A n n u a l R e p o r t 1 9 9 733

IntroductionGenetic transformation has been the subjectof much discussion in recent years. On oneside it has been described as a “bandwagon”and a useless bunch of “molecular tricks”,while on the other side, it has beensuggested that genetic transformation willresult in a second “green revolution” of suchmagnitude that all problems of foodinsecurity will become a thing of the past.The reality of course is likely to besomewhere in between these two extremes.While it is true that genetic engineeringalone will not provide the solution to feedingan ever expanding human population inyears to come, it is also clear that geneticengineering techniques already play animportant role in crop improvement and, incombination with conventional breeding,will provide the tools with which newimproved varieties will be developed at afaster pace in the future.

It is the intention of this paper todescribe briefly genetic transformation andidentify some of the constraints whichremain to be overcome. The paper will thendiscuss the potential impact that genetictransformation may have on the futureimprovement of banana and plantainproduction, and identify those areas whereit is believed genetic transformation has animportant role to play.

Plant genetictransformationMore than 120 plant species have alreadybeen successfully transformed and theseinclude representatives from the majorvegetables, fruits, trees, pastures andornamentals. Over 60 species are involved infield tests in 45, mainly industrialized,countries. In the USA, insect-resistantcotton and maize, virus resistant squash,several herbicide-resistant transgenic crops

including soybean are commerciallyproduced and different geneticallyengineered tomatoes with extended shelflife are sold in supermarkets. In total, thesetransgenic crops were grown on more than12.5 million hectares in 1997 and this area isexpected to reach some 29 million hectares(equivalent to about 40% of arable land inEurope) in 1998. These figures demonstratethat plant transformation has now become aroutine methodology whereby the key issueis no longer the production of individualtransgenic plants but the generation oftransgenic populations with a uniform andpredictable pattern of transgene expression.In other words, plant transformation itselfhas followed a natural progression,developing from early promises of potentialduring the experimental phase (from theend of 80’s to the mid 90’s) to an industrialscale (second half of the 90’s) with realexamples of practical applications.

It should be clear that a major drivingforce behind investment in transformationresearch was the desire to have betterintellectual property protection on the end-products than was possible with varietiesbred through conventional means. To thisend, breeding companies have formedalliances with biotechnology companies, inorder to ensure access to the technologiesand end-products. As a result, manycompanies that are actively involved in plantbreeding have become leaders in one ormore fields of biotechnology and theirinventions in the quickly evolving, but costlyfield of plant transformation are frequentlyprotected by patents.

Technical aspectsPlant transformation can be defined as theintroduction and stable integration of genesinto a plant nuclear genome and theirexpression in the transgenic plant. Planttransformation involves the introduction of

Focu

s Pap

er IIGenetic transformation for

the improvement of bananas - a critical assessmentL. Sági, S. Remy and R. Swennen

Laboratory of Tropical Crop Improvement, KUL

34A n n u a l R e p o r t 1 9 9 7

Focu

s Pap

er II at least two genes (a selectable marker gene

and a gene of interest), but the number ofsimultaneously introduced genes may reachmany more.

During the last decade, wide (andsometimes wild) ranges of transformationtechniques have been tested. This has beenparticularly the case for the transformationof graminaceous monocotyledons, whichwere previously considered to berecalcitrant. It is now however clear thatprobably all the major regenerable cell typescan be transformed through one of just twotechniques, either by particle bombardmentor through co-cultivation withAgrobacterium tumefaciens.

Particle bombardment (or biolistictransformation) uses accelerated metal(gold or tungsten) microprojectiles coatedwith DNA to penetrate and deliver foreigngenes into plant cells, which are thenselected and regenerated into plants. On theother hand, Agrobacterium tumefaciens, a soil bacterium, transforms its hosts byintegrating a segment (called T-DNA) of itstumor-inducing plasmid into the nucleargenome. The transfer of this T-DNA isregulated by a complex process with theinvolvement of numerous bacterial genes(the majority called virulence genes) thatare located outside the T-DNA. Thisinteresting feature allows for the use of T-DNA as a vehicle to introduce virtually anygene into plant cells.

