Author: Oselebe, H O; Obi, I U; Uguru, M I
Date published: August 1, 2010
Journal code: ACSC
(ProQuest: ... denotes formulae omitted.)
Bananas and plantains (Musa spp) are major staple foods in the tropic and sub-tropic regions of the world, where they provide about 25 % of the carbohydrates and 10 % of the daily calorie intake for over 70 million people. Cultivated varieties are mainly triploids (2n = 3x = 33), derived from intraspecific hybridisations within M. acuminata (A genome) and interspecific hybridisation between M. acuminata and M. balbisiana (B genome). Plantains and bananas have a peculiar biology, which tend to impede rapid breeding progress. These include irregular meiotic behaviour due to triploidy (Ortiz and Vuylsteke, 1996), ploidy and genome variation in crosses (Tenkouano et al., 1998a) partly due to variation in ploidy level of gametes produced during micro and megasporogenesis in male and female parents (Oselebe and Tenkouano, 2009; Oselebe et al 2006a). Others include long generation cycle (18 - 24 months from seed to seed), large land requirement (6m^sup 2^ per plant) and cost of extensive field evaluations for large populations. Hypothetically, the breeding of triploid hybrids of Musa could be approached from several angles: 4x - 2x, 2x - 4x and 2x - 2x crosses. However, in these crosses, genome and ploidy variation would be common and may complicate the ability to predict both the outcome of crosses and the performance of product of such crosses, for important economic traits (Tenkouano et al., 1998a). An understanding of the pattern of these variations is critical to the formulation of procedures for the prediction of hybrid performance prior to complex field evaluations that may be time consuming and costly. The prediction of hybrid performances is difficult due to lack of appropriate genetic models. In most cases, hybrid performances are predicted using genetic models based on disomic inheritance of traits in families derived from nested or factorial mating schemes. In this instance, progeny performance is predicted from mid-parent mean with the assumptions that there is an equal contribution of parents to their offspring, that the parents are inbred and unrelated and that the traits under consideration are strictly determined by additive inheritance (Panter and Allen, 1995; Bernado et al., 1996). These assumptions may not be suitable for Musa populations, because maternal and paternal contribution to the triploid offspring may not be equal (Tenkouano et al., 1998a and b). In addition, the assumption of inbred and unrelated parents may not be met in most Musa breeding populations, due to a 'background noise', resulting from the development of the breeding lines from very few parents as a result of low male and female fertility of the landraces. The determination of genetic parameters for predicting hybrid performance in Musa would therefore, be dependent on defining the relationships of progeny to parents in different crosses and an understanding of the parental contributions to offspring. This will lead to the definition of efficient strategies to obtain triploid clones from tetraploid-diploid, diploid-tetraploid or diploid-diploid parents. The objective of this work was to estimate genetic relationships based on meiotic mechanisms in parents and redefine models for predicting hybrid performance for different crosses.
Materials and methods
Study site and genetic materials
The study was carried out at International Institute for Tropical Agriculture (IITA), High Rainfall Station Onne, Southeastern Nigeria (4o 43' N, 7o 01' E 10 masl). IITA Onne Station is found in the densely populated Niger Delta region of Rivers State. Detailed characteristics of the station have earlier been described (Ortiz et al., 1997). The progenies for evaluation in this experiment were developed from two mating schemes: factorial-mating involving 4x - 2x and 2x - 2x parents and bi-parental crosses involving 2x and 4x individuals as female and male parents, respectively. Musa primary tetraploid plantain hybrids including OSH20, OSH27, OSH35, OSH45 and OSH46 were crossed to five primary diploid banana hybrids OSH38, OSH49, OSH52, OSH60 and OSH63 in a 4x-2x breeding scheme. In addition, four diploid banana hybrids: OSH52, OSH53, OSH60, and OSH63 were crossed to three diploid plantain hybrids: OSH10, OSH29, OSH31 and two diploid banana hybrids: OSH38 and OSH53. Also three diploid plantain hybrids: OSH12, OSH29 and OSH62 were crossed to three tetraploid plantain hybrids: OSH32, OSH42, OSH51, in a 2x - 4x crossbreeding scheme. These parental lines were developed at the study site. Pedigree details of these Musa lines are shown in Table 1. At maturity zygotic embryos were extracted from seeds and germinated in vitro. The parental clones were micropropagated clonally using meristematic tissues from shoot tips. Seedlings were acclimatized in the nursery prior to transfer to the field. Progenies and parental clones were established in the field under alley cropping with parental clones as controls.
