Broadly speaking, there are two alternative approaches to the manipulation of genetic expression through transformation:
Transformation with novel genes to confer a new trait, or with modified genes to enable more effective performance of an existing function. These traits, e.g. herbicide tolerance, insect resistance and cold tolerance, are discussed below.
Transformation with genes which lead to the blocking of protein synthesis, through either antisense or ribozyme technology. With the antisense approach, a gene is transcribed in reverse direction, and an mRNA is produced that is complementary to, and thus able to bind, the normal message - blocking protein synthesis. Objectives can include reducing production of unwanted products by targetting particular enzymatic steps in a pathway, or viral resistance by targetting viral gene expression and/or replication. The expression of antisense RNA is a powerful method for the down-regulation of expression of the corresponding sense RNA (Harms 1992), but requires delivery of a large excess of antisense over sense mRNA because the antisense/mRNA hybrids are kinetically constrained (Ratner 1989). With the ribozyme approach, the RNA molecules that are synthesized have the capacity to cleave mRNA's of target genes - essential genes of insects or perhaps viruses. Being catalytic molecules, ribozymes are likely to be more effective at destroying the translational capacity of mRNAs than antisense transcripts (Harms 1992).
Successful application of genetic engineering involves a number of basic components:
identification and cloning of a gene of interest
addition of control sequences promoting appropriate expression
incorporation of a marker gene
transfer to target tissue which is competent for both transformation and regeneration
selection and regeneration of plants
confirmation of appropriate expression
Transformation methods were described in detail by Potrykus (1990). Included in these descriptions are some which are unproven or of limited use, e.g. the use of virus vectors, incubation of tissues in DNA solutions, pollen transformation, microlasers, electrophoresis, and macro and microinjection. Widely proven techniques for the production of transgenic plants are Agrobacterium mediated transfer, and the direct gene transfer methods - bombardment with DNA-coated microprojectiles, and electroporation or polyethylene glycol (PEG) treatment of protoplasts (Birch 1993).
Agrobacterium is a soil pathogen that naturally parasitizes plants by inserting a part of its own DNA into the host's genome. A. rhizogenes and A. tumefaciens transfer a portion of one of their large plasmids (termed Ri and Ti respectively) to the chromosomal DNA of a plant when the bacterium infects the wound. The transferred DNA (T-DNA) encodes functions that induce root formation in the former case and tumour formation in the latter. These natural transformation systems have been used to introduce non-T-DNA genes into plants, by inserting the foreign gene into the T-DNA itself, or by using a binary vector system, where the foreign gene is carried by an artificial T-DNA on a second plasmid or in a second bacterium (Tepfer 1990). One of the most important findings was that disarmed T-DNA (i.e. without functional oncogenes) can be transferred and integrated into the plant genome, allowing regeneration of transgenic plants (Binns 1990). Uncertainties remain concerning how the T- DNA crosses the cell wall, and how it becomes integrated. Under optimal conditions, transformation efficiency is very high and genes are stably integrated, but variation in host sensitivity has been a limitation. Agrobacterium infects nearly all dicots, and some conifers, but generally not monocots (Grimsley 1990, Laimer da Camara Machado 1992, Potrykus 1990). Even within a species, variation in sensitivity to infection can be substantial (Potrykus 1990, Binns 1990, Bergmann & Stomp 1992). In a study with Pinus radiata, host genotypes differed significantly in susceptibility to infection, although bacterial strain and bacterial strain × host genotype effects were not significant. A conclusion of this study was that a major component of the host-pathogen interaction may be the genetic and environmental factors controlling host plant cell division at the time of inoculation (Bergmann & Stomp 1992).
Protoplast methods basically involve removal of cell walls to yield a protoplast suspension, and then the use of chemical (PEG) or electrical (“electroporation”) treatments to increase membrane permeability to DNA transfer (Lindsay & Jones 1990). While the use of protoplasts eliminates a major barrier to the passage of DNA (the cell wall), regeneration of plants from protoplasts is difficult for many species.
Biolistic methods involve the acceleration of small metal particles coated with DNA to deliver DNA into plant tissues, through the cell wall. Acceleration of gold or tungsten particles is achieved by an explosive charge or by gas flow, the latter reportedly cheaper and more efficient (Finer et al. 1992, Takeuchi et al. 1992). Helium gas is used because of low mass and high diffusivity. The biolistic approach is relatively simple technically, avoids the requirement for removal of the cell wall, and is good for producing transient expression signals. The transition from transient to stable integrative transformation, however, is very inefficient. Transformation of chloroplasts has been achieved by particle bombardment (Potrykus 1990).
