Banana Bunches at FHIA
Diploid Breeding at FHIA and the Development of Goldfinger (FHIA-01)
Phillip Rowe and Franklin Rosales
In 1959, executives of United Fruit Company had just witnessed the final elimination of the Gros Michel banana variety by race 1 of Fusarium wilt. Fortunately, the Cavendish clones are resistant to race 1 and were planted to save the export industry. However, no other natural variety would be available to replace Cavendish if it too were later destroyed by an uncontrollable disease. These farsighted men started the current FHIA breeding program (the program was donated to the Fundación Hondureña de Investigación Agrícola (FHIA) in 1984) to develop a “man-made” banana which would be resistant to such an anticipated new disease. This disease now exists in the form of race 4 of Fusarium wilt. In addition, black Sigatoka has the potential to eliminate Cavendish for export and the burrowing nematode could greatly reduce productivity. Black Sigatoka could become resistant to all approved fungicides which are used as chemical control measures, and additional nematicides could be banned (two have already been banned by regulatory agencies) due to their dangers to consumer health and the environment.
While protection of the export trade was the main reason for beginning the breeding program, the hybrids developed in this program now also provide the only practical solution to the drastic yield reductions (up to 50%) presently being caused by black Sigatoka on the plantains and cooking bananas grown for local consumption. About 90% of the world production of bananas and plantains is for domestic food, and most of the major producing areas have become infested with this disease in the last few years.
Current financing of the breeding program is by the International Development Research Centre (IDRC) of Canada, the Windward Islands Banana Growers Association (WINBAN), the government of Honduras, the government of Ecuador, the Natural Resources Institute (NRI) of the United Kingdom, and the United Nations Development Program/International Network for the Improvement of Banana and Plantain (UNDP/INIBAP).
Diploid Breeding – History and latest developments
The most important activity in banana and plantain breeding has always been development of agronomically superior disease-resistant diploid hybrids. These diploids are the main source of genetic variability in breeding new commercial type hybrids since the triploids they are crossed onto are fixed in terms of genetic make-up. Progenies from crossing diploids onto the different seed-fertile triploid banana, plantain and cooking banana clones are tetraploids. Selected tetraploids are evaluated as potential new commercial hybrids or as parental lines in subsequent cross-pollinations.
Development of SH-2095, which was the first diploid with outstanding bunch features, was the result of the most exhaustive efforts which have been made in the different facets of genetic improvement. SH-2095 was the one agronomically superior hybrid selected during the first ten years of the program, and it is in the pedigrees of most of the subsequent diploids which have been selected from the segregating populations.
The hybrid which required the second most intensive cross-pollinations for its development was the SH-3142 nematoderesistant diploid. SH-3142 was the one hybrid selected from the few obtained by pollinating 10,000 bunches of the almost sterile Pisang Jari Buaya accession which is resistant to the burrowing nematode. In contrast to Pisang Jari Buaya, SH-3142 is readily usable as both a pollen and seed parent in cross-pollinations.
In addition to being resistant to the burrowing nematode, SH-3142 was found to be resistant to the race 4 of Fusarium wilt from Taiwan in the initial screenings for resistance to this disease. This hybrid is one of the parental lines of the SH-3362 diploid which is resistant to the race 4 in Australia. Now, a commercial type tetraploid hybrid derived from crossing SH-3142 onto the Dwarf Prata triploid has shown apparent resistance to race 4 in Australia. This tetraploid, which has been given the code name of FHIA-01, is of vital importance since race 4 has the potential to destroy the current Cavendish commercial cultivars and the only control measure for this disease is genetic resistance. FHIA-01 is discussed in detail later.
Individual diploid hybrids with improved bunch characteristics and resistance to one particular disease have been available for several years. A primary objective in diploid breeding has been to develop agronomically superior hybrids with multiple disease resistance. The outstanding parental line for this series of crosses is the recently selected SH-3723. SH-3723 has exceptional bunch features (Fig. l) and has possibilities of being resistant to the two major diseases and the burrowing nematode. It is known that SH-3723 is resistant to black Sigatoka and resistances to race 4 of Fusarium wilt and the burrowing nematode are in its pedigree from both its SH-3248 and SH-3362 progenitors.
The one weakness of SH-3723 is that it has poor pollen and cannot be used as a male parent. It was learned this year that this hybrid produces an average of about two seeds per bunch when pollinated, and it is being multiplied for massive pollinations to obtain adequate quantities of progenies for subsequent evaluation and selection. It is expected that some of the hybrids produced from crossing onto SH-3723 will have the desirable agronomic and disease resistance features of this parental line and be pollen-fertile. Such selected hybrids with SH-3723 parentage would provide possibilities for using them in cross-pollinations onto seedfertile triploids to develop new commercial-type bananas, plantains and cooking bananas with multiple disease resistance.
