA

Gros Michel

Gros Michel

ITC0484

1

1

0

1

1

0

A1/A2/A4a

23

AAA

Ibota

Yangambi km5

ITC1123

1

1

0

1

1

0

A1/A2/A4a

The presence of two different types of gene-pools was identified in ten accessions. A1/A4 combination was found in four acuminata triploids (17, 18, 21 and 22) including Lujugira/Mutika (21 and 22), in one AAB (31) and one AA diploid cultivar (12) as well. Single AAB accession contained A2/A4 combination (29). In these cases again, the A4 component was of the subgroup represented of ssp. banksii, ssp. errans and ssp. microcarpa (A4a). In three AAB cultivars (24, 25, and 26), A1/A2 combination was identified.

Three different types of acuminata gene-pools were identified in five triploid acuminata cultivars and unexpectedly in one AAB hybrid. According to the subgroups in gene-pool A4 Gros Michel and Yangambi Km5 were the combination of gene-pools A1, A2 and A4a, while the three Cavendish accessions (14, 15 and 16) along with the AAB accession Pisang Rajah (32) represented A1, A2 and A4b combination of the acuminata genomes.

Discussion

In the present study, we compiled the sequence of the full length ribosomal locus of Musa acuminata and about half of Musa balbisiana. Comparative sequence analysis of the overlapping region in the two species revealed that the highest sequence divergence between M. acuminata and M. balbisiana was localised in the 5′ETS region. Therefore this region was selected for analysis of sequence divergence in wild type genotypes. The sequence analysis of this rDNA region proved to be powerful in discriminating M. acuminata wild types but revealed no sequence diversity in M. balbisiana giving additional support for the general belief that the genome of M. balbisiana is genetically less diverse.

Four gene-pools represent M. acuminata wild types

Six to nine subspecies have been defined within M. acuminata wild types according to morphological classification (Shepherd 1990; Simmonds and Shepherd 1955). Crouch et al. (1999) demonstrated that different (RAPD, VNTR and AFLP) molecular markers may provide different aspects of genetic relationship in Musa species. Based on microsatellite and RFLP markers, wild diploid M. acuminata genotypes were classified into four groups, i.e. banksii, zebrina, malaccensis and burmannica/burmannicoides by Carreel et al. (1994) and Grapin et al. (1998). Later Ude et al. (2002) suggested three genetic subspecies in M. acuminata based on AFLP genotyping of acuminata wild types.

In the present study we used the sequence information of the 5′ETS region, which has been widely used to infer phylogenetic relations of closely related species since the concerted evolution of the individual rDNA loci within a species allowed their use as single copy genes (Sallares and Brown 2004; Sanderson and Doyle 1992). Our results confirmed the classification suggested by Carreel et al.(1994) and Grapin et al. (1998) by suggesting four groups/gene-pools within M. acuminata wild type genotypes after analysing eight morphologic subspecies represented by ten individuals. One gene-pool was formed by ssp. microcarpa, ssp. banksii and ssp. errans representing the Papua New Guinea distribution area, while ssp. burmannica, ssp. burmannicoides and ssp. siamea formed an other gene-pool representing the North-western distribution area. Two additional gene pools were represented by ssp. zebrina, which are the subspecies of the eastern part of the distribution area (Indonesian islands) and ssp. malaccensis representing peninsular Malaysia, respectively. When summarising the above listed previous ones and our results based on direct comparative analysis of the genomic DNA of M. acuminata, it is indicated that subspecies identified by morphologic traits may be possibly grouped into four gene-pools representing geographic areas and the present day subspecies representing possible geographic varieties only. However, the analysis of a larger number of M. acuminata wild types will be necessary to reveal all gene-pools present in this species.

Genetic structure of the accessions

The highly variable nature of the 5′ETS region made possible the development of an easily applicable marker system capable for identifying the acuminata component(s) present in the ten triploid (AAA), ten AAB, eight ABB, three AA cultivars, two AB cultivars, one AS hybrid and one tetraploid ABBT hybrid. A4 was the most frequently identified gene-pool present in 23 out of the 35 analysed genotypes (65%). A1 and A2 gene-pools were equally represented being 17 (48.5%) and 15 (42.8%) times present in the analysed genotypes. However, gene-pool A3 representing the Northern distribution area of M. acuminata by ssp. burmannica, ssp. burmannicoides and ssp. siamea did not contribute to the acuminata component of the analysed accessions. Considering that the present day accessions are mostly the result of farmers’ and breeders’ selection of the preferential contribution of gene-pool A4 represented by ssp. microcarpa, ssp. banksii and ssp. errans may suggest that this gene-pool provides advantageous phenotypic traits in this respect like parthenocarpy, which possibly arose from M. acuminata ssp. banksii and/or ssp. errans as suggested by Carreel et al. (2002).

The most popular genotypes for production like Cavendish and Gros Michel showed the highest degree of “hybridisation” possessing three different acuminata components in their genome. All these genotypes contained A1 (zebrina) and A2 (malaccensis) gene-pool components along with A4 subgroup ssp. microcarpa, ssp. banksii and ssp. errans type rDNA (A4a) in Gros Michel and Yangambi KM5, while in Cavendish the gene-pool A4 component was similar to the rDNA identified in the putative malaccensis hybrid (A4b).