Whichever technique is used, the precisetransformation conditions for a specificregenerable cell type require optimizationfor each species, and possibly even, eachvariety.

A critical factor in transformation is theselection of transformed cells, since for eachtransformation event, the introduced gene(s) will be incorporated into only afraction of the cells subjected totransformation. This selection is performedwith the help of a selectable marker genewhich confers resistance to chemical agents,such as antibiotics, herbicides, or aminoacid/sugar analogues, which are otherwisetoxic to plant cells.

The transformed nature of the transgenicplants must be verified. This is done at themolecular level by DNA hybridizationanalysis to demonstrate incorporation of thetransgenes in primary transformants as wellas in their progeny. If large numbers ofplants are to be analysed a PCR pre-screening may be carried out. Verification of transformation must also be done byenzymatic or immunological assay orphenotypical observation for the geneproduct to show the expression (or on thecontrary, the silencing) of the integratedgenes in parallel with correct controls.

Specificcharacteristics of genetictransformation

Source of variationOne advantage of genetic transformationover conventional breeding lies in theavailability of sources of genetic variability.While conventional breeders are limited tothe gene pool that exists in sexuallycompatible species, this barrier does notapply for genetic transformation. Genesisolated from lower organisms (viroids,viruses, bacteria and fungi) as well as fromhigher animals (including human genes) canbe introduced into plant cells, and suchgenes have been shown to be correctlyexpressed in the transformed plants. Thisallows for the generation of transgenicplants expressing useful traits, which couldnot be achieved by classical breeding. Thelist of preliminary yet encouraging examplesis wide ranging, and includes the productionof vaccines as well as biodegradable plasticsin such plants.

Introduction of large DNA segmentsIt is often assumed that planttransformation involves only single genes,and is therefore of limited usefulness in cropimprovement, as most agronomicallyimportant traits are multigenic. However, ithas been shown that up to 600 kb DNA canbe introduced into soybean cells with oneshot by particle bombardment and similarly,Agrobacterium efficiently delivers andintegrates at least 150 kb foreign DNA intothe plant genome. This size range is morethan sufficient for the introduction ofmultiple genes. Since it is known thatsimilar sizes of DNA can be incorporated inthe genome of prokaryotic as well aseukaryotic organisms such as yeast, this isnot altogether surprising.

Constraintsremaining

Somaclonal variationAn important scientific concern relates tothe tissue culture component oftransformation. In many culture systems,the rate of somaclonal variation, i.e. randomgenetic changes that occur during in vitroculture, can be correlated to the time theexplants spend in tissue culture. This isparticularly the case for banana.Regeneration systems requiring a minimumof time in culture, but still compatible with

transformation, are therefore needed. Itshould be noted that somaclonal variation isinherent to in vitro culture and is notrelated to the transformation techniques.Therefore, somaclonal variation could beeliminated by a tissue culture-freetransformation method, but at present, thisis available only for Arabidopsis thaliana.

Gene expressionHighly variable levels of transgene expressionand complete suppression (also called genesilencing) are also a cause for concern.Variable gene expression is thought to be dueto the random integration of the transgeneinto the plant DNA, thus resulting in differentgenetic contexts around the transgene in theindependent transgenic plants. On the otherhand, gene silencing appears to be related tomulticopy integration and/or DNAmethylation, although the precise mechanismremains to be resolved.

The major consequence of unpredictabletransgene expression is the cost of having toscreen large numbers of plants for thedesired transformants and the continuingevaluation of stability in subsequentprogenies.

Transformation efficiencyThe constraints outlined above call for aprecise definition of transformationefficiency. In the development oftransformation technologies, scientists striveto achieve high transformation frequencies,i.e. they try to generate as many transformedplants as possible. However, in practice theissue is not the absolute number oftransgenic plants but the number of usefultransgenic events that fulfil the requiredconditions. Therefore, precise genetargeting, which is at present more advancedfor the transformation of animal species, isalso needed for the improvement of planttransformation efficiencies.