The inherent danger in having only one source of resistance to a disease is that no alternative would be available if the pathogen were to become capable of attacking that source. However, the poor bunch features of most of the accessions which provide an array of genetic diversity for disease resistance have handicapped development of agronomically superior hybrids from the different sources of resistance. All the accessions which are resistant to black Sigatoka have very inferior bunch characteristics. Until last year, the only bred diploid which was resistant to black Sigatoka and also had advanced bunch features was SH-3437. The resistance of SH-3437 came from the IV-9 accession of Musa acuminata ssp. burmannica.
Last year, the SH-3681 diploid with the II-357 M. a. ssp. malaccensis genes for resistance to black Sigatoka was selected after exhaustive long term efforts to incorporate the resistance of this accession into hybrids with improved bunch qualities. Development of SH-3681 is considered a major accomplishment in providing an additional breeding line with resistance to this disease. However, in view of the current difficulty and expense in controlling black Sigatoka, additional sources of resistance would give even greater security for breeding resistant commercial hybrids.
The first progenies from cross-pollinations between agronomically superior diploids and the parthenocarpic Lidi and two wild M. a. spp. siamea accessions which are resistant to black Sigatoka were evaluated this year. Positive results from these crosses readily indicate the value of the SH-3362 diploid (which has an outstanding bunch size) for crossing with these resistant accessions which have inferior bunch features. The influence of SH-3362 is expressed in the bunch features of SH-3745 which was derived from the 1-131 (siamea) x SH-3362 cross. SH-3745 is highly resistant to black Sigatoka and was selected out of a population of 250 segregating hybrids from crosses onto the I-131 and II-334 (siamea) accessions. By contrast, the SH-2989 diploid (which was derived from crosses onto the burmannica accession several years ago) was the only plant with an improved bunch from a population of 2,500 hybrids. From further crosses with SH-2989, the exceptional black Sigatoka resistant SH-3437 diploid was developed. Similarly, now that SH-3745 has provided the siamea genes for resistance in a hybrid much superior to the accessions, it is expected that further crosses with SH-3745 will also result in outstanding hybrids with this new source of resistance.
A total of 21 additional diploids were selected from the segregating populations this year. Seven of these selections were from the SH-3437 x SH-3362 cross made in breeding for resistance to both black Sigatoka and race 4 of Fusarium wilt.
Breeding Apple-Flavored Bananas
In Asia, Brazil, and Australia where consumers have a choice between Cavendish and the apple flavored (more tart) fruit of the Prata and Silk cultivars, the latter is preferred. All the clones which have this apple-like flavor are susceptible to Fusarium wilt and have small bunches. Thus, they are not suitable for export.
The Dwarf Prata (AAB) clone was brought from Brazil in 1981 for evaluation as a fixed triploid parental line in breeding disease-resistant hybrids with this unique flavor. Now, the SH-3481 hybrid, which was derived from crossing the SH-3142 nematode-resistant diploid onto Dwarf Prata, is considered to be the best tetraploid produced to date. This hybrid was given the code name of FHIA-01 for its entry in the International Musa Testing Program (IMTP) sponsored by INIBAP for testing the reaction of seven FHIA hybrids to black Sigatoka in six Latin American and African countries. FHIA-01 is popularly known as Goldfinger. Its exceptional plant and bunch features (Fig. 2) with no treatment to control black Sigatoka make FHIA-01 an excellent candidate to become the first bred hybrid to be grown commercially.
The desirable features of FHIA-01 include the following:
1. Resistant to race I of Fusarium wilt
2. Highly resistant to black Sigatoka
3. Resistant to race 4 of Fusarium wilt
4. Tolerance to the burrowing nematode
5. Large bunch size.
6. Strong plant, support large bunches with no propping.
7. Good plant architecture.
8. Good post-harvest green life.
9. Strong neck.
10. Apparently resistant to crown rot.
11. Good flavor, distinctively more tart than Cavendish when first ripened and becomes similar to Cavendish in the advanced stages of ripeness.