The applicability of the presented marker system on revealing the genomic structure of present day cultivars was supported by some previous publications. Our results support the isozyme study of cultivated bananas and plantain reported by Lebot et al. (1993) that genes contributed by M. acuminata to AAB Popoulou/Maia Maoli (No. 30) were similar to banksii complex in PNG. Our results also confirmed the findings of Bhat et al. (1995) who suggested that Safet Velchi (ABcv; 34) was closely related to M. acuminata ssp. malaccensis. In this case our marker system indicates the presence of balbisiana gene-pool along with A2 gene-pool, which is represented by ssp. malaccensis. Similarly Auvuchanon et al. (2001) revealed that ‘Namwa’ (ABB) had an acuminata chromosome set, which hybridised with ‘Kluai Pa Pattalung’ (derived from ssp. malaccensis) by using GISH. In accordance to this finding, our marker system also verified the presence of malaccensis type gene-pool in ‘Namwa Khom’ (ABB; 43). Hamon et al. (2003) reported that Lujugira/Mutika (21 and 22) AAA genotypes were genetically close to ssp. banksii and zebrina, which was confirmed by our findings that along with A1 gene-pool represented by ssp. zebrina, gene-pool A4a was also participating, which was represented by ssp. banksii, ssp. errans and ssp. microcarpa, however, the ratio of the gene-pools are not known yet. Recently Swangpol et al. (2007) reported that using non-coding cpDNA sequences revealed that ‘Hom Thong’ (Gros Michel group), an AAA triploid cultivar, is likely an intersubspecific hybrid between wild M. acuminata ssp. malaccensis and ssp. banksii. Our results confirm this finding, but by use of a nuclear marker system we could reveal the presence of a third acuminata component A1 representing ssp. zebrina type gene-pool in this cultivar. Therefore, we may say that the developed rDNA based marker system could be a valuable tool in elucidating the genetic structure of the acuminata component of hybrid cultivars. However, the precise identification of balbisiana component was still not possible.

Dissident genotypes

During this study we revealed that the ssp. malaccensis genotype (ITC0250) was a putative hybrid, since it possessed two different types of ribosomal loci. One type represented a distinct gene-pool, which we considered latter as malaccensis type; while the other one was related to gene-pool A4 represented by ssp. microcarpa, ssp. banksii and ssp. errans. However, an additional SNP could differentiate the allele present in ssp. malaccensis against the rest of the subspecies representing the A4 gene-pool.

Apart from the above mentioned M. acuminata ssp. malaccensis accession, several other unexpected genotypes were observed. Minor component balbisiana rDNA was detected in M. acuminata ssp. banksii (ITC0766) and acuminata type rDNA sequences were observed in M. balbisiana Lal Velchi (NEU0051). Preliminary quantitative PCR analysis of the ssp. banksii (ITC0766) revealed that about 3% of the rDNA loci represent balbisiana type rDNA and additionally a genomic clone of Pyruvate decarboxylase (accession number FJ268999) gene also showed the presence of balbisiana sequences (unpublished). This gene yields an RsaI site present in the balbisiana type allele (data not shown). Since in Eumusa only one chromosome pair carries 18S–5.8S–26S rDNA loci, the presence of ‘illegitime’ genome possibly reflects ancient hybridisation events of the two species and subsequent recombination of the two genomes followed by several back crosses resulting in the present day “wild type” genotypes (Fig. 3c). The example of ssp. banksii is the first molecular evidence to our knowledge on the recombination of acuminata and balbisiana genomes in putative AB hybrids, which were also evidenced earlier by cytology (see for review Shepherd 1999). A similar hybridisation event could explain the presence of three acuminata gene-pools in the formation of the AAB hybrid Pisang Rajah (ITC0243), where one of the acuminata ancestors could have been hybrid genotype like ssp. malaccensis in our case. On the other part in case of Lal Velchi, a second DNA sample originating from J. Dolezel’s laboratory (Olomouc, Czech Republic) the presence of the acuminata component could not be verified, possibly indicating a mistaken identity of the original sample.

On the contrary, no acuminata type rDNA loci were detected in two ABB hybrids, Kluai Tiparot (ITC0652) and Simili Radjah (Peyan) (ITC0123). No acuminata components in Kluai Tiparot could be detected by use of additional acuminata/balbisiana rDNA specific markers or by the analysis of the pyruvate decarboxylase gene (data not shown). These results were confirmed on an independent DNA sample originating from IAEA Laboratories Seibersdorf/Austria. These results may indicate that Kluai Tiparot (ITC0652) either is not an acuminata/balbisiana hybrid or that its genome harbours an incomplete acuminata genome, which does not include the loci represented by the used markers. However, the analysis of the pyruvate decarboxylase gene for Simili Radjah (Peyan) verified the presence of acuminata genome elements, possibly indicating the presence of an incomplete acuminata genome (data not shown). Concerning these two cultivars, the absence of acuminata rDNA was reported by Nwakanma et al. (2003) as well.

In the present paper we described the use of the rDNA loci for the determination of acuminata gene pools participating in hybrid genotypes. The 5′ETS sequences presented in this paper may serve as starter for building a sequence database, which could enable us to refine the identification of genomes in hybrid accessions. However, the observations obtained with dissident genotypes suggest that similar marker system(s) are needed in representing other parts of the musa genome (chromosome or chromosome arm specific markers) to get a better insight into the organisation/origin of the hybrid genomes.

The existence and high frequency of unexpected genotypes shows the plasticity of the banana gene pool. Both the transitory nature of species boundary of closely related species like M. acuminata and M. balbisiana allowing hybrid formation and the overlapping distribution areas of subspecies/varieties contributed to the high level of genetic variability of the present day cultivars. But the fine structure of these genomes can be traced by DNA based molecular tools as published by Raboin et al. (2005) and the present paper.

Acknowledgments

The present work was funded by Bioversity International (formerly International Network for Improvement of Banana and Plantain (INIBAP)). The authors are grateful to Prof. Benchamas Silayoi and Assoc. Prof. Dr. Somsak Apisitwanich for guidance and technical advice in bananas and Assoc. Prof. Dr. Jaroslav Doležel for kindly providing us with DNA from some banana wild type species.

Copyright information

© Springer Science+Business Media B.V. 2009