Pollen samples were collected from parental clones between 7:30 and 10:30 a.m. Pollen grains were dislodged from the stamen, spread on a microscope slide and stained with acetocarmine glycerol gelly (Marks, 1954). One hundred pollen grains each (from four microscopic fields) from eight slide preparations per parent were observed under a Leitz Diaplan binocular light microscope (×400 magnifications). The diameters of ten randomly selected deeply stained grains were measured with the aid of a graduated eyepiece. The most frequent pollen diameter was considered to be that of normal haploid (n) pollen size. A mathematical approach to the evaluation of pollen diameter was used (Veronessi et al., 1988; Tondini et al., 1993). The diameter of n pollen of tetraploids was comparable to 2n pollen diameter of diploids and bigger than n pollen of diploids. Pollen grains were classified as unreduced (2n) or giant pollen, if they had about 1.26 (i.e. those ≥ 160 µm) or 1.59 times the linear dimensions of haploid pollen, respectively (Darlington, 1937). The number of haploid (n), diploid (2n) and giant (4n) pollen grains was recorded per parent. This was used as an estimate of the products of microsporogenesis and an indication of the range of gametes produced by parents of specific ploidy levels. Data on pollen quantity, variability and size were collected and subjected to the analysis of variance using the GLM procedure in SAS.
The ploidy status of the breeding populations was determined by flow cytometry (FCM) analysis following to procedures outlined in Oselebe et al (2006b), based on the postulate that variation in DNA content can be equated with variation in ploidy status. To release cell nuclei, about 50 mg of Musa leaf mid-rib was chopped with sharp razor blade in a petri dish with 0.5 ml ice-cold Otto 1 buffer (0.5 M citric acid monohydrate, 0.5% Tween 20). A further 0.5 ml Otto 1 buffer was added and the suspension mixed with a pipette. The suspension of the released nuclei was filtered through a 50 ìm nylon mesh and kept at room temperature. To stain the DNA, 2ml Otto 11 buffer (0.4 M anhydrous Na^sub 2^HPO^sub 4^) containing 4ìg/ml DAPI (4-6-diamidino-2-phenylindole) was added. Fluorescence of DAPI-stained nuclei was analysed using a Partec PAS 11 Flow cytometer (Partec GmbH, Germany) to ascertain the nuclear DNA content and determine the ploidy status of the genotypes (Dolezel 1997). The Flow cytometer was standardized using a known diploid 'Calcutta 4'and a known triploid ('Obino l'Ewai') as an internal reference standard (Pillay et al., 2000). Gain of the instrument was adjusted so that the G^sub 1^ peak of nuclei isolated from the control diploid plant was on channel 50, while that of the triploid was on channel 75 and setting kept constant during analysis of samples prepared from the breeding populations to compare their peak or histogram (Fig.1). Peaks on higher channels would indicate higher ploidy levels (Pillay et al., 2000). Ploidy levels of the progenies of crosses were recorded and ploidy frequencies expressed in percentages.
Predicting Hybrid Performance
A modified mid-parent approach for predicting hybrid performance was proposed (Tenkouano et al., 1999b), based on the following equation: ...
Where, H^sub ij^ indicates the expected value of the hybrid produced from the ith and jth parents. P^sub i^ and P^sub j^ are the observed values of the ith and jth parent, respectively. There are two components of the equation: a) an additive component, which is the weighted average of parental phenotypes, and b) a multiplicative component, which reflects heterotic or inbreeding effects. The terms c^sub i^ and c^sub j^ are the relative contributions of parents i and j, respectively to their progeny. In a disomic situation, genomic contribution of parents to the offspring are equal, i.e. ci = cj (Panter and Allen, 1995; Bernado et al., 1996). In Musa however, it is postulated that secondary triploid Musa hybrids receive two chromosomes from their tetraploid maternal parent for every chromosome donated by their paternal diploid parent. Therefore, (c^sub i^ ≠ c^sub j^) provided that 2n gametes were not formed. The terms fii and f^sub jj^ represents the probability of two alleles being identical at any locus in parents i and j, respectively and is indicative of their homozygosity (inbreeding) level. F^sup ij^ refers to the coefficient of relationship (similarity) among parents i and j. The three terms: f^sub ii^, f^sub jj^ and f^sub ij^ may be calculated using pedigree information, or molecular data, or a combination thereof (Tenkouano et al., 1998). The logarithm term included is analogous to Nei's (1972) formula for the calculation of genetic distance and reflects the postulate that the difference between the expected progeny performance and the mid-parent value would vary as a function of genetic distance among parents. If parents were similar genetically, the logarithm term would approach zero and the performance of the progeny would be expected to approach the mid-parent value for the trait under consideration. This is the case from the progeny obtained from selfing an inbred line, which will be expected to express the same genotypic value as the inbred line. In contrast, crossing two unrelated parents should lead to an F^sub 1^ that expresses hybrid vigour, the magnitude of which would depend on parental contribution and inbreeding status. In other to predict hybrid performances, estimates of relationships between parents and progenies generated from pollen analysis of parental lines and ploidy analysis of progenies from different crosses would be substituted in the modified mid-parent formula for predicting hybrid performances in inter-ploidy crosses.