Management and field release of geneticallytransformed plantsOnce transgenic plants have been produced,their further management and field releaserequires the smooth coordination of a largenumber of factors. Field release must beauthorized by national governments throughregulatory bodies. In many developingcountries, operational regulatory bodies arenot yet in place, and this causes delays inimplementing field testing of transgenicplants. Field testing should involve an activebreeding programme to ensure that plants

A n n u a l R e p o r t 1 9 9 735

are evaluated correctly and preciseobservations are taken. Such collaborationwith a breeding programme should also formthe basis for integrated breeding activities,where plant transformation providestransgenic parental lines that, after fieldtesting, could be incorporated into existingcrossing and selection programmes.

Why is genetictransformationimportant in Musa?Genetically resistant banana and plantainvarieties are the basis upon whichsustainable improved production can bedeveloped. Since a very wide range ofvarieties of bananas and plantains are grownworldwide, each adapted to a specificecoregion and selected for specific eating orcooking qualities, a similarly wide range ofimproved varieties is required.

Classical breeding of bananas is hamperedby long-generation time, triploidy andsterility of most edible cultivars. Sources ofresistance to many of the major pests anddiseases are known in both landraces andwild species. However, landraces are oftensterile and cannot be used in breeding, whilecrosses involving wild species result in thetransfer of many unwanted traits togetherwith the desired resistance genes.

Furthermore, there are certain diseasesfor which sources of resistance are notknown, an important example being bananabunchy top virus (BBTV).

Genetic transformation provides anopportunity for single genes or genecombinations, (such as those associatedwith disease resistance) to be extractedfrom the genome of the source organism andtransferred directly into the desired variety.This allows the variety in question to retainall its original characteristics, with thesimple addition of the desired trait.

Furthermore, since most banana cultivarsdo not produce seeds under naturalconditions, crosses with other varieties orspecies will not occur. In these cases, theintroduced gene remains confined to thevariety in which it has been introduced.

Status oftransformation in Musa

Two systems have resulted in theproduction of transgenic banana plants:particle bombardment of embryogenic cellsuspensions and Agrobacterium-mediatedtransformation of in vitro meristems.

Particle bombardment of embryogeniccell suspensionsThis technology combines particlebombardment transformation with aregeneration protocol using highlyregenerable suspension cultures containingfast-dividing and totipotent somatic cells. Itappears that this technology is limited onlyby the availability of cell cultures with asufficiently high capacity for plantregeneration. An average of 2-3 transgenicplants per bombardment can beregenerated from a cell suspension with agood morphogenic potential. Transformedplants are ready to be established in thegreenhouse six to eight months afterbombardment. This method is being usedwith some considerable success by severalresearch teams worldwide. At KUL alone,close to 900 independent transgenic lineshave been produced from the AAB plantaincv. ‘Three Hand Planty’and the Cavendish-type dessert (AAA) banana cv. ‘Williams’.These plants were transformed with eitherof two selectable marker genes and anumber of chimaeric genes including threedifferent antifungal genes and variousgenes isolated from banana viruses such asthe coat protein gene of banana bractmosaic virus. More than 200 independentlines have been transferred so far to thegreenhouse and a significant proportionhave been multiplied and prepared for fieldtesting.