12. Diced fruit does not oxidize.
13. Cold tolerant.
14. Ripens to a golden yellow without refrigeration or ethylene treatment.
15. Excellent flavor and texture when boiled green.
FHIA-01 has a plant height very similar to that of the semi-dwarf Valery which was the most widely grown Cavendish cultivar before it was replaced by the shorter Grande Naine. However, FHIA-01 is a stronger plant than Grande Naine and this plant sturdiness is expected to compensate for the taller height in with standing damage from strong winds. In other comparisons with Grande Naine (which is receiving the prescribed chemical disease control treatments), FHIA-01 is slightly slower to flower and the fruit reaches harvest grade about two weeks later. However, the productiveness of FHIA-01 is far superior to that of Grande Naine if no fungicide or nematicide applications are made.
The disease-resistances and hardiness of FHIA-01 could result in its being even more important for domestic consumption than for export. It can be grown by small holders in countries where Cavendish could not be grown because the expensive pesticides required were beyond the economic means of these farmers. In addition to the six countries already mentioned where it has been tested in the IMTP plots, this hybrid is being evaluated extensively in Australia. It has also been distributed to many other countries by INIBAP for local evaluation. Indeed, FHIA-01 could soon become the universal dessert banana, especially since it also has shown tolerance to marginal soil fertility conditions and extended periods of inadequate rainfall.
To illustrate the potential impact of FHIA-01 in East Africa where boiled or steamed cooking bananas are a dietary staple for 20 million people, bunch sizes of this hybrid are about twice as large as those of the popular Nyamwihogora East African cooking banana at sea level. If FHIA-01 is adapted to the higher altitudes of this region, it could be a productive black Sigatoka-resistant cooking as well as dessert banana. Dessert bananas are not currently readily available in East Africa since their traditional highland cooking bananas are not very appetizing for eating raw when ripe.
Improvement of Cavendish Banana Cultivars through Conventional Breeding
J.F. Aguilar Morán
Fundación Hondureña de Investigación Agrícola
Abstract
In their article “Banana breeding: polyploidy, disease resistance and productivity”, Stover and Buddenhagen (1986) reported the results of the evaluation of female fertility in Cavendish banana cultivars. They showed that the pollination of a few hundred bunches of ‘Valery’ (AAA) and other Cavendish clones with pollen from diploids did not yield seed. The authors concluded that “the apparent seed sterility of Cavendish cultivars (without any research to determine or overcome the blocks) precluded their use as female parents in conventional breeding programs”. The scientific community accepted these observations as fact and did not carry out additional tests, because the commercial cultivars of banana for export are all triploid and parthenocarpic. The triploid condition of the Cavendish banana causes them to produce many sterile eggs, and the process of parthenocarpy allows the development of fruit without ovule fertilization. On the assumption that Cavendish cultivars have low fertility, the Banana and Plantain Breeding Program at the Honduran Foundation for Agricultural Research (FHIA), starting in 2002, pollinated 20,000 bunches, approximately 2 million fingers, of the Cavendish cultivars ‘Grand Naine’ and ‘Williams’ with pollen from 10 Cavendish cultivars for the development of Cavendish tetraploids. As a result, 200 seeds with 40 viable embryos were obtained, from which 20 tetraploid hybrids were developed. These results confirmed the assumption that Cavendish cultivars have low fertility, which allows their use in conventional breeding methods to create new progenies. The selected tetraploid progenies were crossed with improved FHIA diploids for the development of second generation triploid hybrids. As a result of this cross, two hybrids with resistance to black leaf streak and Fusarium wilt race 1, have been preselected. These hybrids exhibit similar performance to known Cavendish cultivars.
INTRODUCTION
From its early beginnings in the 1970s, the banana industry relied on the ‘Gros Michel’ (AAA) cultivar. However, because of its susceptibility to Fusarium oxysporum f. sp. cubense race 1, the causal agent of Fusarium wilt, this cultivar was replaced in the 1950s with cultivars from the Cavendish subgroup (AAA genome). In 2003, Pearce reported on the probable disappearance of the Cavendish banana within a period of 10 years. This article generated a lot of attention with the media, the banana research community, banana producers and consumers. The reason behind this gloomy forecast was the susceptibility of the Cavendish cultivars to Fusarium oxysporum f. sp. cubense tropical race 4 (Foc TR4), to Mycosphaerella fijiensis, the causal agent of black leaf streak (BLS), and to nematodes.
Considering the needs of the banana export industry to substitute existing Cavendish cultivars for Cavendish-like cultivars with resistance to BLS and Foc TR4, the Honduran Foundation for Agricultural Research (FHIA) initiated work to develop such new cultivars, building on the 52 years of experience of the program in conventional banana breeding. The FHIA banana breeding program, even though not located in a region where Foc TR4 is present, has developed three banana hybrids resistant to Foc TR4 (FHIA-01, FHIA-18 and FHIA-25) (Daly et al., 2007). These three hybrids all have the FHIA-improved diploid SH-3142 (resistant to BLS and Foc TR4) as their male parent.