Results and discussion
The aim of breeding schemes involving 4x and 2x lines in Musa populations is to produce 3x progenies that have superior agronomic and post-harvest characteristics. In order to predict the outcomes of 2x - 2x, 2x - 4x and 4x - 2x breeding, it was important to examine the frequency distribution of ploidy classes in gametes produced by the 4x and 2x parents which would be an indication of their contribution to their offspring. This was based on the postulate that parental contribution to progeny in 4x - 2x crosses would be 1:1 for 2x offspring, 2:1 or 1:2 for 3x offspring and 3:1 or 2:2 for 4x offspring. Significant differences among 2x and 4x genotypes were observed for mean pollen diameter (P < 0.01) and pollen stainability (P < 0.05), but not for pollen quantity and most common pollen diameter (Table 2). There were no significant sample effects over time and sample by genotype interaction was also not significant. The tetraploid parents had a higher mean pollen diameter (128.2 ìm) than the diploid lines (117.7 ìm), while the diploids had a higher stainability value (97.7%), compared with the tetraploids (Table 3). The significant difference in pollen diameter between diploid and tetraploid parents is analogous to their DNA content (Ortiz et al., 1998) and supports the postulate of Dessaw (1988) that pollen derived from normal microsporogenesis should contain chromosome numbers that reflect the ploidy state of the sporophyte. Based on this postulate, the mean pollen diameter should therefore be a function of the genome size, because the nucleus and cytoplasm increase as chromosome number increases. The higher stainability value of the diploids in this study may be indicative of normal microsporogenesis, balanced gametes and hence pollen viability. Only well rounded and deeply stained pollen grains were counted and regarded as being viable in line with Ortiz et al., (1998). On the other hand, meiotic irregularities in 4x microsporogenesis may have led to the production of higher number of unbalanced gametes, hence, lower stainability values. These irregularities may be related to pairing abnormalities during meiosis leading to unbalanced and/or aborted pollen grains, and reduced stainability. Although the 4x parents had a reduced stainability value, there was no significant difference in pollen production between them and the 2x parents. Of a particular note, however, is the non-significant difference in the most common pollen size between the two parental lines. Normally, the most common pollen size is regarded as the haploid (n) pollen of a particular clone. 2n pollen of diploids is comparable to n pollen of tetraploids and bigger than n pollen of diploids (Verronesi et al., 1988). This implies that the diameter of n pollen of tetraploids should be equal to 2n pollen of diploids, each having the same chromosome content (2n = 2C for 2x parent; n = 2C for 4x parent). Nonsignificant difference in the n pollen diameter between the diploids and the tetraploids may therefore indicate a predominance of 1x pollen in the tetraploid male parent which could lead to ploidy polymorphism in progenies from crosses in Musa populations having 4x as a male parent.
Diploid individuals may produce gametes with 1, 2, or 4 times the basic chromosome set, considering only euploid products. Whereas predominating gametes are normal (n = 1C) from diploid Musa lines, restitutional gametes do occur and are widespread among accessions and hybrids (Ortiz, 1997), resulting from failure of first meiotic division (first divisional restitution [FDR]) or second meiotic division (second divisional restitution [SDR]) or of both, referred to as double restitution (Tavoletti et al., 1991). From the later cases, 2n gametes (2n = 2C) and 4n gametes (2n = 4C) are formed. Ortiz (1997) also suggested that at least one dominant allele may be involved in the inheritance of 2n pollen in Musa. Other factors that may contribute to the formation of gametes with varying chromosome sets (n = 1x = 1C, n = 3x = 3C and n = 6x = 6C) include spindle disorientation, abnormal cytokinesis, incomplete homeology of the genomes that make up the genotype of the line or highly divergent genome sets which could reduce pairing (Ude et al., 2002). Disjunction in this later case for different configurations may lead to the formation of gametes with 1n, 3n and 6n chromosomes. Upon fusion of gametes from 4x and 2x individuals, progenies with 1 to 12 sets of the basic chromosome complements may therefore be recovered (Table 4).