Comparative features of the major transformation methods were tabulated by Birch (1993):
Table 1. Features of Methods for Transgenci Plant Production (Birch 1993)
|Special T-DNA Constructs||Yes||No||No|
|Suitability for stable transformation:|
|Protoplast regeneration required||No||No||Yes|
|Tissue culture duration||Short||Short||Long|
Promoters are DNA sequences which initiate the transcription of adjacent gene coding sequences and which are modulated by additional adjacent DNA sequences known as enhancers (Strauss et al. 1991, Whetton & Sederoff 1991). Most commonly used experimentally is the cauliflower mosaic virus 35S promoter, a promoter active in most stages of development and in most plant tissues. Effective applications of genetic engineering, however, are frequently likely to be dependent on the use of promoters specific to particular tissues or developmental events. It is desirable, for example, for insect resistance to be expressed only after insects begin to feed and in the tissues under attack - requiring the splicing to the resistance gene of a combination of suitable promoters. The approach to finding such promoters is to identify any gene expression with the required tissue or developmental specificity, and then to isolate the regulatory elements and test for specificity. Many promoters directing specific patterns of gene expression are known:
wound-inducible - important in relation to transformation for insect resistance e.g. the potato and tomato proteinase inhibitor promoters (Klopfenstein et al. 1991, McNabb et al. 1990, McNabb et al. 1991), and the TR promoter from A. tumefaciens (Van Montagu 1992a)
ABA-inducible - the EM promoter from wheat (Charest et al. 1992)
specific to epidermal and xylem cells - the Phaseolus vulgaris pal promoter (Beachy et al. 1992)
specific to leaf mesophyll cells - the promoter from the gene encoding the small subunit of ribulose biphosphate carboxylase (Beachy et al. 1992, Campbell & Neale 1992)
specific to phloem - the A. rhizogenes roIC promoter (Beachy et al. 1992)
specific to roots (Han et al. 1992)
specific to the tapetal region (Leemans 1992)
This field is moving rapidly, but much remains to be done. An important question is the extent to which promoters will fuction across wide taxonomic boundaries. Some results are encouraging - in particular the demonstrated evolutionary conservation of tissue specific expression of ribulose biphosphate carboxylase activity between conifers and angiosperms (Campbell & Neale 1992).
Both experimentally and for the practical application of genetic engineering, a method of verifying transformation or selecting transformed from untransformed tissues is necessary. For experimental purposes, success of Agrobacterium mediated transfer has often been judged by expression of bacterial genes transferred - tumour formation, formation of hairy roots and production of opines (characteristic amino acids). More generally though, the addition of marker genes to the constructs is required. Marker genes employed include:
B-glucuronidase (GUS) - determined using a characteristic but lethal assay
Bar gene conferring resistance to the herbicide Basta (phosphinothricin)
neomycin phosphotransferase II (NPTII) conferring resistance to neomycin/kanamycin
firefly luciferase (a visual assay)
Herbicide or antibiotic assays which kill or suppress non-transformed cells, while allowing proliferation of transformed tissues, are attractive, but sensitivity of tissues to the selecting agent is critical. A high incidence of untransformed “escapes” has been demonstrated in some studies (Ellis, McCabe et al. 1992).
The insecticidal properties of crystal proteins produced by this bacterium have long been known, B.t. based bioinsecticides have been used since 1935, and the current annual market is estimated at $US 100 million (van Montagu 1992a). 22 different crystal proteins and their corresponding genes have been identified. Based on structural relationship and spectrum of activity, these are grouped into Lepidoptera-specific, Lepidoptera and Diptera specific, Diptera specific, and Coleoptera specific (van Montagu 1992a). Specificity of the proteins is determined mainly by their binding to specific receptors located in the membrane of the midgut of larvae. Insecticidal activity involves two steps: - binding to the receptor and then toxicity caused by integration into the membrane and creation of a pore (Frutos unpublished). One of the first successful applications of plant transformation technology for crop improvement, insect resistant tobacco, tomato, potato, and cotton engineered with four different crystal genes has been reported. Compared to other genes transferred to plants, insecticidal crystal protein genes are weakly expressed in transgenic plants (van Montagu 1992a). B.t. based insecticides have been used on some forest crops, and development of trees that can produce the toxin is underway in a number of species (Strauss et al. 1991). Development of resistance to B.t. toxins, e.g. in Lepidoptera, has been reported. Such resistance is not due to an active process like destruction of the protein, but to a “passive” process - individuals selected for resistance are those with a modified receptor, unable to be recognized by the activated toxin (Frutos, unpublished).
These proteins, occurring naturally in plants, animals and microorganisms, operate by binding tightly to proteinases and inhibiting their function, thus reducing the effective concentration of digestive enzymes. Examples are the serine proteinase inhibitors, blocking the action of common animal proteinases such as trypsin and chymotrypsin. At least six unrelated families of serine proteinase inhibitors are known, varying in reactive sites and the enzymes inhibited (Strauss et al. 1991). Reduced feeding by insects was reported for hybrid poplar transformed with a wound-inducible potato proteinase inhibitor (McNabb et al. 1990, Klopfenstein et al. 1991).
Several chitinase genes have been cloned and some transferred into tobacco, but effects on insect pests are yet to be reported (Strauss et al. 1991). Plant secondary products are a major means by which trees protect themselves against insects, and potential exists for importing encoding genes from suitable sources - e.g. the neem tree. Many natural resistance mechanisms in trees, however, are under polygenic control, are of unknown molecular basis, and are not likely to be amenable to genetic engineering in the near future (Strauss et al. 1991).