The objective of this work was to develop Cavendish tetraploids derived from the crossing of Cavendish cultivars, to increase the likelihood of having a high concentration of Cavendish desirable traits. These tetraploids were then used as female parent in crosses with SH-3142 to produce second-generation triploid Cavendish hybrids.
FEMALE FERTILITY OF COMMERCIAL CAVENDISH CULTIVARS
The use of conventional breeding in the development of disease-resistant Cavendish hybrids has received very little attention since it was reported by Simmonds (1962), Stover and Buddenhagen (1986) and Stover and Simmonds (1987) that Cavendish cultivars are female sterile. Breeding to develop new dessert bananas has thus depended largely on ‘Gros Michel’ or its shorter version mutant ‘Highgate’.
According to Simmonds (1962) and Stover and Simmonds (1987), the majority of banana fruits are sterile due, probably in varying degree, to specific female sterility genes, triploidy and chromosomal structural change. The relative importance of each factor depends on the nature of the clone under investigation. Seed development in many edible bananas cultivars depends not only upon maternal conditions but also upon pollen availability.
Stover and Simmonds (1987) indicated that the sterility of the clones cultivated for the international trade is a consequence of the high inherent female sterility. Within the Cavendish group, seed production was not observed in the many thousands of pollinations conducted. By contrast, ‘Gros Michel’ can produce one or two seeds per bunch if pollinated or if grown next to a pollen source.
Stover and Buddenhagen (1986) reported in their review the results of trials to determine the female fertility of Cavendish cultivars. In testing for seed fertility by crossing with male fertile diploids, the authors found that the Cavendish cultivars never yielded seeds. Most extensive pollination attempts were made with the Cavendish cultivar ‘Valery’, in which a few hundred bunches were pollinated. Several other Cavendish clones were also pollinated to a more limited extent, but also without success. The authors concluded that “it is this apparent seed sterility (without any research to determine or overcome the blocks) that has precluded the use of the “ideal triploids” as parents”. The first exception to the absence of Cavendish input into conventional banana breeding was the successful yield of seed from a wild Musa acuminata ssp. malaccensis when pollinated with ‘Valery’ pollen in 1964. Several dwarf diploid hybrids were obtained (the dwarf character from ‘Valery’) which were further used in diploid breeding and are now in the pedigree of all the advanced dwarf diploids (Stover and Buddenhagen, 1986). A second exception came in the 1990s when FHIA obtained one seed from the cross of ‘Williams’ as female parent and ‘SH-3142’ as male parent; this seed developed into a tetraploid plant that was named ‘FHIA-02’.
From 1960 to 2002, conventional breeding was not considered a viable method for the development of a Cavendish replacement resistant to BLS and/or Foc TR4, because of the assumption that Cavendish was female sterile. In 2009, it was reported that the sequencing of the banana genome would generate knowledge for the development of a transgenic banana with resistance to Foc TR4 (Grim, 2009), but though developing a replacement for Cavendish through transgenesis is possible, the banana industry would again be dependent on a single cultivar. However, in 1992, Phillip R. Rowe, then head of the banana breeding program at FHIA, received from Frederick Novak of the International Atomic Energy Agency (IAEA), the banana clone ‘Novaria’, an irradiated mutant derived from ‘Grand Naine’, to assess its resistance to BLS and its commercial value under growing conditions at La Lima, Honduras. After the evaluation, Rowe used ‘Novaria’, which was susceptible to BLS, to make crosses with improved male diploids. These crosses produced seeds with no endosperm or embryo. Rowe thought that the fertility of ‘Novaria’ was due to the effects of irradiation. This information provided the basis to evaluate the fertility of the Cavendish clones and a program for the development of a Cavendish replacement through conventional breeding was initiated in 2002. The first task of FHIA’s breeding program was to determine rates of male and female fertility of Cavendish cultivars, to subsequently outline strategies for hybrid development.
EVALUATION OF MALE FERTILITY OF CAVENDISH CLONES
To determine the male fertility of commercial cultivars of Cavendish, the diploid wild Musa acuminata ‘Calcutta IV’ was used as female parent and Cavendish cultivars were used as male parents. ‘Calcutta IV’ is not parthenocarpic, and the fruit only develops if at least one ovule is fertilized. When ‘Calcutta IV’ is pollinated with fertile diploid pollen, it produces 1,500 to 2,000 seeds per bunch. Plants of ‘Calcutta IV’ that were pollinated with Cavendish produced an average of 900 seeds per bunch. This first result indicated that commercial cultivars of Cavendish are not male sterile but have intermediate male fertility.