Ploidy variation in progeny
Based on analysis of microsporogenesis, two postulates was drawn: crosses between 4x and 2x parents would predominantly produce 3x progenies in the 4x - 2x direction and 2x progenies in the 2x - 4x direction, assuming that macrosporogenesis was normal. Progenies from 2x - 2x crosses would also produce predominantly 2x progenies. Results of the FCM analysis from this study were complementary. 4x - 2x crosses produced predominantly 3x progenies, whereas 2x - 4x crosses gave predominantly 2x progenies. Individuals recovered from 2x - 2x crosses were predominantly diploids (Fig 2). This suggested a pattern of unequal contribution of the parents with respect to 3x progeny (n = 2x = 2C by the 4x parent plus n = 1x = 1C by the 2x parent) from 4x - 2x crosses. The reverse was the case for 2x - 4x crosses where parental contribution was equal with respect to the progeny (n = 1x = 1C plus n = 1x = 1C).
Predicting hybrid performance
Progress in breeding for higher yield in Musa would be dependent on identifying and accessing genes and gene combinations most likely to produce higher heterosis and the selection of appropriate parental lines whose combinations are likely to lead to hybrids of the desired ploidy level and heterosis. The modified mid-parent approach (Tenkouano et al., 1999b) could be used to estimate hybrid performance with the substitution of estimates of parental contribution based on present results. In 4x - 2x crosses which constitute the most reliable means of generating triploid hybrids (Oselebe et al 2006b), in which 3x progenies are predominantly recovered (> 90 %), and in conformity with the assumption of normal meiosis in both parents, hybrid performance could be predicted using the following equation: ...
The above equation may also apply for 4x - 4x crosses based on our findings that 4x male parents produces predominantly n = 1C gametes.
Other specific applications of this formula may suffice for different crosses. In disomic situations, there may be two options, which would be dependent on segregation outcomes during meiosis. In a breeding scheme that utilizes 2n gametes produced through unilateral sexual polyploidization in 2x - 2x crosses to produce 3x progenies, unequal parental contribution (n = 1x = 1C + n = 2x = 2C or n = 2x = 2C + n = 1x = 1C) would be expected. In this breeding scheme, hybrid performance could be predicted using the equation below: ...
In a normal situation, parental contributions to progeny will be equal from 2x - 2x crosses (Panter and Allen, 1995; Bernado et al., 1996), and hybrid performance could be predicted thus:
This equation would also apply to 2x - 4x crosses where equal contribution of parents to their progeny is expected. This is substantiated by present results in relation to 4x microsporogenesis and the preferential survival of monoploid gametes. In all the cases, the terms f^sub ii^ and f^sub jj^ denotes the probability that two alleles are identical at any locus in parents i and j, respectively; f^sub ij^ represents the coefficient of relationship among the parents. All the three values would be calculated from various combinations of pedigree and molecular data (Tenkouano et al., 1999a). The ability to predict hybrid performance prior to field trials would reduce the cost of Musa breeding, both in terms of labour, time and money. It is worth noting that, it may not be easy to predict accurately the outcome of any cross in Musa. However, outcomes with the greatest percentage most of the time may be taken as the normal situation.
The authors thank Dr Abdou Tenkouano for his invaluable contributions to this research and Mr Christian Dimkpa, Ms Josephine Okoro and Mr Abraham Shaibu for assistance. Financial support from the Directorate General for Development Cooperation (DGDC, Belgium) and the United States Agency for International Development (USAID) is gratefully acknowledged.