Agrobacterium-mediatedtransformationThe Agrobacterium-based bananatransformation system as originallyreported, used explants containing theapical meristem or corm meristematictissues which were wounded bymicroparticle bombardment with uncoatedparticles. After a brief recovery period, thesemeristems were subsequently co-cultivatedwith Agrobacterium tumefaciensharbouring the plant transformation vectorin the presence of acetosyringone, a knowninducer of the Agrobacterium virulencegenes. Antibiotic resistant plants wereregenerated and DNA hybridizationdemonstrated that the transgenes wereincorporated into high molecular weightgenomic DNA and no residualAgrobacterium persisted in the plants. Inaddition, plantlets that had undergonemultiple rounds of propagation maintainedthese genotypic and phenotypic traits. Themain advantages of this method are (i) theease and speed of transgenic plantregeneration, (ii) the lack of requirementfor high-tech equipment or sophistication intissue culture and (iii) the assumption thatonly the desired DNA sequences are

transferred to recipient cells. However, thegenotype specificity of this technique inbanana needs to be thoroughly evaluated. Inaddition, several independent laboratorieshave found that a high incidence ofchimaeric transformants is characteristic ofthis method. This is probably due to the verylow chance that all cells contributing to theformation of new organs are accidentallytransformed in the organized explants.Recently different research groups haveobserved that regenerable cell suspensionscan also be transformed by Agrobacteriumat a high frequency. However, theinvolvement of cell suspensions makes thismethod slow and more expensive, thus lessattractive for routine use.

Introduction of multiple genesLong-term and multiple disease resistance islikely to be achieved by integrating severalgenes with different targets or modes ofaction into the plant genome. Technically,this can be done either in severalconsecutive steps or simultaneously. Particlebombardment of embryogenic cell cultures,the present method of choice for genetictransformation of banana, relies on the time-consuming development of highlyembryogenic cultures from a number oftarget cultivars, which makes consecutivetransformations impractical. It has beenshown at KUL that simultaneous genetransfer into banana can be performed byco-precipitation of a mixture of chimaericgene constructs onto microparticles beforebombardment. Transgenic plantsregenerated after bombardment with suchconstructs contained unlinked genes at afrequency of 70 – 80%. As a result of theseco-transformation experiments, transgenicbanana plants containing up to six differentgenes have been obtained. Theseobservations indicate that simultaneousbombardment of different plasmidmolecules may be a convenient way for theintroduction and perhaps co-expression ofmultiple genes in banana plants.

Currentapplications of genetictransformation in Musa

Fungal disease resistanceSince the most significant damage to bananaproduction is caused by fungal pathogens,the introduction of genes conferringresistance to fungal pathogens is a primaryresearch target. Efforts are presently

Focu

s Pap

er II

36A n n u a l R e p o r t 1 9 9 7

Focu

s Pap

er II focused on the recently described

antimicrobial proteins (AMPs) which arestable, cysteine-rich small peptides isolatedfrom seeds of diverse plant species. TheseAMPs have a broad anti-fungal spectrum andshow high in vitro activity against fieldisolates of Mycosphaerella fijiensis andFusarium oxysporum, the two main fungalpathogens of Musa, while they exert notoxicity to human or plant cells.

To date, five AMP genes have beenintroduced at KUL into embryogenic cellsuspensions of banana and plantain usingparticle bombardment. Large-scalemolecular and biochemical characterizationof the transgenic lines confirmed that a vastmajority contained the foreign genes. Forexample, PCR-screening of a total of 776putative transgenic plants revealed thatmore than 90% of them contained theselectable marker gene as well as the AMPgenes. Similar results were obtained by DNAhybridization of 56 plants, which showedthat 89.3% had the foreign geneincorporated in their genome.

To study the expression of antifungalproteins in the transgenic plants, the AMPgenes were placed under the control ofvarious regulatory sequences (promotors)which had previously been shown to drivehigh gene expression in banana as well as inother monocots. Using specific antibodies,the concentration of two antifungal proteinsin total leaf proteins was determined intransgenic plant extracts and results rangedfrom 0.05% to more than 1% depending onthe promoter used. Addition of theseextracts to germinating spores of a fieldisolate of Mycosphaerella fijiensis, thecausal agent of black Sigatoka disease,resulted in a significant inhibition of fungalgrowth making these plants likelycandidates to control this pathogen underfield conditions.