The results of these trials provided the basis for an evaluation of the pollen fertility of Cavendish cultivars, using diploid, triploid and tetraploid females with high fertility. As a result of these crosses, diploid, tetraploid and triploid hybrids were generated, respectively. It was also shown that crosses with diploid females generated the largest number of seeds and hybrid plants, followed by crosses with tetraploid females and finally triploid females.
These trials showed that the pollen of Cavendish cultivars was fertile and that pollen meiosis occurred. As a result of this division, several types of male gametes should be produced, of which probably only the haploid gametes (n) are viable during the process of pollination and fertilization. This conclusion was reached because there were no triploid or tetraploid seedlings in the offspring of diploid females, no pentaploid or hexaploid seedlings in the offspring of triploid females, and no tetraploid or pentaploid seedlings, in the offspring of tetraploid females.
The above results have helped to formulate breeding strategies for the development of a Cavendish replacement. If the pollen from a triploid has the ability to fertilize an egg from a triploid cultivar, and as a result of this cross, a tetraploid hybrid can be obtained, then it is possible to improve triploid females. This situation is different from that understood by Simmonds (1962) and Rowe and Richardson (1975), who believed that banana breeding was essentially male breeding, because it was impossible to improve female triploids. The development of intermediate breeding materials or female tetraploids which will be crossed with improved diploids resistant to Foc TR4 can result in the development of second-generation triploid hybrids with new or similar organoleptic traits to the initial triploid Cavendish.
EVALUATION OF FEMALE FERTILITY OF CAVENDISH CLONES AND
DEVELOPMENT OF FIRST AND SECOND-GENERATION CAVENDISH-TYPE
HYBRIDS
The female fertility of Cavendish cultivars was confirmed at FHIA through the pollination of 20,000 bunches of Cavendish with pollen from 10 male Cavendish parents. As a result of these crosses, 200 seeds were obtained. From these seeds, 40 embryos were rescued and 20 plantlets were developed. This result shows that, although the female fertility of the Cavendish cultivars is very low, these cultivars should not be classified as sterile.
The tetraploid hybrids developed (Cavendish × pollen donor) contain the three sets of Cavendish chromosomes and one set of chromosomes from the male parent. This makes it possible to develop diploid gametes from the tetraploid female parent which would have two of the three chromosome sets of Cavendish and would be fertilized with the haploid gamete from the male parent in order to generate second-generation triploid hybrids with 66% Cavendish genes.
During 2009, tetraploid females, derived from Cavendish cultivars, were pollinated with the improved diploid SH-3142 developed by FHIA. From the field evaluation of these hybrids, we preselected hybrids with resistance to BLS and Foc race 1. These hybrids have confirmed the potential of this breeding approach to improve Cavendish.
CONCLUSIONS AND PROSPECTS
Our work has shown that Cavendish cultivars have intermediate male fertility, and should thus not be classified as male sterile, as previously reported. The rate of female fertility of Cavendish cultivars, depending on biotic and abiotic factors, is approximately one seed for every 100 pollinated bunches. Crosses between triploid Cavendish cultivars can thus generate tetraploid Cavendish hybrids, showing that it is possible to improve Cavendish triploids through conventional breeding methods. The development of tetraploid hybrids with three complete sets of chromosomes from the Cavendish mother and one set of chromosomes from the male parent is the basis for the creation of secondgeneration triploid Cavendish type hybrids.
A sustainable banana industry requires the continuous development of improved Cavendish-like cultivars, to overcome new pest and disease threats and/or possible loss of resistance to existing constraints. A continuous program to develop improved males and females is needed, supported by long-term public and/or private financial support. Based on our research, the development of more than 20 additional female-fertile tetraploids derived from crosses between Cavendish and a male donor is needed to produce a minimum of 2,000 second-generation triploid hybrids, which would provide the opportunity to pre-select 200 hybrids resistant to BLS and Foc race 1. The resistance of these pre-selected hybrids should be tested under field conditions in Asia or Australia, where Foc TR4 is present. The selected male parent of the second-generation triploid hybrids needs to be resistant to BLS and Foc (race 1 and TR4).
Our results indicate that FHIA’s method of banana improvement is appropriate for the development of new Cavendish hybrids through the re-synthesis “de novo” of a Cavendish with resistance to BLS and Foc TR4.