Bernardo R, Murigneux A, Karamura Z (1996) Marker-based estimates of identity by descent and alikeness in state among maize inbreds. Theor. Appl. Genet. 93: 262-267
Darlington CD (1937) Recent advances in cytology (2^sup nd^ Ed). P.Blakinston's sons and Co, London
Dessauw D (1988) Etude des facteurs de la sterrilite du bananier (Musa spp.) et des relations cytotaxonomiques entre M. acuminate et M. balbisiana Colla. Fruits 43:539- 558, 615-638, 685-700
Dolezel J (1997) Application of flow cytometry for the study of plant genomes. J. Appl. Genet. 38(30): 285-302
Marks GE (1954) An acetocarmine glycerol jelly for use in pollen fertility counts. Stain Tech 29:277
Nei M (1972) Genetic distance between populations. Amer. Nat. 106: 283-292
Ortiz R (1997) Occurrence and inheritance of 2n pollen in Musa. Ann. Bot. 79:449-453
Ortiz R, Ulburghs F, Okoro JU (1998) Seasonal variation and apparent male fertility and 2n pollen production in plantain and banana. HortScience 33: 146-148
Ortiz R, Austin PD, Vuylsteke D (1997) IITA High Rainfall Station: Twenty years of research for sustainable agriculture in West African humid forest. HortScience 33(6): 969-972
Ortiz R, Vuysteke D (1996) Recent advances in Musa genetics, breeding and biotechnology. Plant Breed. Abst. 66: 1355-1363
Oselebe HO, Tenkouano A (2009) Ploidy versus gender effects on inheritance of quantitative traits in Musa species. Australian J. Crop Sci 3(6):367-373
Oselebe, HO, Tenkouano A (2008) Combining ability for important agronomic traits in Musa L. diploid populations. J. Agric, Biotech and Ecology, 1: 93 - 104
Oselebe HO, Tenkouano A, Pillay M (2006a) Ploidy variation of Musa hybrids from crosses. African J. Biotech. 5 (11): 1048 - 1053
Oselebe HO, Tenkouano A, Pillay M, Obi IU, Uguru MI (2006b) Ploidy and genome segregation in Musa L. breeding populations assessed by Flow Cytometry and Random Amplified Polymorphic DNA Markers. JASHS 131(6): 780 - 786
Panter DM, Allen FL (1995) Using best linear unbiased predictions to enhance breeding for yield in soybean: 11. Selection of superior crosses from a limited number of field trials. Crop Sci. 1: 215-222
Pillay M, Nwakanma DC, Tenkouano A (2000) Identification of RAPD markers linked to A and B genome sequences in Musa L. Genome 43:763-767
Tavoletti S, Mariani A, Veronesi F (1991) Cytological analysis of macro- and microsporogenesis of diploid alfalfa clone producing male and female 2n gametes. Crop Sci. 31: 1258-1263
Tenkouano A, Crouch JH, Crouch HK, Vuylsteke D, Ortiz R (1999a) Comparison of DNA marker and pedigree-based methods of genetic analysis of plantain and banana (Musa spp.) clones. I. Estimation of genetic relationships. Theor. Appl. Genet. 98: 62-68
Tenkouano A, Crouch JH, Crouch HK, Vuylsteke D, Ortiz R (1999b) A comparism of DNA marker and pedigree methods for the genetic analysis in platain and banana (Musa spp.) clones. 11. Predicting hybrid performance. Theor. Appl. Genet. 98: 69-75
Tenkouano A, Crouch JH, Crouch HK, Ortiz R (1998a) Genetic diversity, hybrid performance and combining ability for yield in Musa germplasm. Euphytica 102: 281- 288
Tenkouano A, Ortiz R, Vuylsteke D (1998b) Combining ability for yield and plant phenology in plantain-derived populations. Euphytica 104: 151-158
Tenkouano A, Crouch JH, Crouch HK, Vuylsteke D (1998) Ploidy determination in Musa germplasm using pollen and chloroplast characteristics. Amer J Hort Sci 33 (5): 889-890
Tondini F, Tavoletti S, Mariani A, Veronesi F (1993) A statistical approach to estimate the frequency of n, 2n and 4n pollen grains in diploid alfalfa. Euphytica 69: 109-114
Ude G, Pillay M, Nwakanma D, Tenkouano A (2002) Genetic Diversity in Musa acuminata Colla and Musa balbisiana Colla and some of their natural hybrids using AFLP Markers. Theor. Appl. Genet. 104:1246-1252.
Veronesi F, Mariani A, Tavoletti S (1988) Screening for 2n gamete producers in diploid species of the genus Medicago. Genet. Agr. 42:187-200
Oselebe, H.O. 1 2 *, I.U. Obi 3 and M.I. Uguru3
1International Institute of Tropical Agriculture, P.M.B. 5320 Oyo Road, Ibadan, Nigeria
2Department of Crop Production & Landscape Management, Ebonyi State University, Abakaliki, Nigeria
3Department of Crop Science, University of Nigeria Nsukka, Enugu State, Nigeria
*Corresponding author: email@example.com