Further research will demonstratewhether these transgenic plants express thefunctional antimicrobial peptides at levelshigh enough to control fungal leaf or rootdiseases in the field. It has been shown thatone AMP gene is also expressed in the fruit,opening the opportunity to create resistanceagainst pre- and postharvest diseases suchas cigar-end rot and crown rot.

Virus resistanceEngineering resistance to banana bunchytop virus (BBTV) is another obviousobjective for banana transformation, sinceno natural resistance to this virus has so farbeen identified in the Musa genepool.Molecular virologists have gathered more

knowledge about BBTV in the last five yearsthan during the preceeding 100 years. Thegenome organization of BBTV, as well as itsmovement within the plant, is now startingto be understood in detail.

The further molecular characterization ofthis virus is ongoing, and this will assist indesigning various transformation strategiesto achieve protection against the virus bygene transfer. One of the strategies nowbeing tested in several laboratories is theexpression of various BBTV genes intransgenic banana plants in order tointerfere with the normal replication,encapsidation or movement of the virus.Another approach is the expression inbananas of heterologous antiviral proteinsthat are known to act by the inhibition ofviral replication or translation.

Similar strategies can also be consideredagainst another recently emerged DNA virus,the banana streak virus whose molecularresearch in laboratories worldwide hasalready revealed a number of interestingfeatures including its integration into thebanana genome.

The future oftransgenic researchin MusaTo date, no genetically transformed bananaor plantain plants are being tested in thefield. This is essentially due, as mentionedearlier, to the lack of functional regulatorybodies in the countries where testing wouldbe carried out. In view of the fact that moreand more transgenic plants are beingproduced by laboratories worldwide, thelack of regulatory mechanisms is likely tocontinue to be a major bottleneck toprogress. It is therefore recommended thatgreater efforts, at both national andinternational level, be made to resolve thissituation.

Major gaps in scientific knowledge relatedto Musa pests and diseases still exist. Inorder for further progress in genetictransformation to be made, these gaps mustbe filled. For instance it is remarkable thatconsidering the importance of fungaldiseases in banana, insufficient knowledge isavailable at the molecular, or even cellularlevel on the pathogenesis of fungi in banana.Much more information in this area must beobtained before studies on the molecularinteraction and signal transduction betweenthe banana host and the fungal pathogencan be conducted.

Besides the more advanced areas offungal and virus resistance, there are severalareas of banana transformation which arealso being investigated and may holdpotential for the future. These includeresistance to pests such as nematodes andinsects, tolerance to cold and water stress,improved fruit quality and the potential useof bananas as bioreactors.

ConclusionsThe genetic transformation of culturedcells and tissues has led not only to theproduction of the first transgenic bananaand plantain plants, but also to thegeneration of a large number of transgeniclines with agronomically useful genes.Several hundreds of independenttransgenic plants transformed with genesencoding antifungal proteins are availablefor field testing. Banana bunchy top,banana streak and bract mosaic viruseshave been isolated and their genomes arebeing characterized. The prospects forimproving bananas through the use ofmolecular techniques therefore are clearlydemonstrated. However, it should alwaysbe remembered that moleculartechnologies cannot stand alone, but arecomplementary to conventional breedingbased on hybridization. Indeed molecularbiology reveals the specific action andinteraction of a particular or a few gene (s)under controlled conditions while breedingdeals with a statistical process where largenumbers of genes are involved. It istherefore essential that Musa breedersand Musa biotechnologists work togetherto accelerate Musa improvement and notcompete against each other.

In view of the limited resources beingdevoted to research into Musaimprovement, and knowing the scale of theproblems to be overcome, it is importantthat donor agencies, and particularly theCGIAR, put emphasis on stimulatingcollaboration between successful breedinginstitutions and advanced biotechnologycenters. One recent initiative that aims tobring together all the major players inMusa improvement research is PROMUSA,the Global Programme for MusaImprovement. This programme waslaunched in 1997 by INIBAP and the WorldBank and provides the necessaryframework within which collaborativepartnerships can be developed andfostered. Further details of PROMUSA areprovided in Focus Paper 1.