Stable resistance to insects is best achieved by the incorporation of multiple resistance genes - preferably including more than one class.
Since 1986, genetically engineered resistance has been reported against many viruses, e.g. alfalfa mosaic virus, cucumber mosaic virus, grapevine chrome mosaic virus, plum pox virus, potato leafroll virus, potato virus X, potato virus Y, soybean mosaic virus, tobacco etch virus, tobacco mosaic virus, tobacco rattle virus, tobacco streak virus (Laimer da Camara Machado 1992), potato virus S, potato virus M, pea early browning virus (Harms 1992) and tomato golden mosaic virus (Bejarano 1991). Representing at least ten different groups, these viruses differ substantially in morphology, genome organization, and replication strategy.
In most cases, resistance has been engineered by incorporating into the host genome a gene from the virus, most commonly, although not exclusively, the virus coat protein gene (Harms 1992, Hoekema et al. 1989, Beachy et al. 1992, Zaitlin 1991). Expression of the coat protein gene in transgenic plant cells interferes with the uncoating of the incoming virus. Resistance is frequently conferred also to other related viruses, and shows considerable resemblance to the phenomenon of virus cross-protection (Harms 1992). It has been suggested that genes leading to the accumulation of several non-functional viral proteins, including capsid proteins, insect transmission genes, proteases, and movement proteins, might also provide resistance against infection, or transmission between plants or between cells (Beachy et al. 1992). In an alternative approach, tobacco resistant to tomato golden mosaic virus was obtained by transformation with an antisense DNA sequence of a gene encoding a protein absolutely required for viral replication (Bejarano 1991).
Coat protein mediated resistance has been demonstrated in field trials, now aged up to four years, in several countries, and it is anticipated that virus resistant transgenic agricultural plants will be released for commercial use in the mid-1990s (Beachy et al. 1992).
Few reports exist of successful engineering of bacterial or fungal resistance. Tobacco transformed with genes for thionins (polypeptides that are toxic to bacterial and fungal pathogens) showed resistance when challenged with bacterial pathogens (Carmona 1991). The Septoria leaf spot and stem canker pathogen of poplar was found to produce proteinases which are inhibited by purified proteinase inhibitor II protein, and studies are underway with transgenic plants to determine the effect of the proteinase inhibitor gene on infection and development (McNabb et al. 1990). An alternative approach is to seek detoxification enzymes in pathogens producing phytotoxins, and to transfer the encoding genes to plants in which resistance is desired - this has been done experimentally with Pseudomonas syringae and tobacco (Harms 1992).
No transgenic plants resistant to nematodes have been reported to date, but a number of lines of research are being pursued (De Waele 1992):
Lytic enzymes and toxic proteins to kill nematodes. Lytic enzymes include chitinases and collagenases, and toxins include the B. t. toxin. Isolates of the latter with activity against nematodes have been identified.
Cloning of genes known to confer resistance, e.g. the Mi gene from Lycopersicon peruvianum which is effective against almost all Meloidogyne species attacking tomato.
Interference in the nematode-plant interaction, e.g. using proteins blocking recognition.
Usually a genetically simple trait with a clear unambiguous phenotype, herbicide tolerance is particularly amenable to genetic manipulation (Chaleff 1986b). Many herbicides act by inhibiting important enzymes. Three resistance strategies are available:
Glutathione-S transferases, for example, catalyze the conjugation of glutathione to an electrophilic centre of hydrophobic herbicides and thereby confer resistance to herbicides of the s-triazine type. Some crop species have these enzymes naturally - e.g. maize and sorghum can detoxify atrazine (Schulz et al. 1990). Transformation with the bacterial Bar gene encoding phosphinothricin acetyl transferase (PAT), which acetylates phosphinothricin, has given rise to phosphinothricin-resistant transgenic plants for both tobacco (Schulz et al. 1990) and wheat (Vasil et al. 1992). A possible source of genes which code enzymes that metabolize nonselective herbicides is the wide spectrum of microbial organisms which detoxify herbicides in the soil (Schulz et al. 1990).
Overproduction of the enzyme.
Glyphosate is a potent competitive inhibitor of the enzyme 5-enol-pyruvylshikimicacid 3- phosphate (EPSP) synthase, which is involved in the biosynthesis of aromatic amino acids. The use of EPSP synthase in combination with a powerful promoter, resulting in overexpression, has resulted in the production of glyphosate resistant petunia plants (Schulz et al. 1990). The herbicide phosphinothricin is a strong inhibitor of glutamine synthetase. Fusing of the alfalfa glutamine synthetase gene to the 35S promoter and transferring to tobacco resulted in over expression of the gene and a large increase in resistance to the herbicide in the transgenic plants (Schulz et al. 1990).
The aroA gene which encodes EPSP synthase has been isolated and sequenced from several organisms. A mutated bacterial aroA gene encoding a resistant EPSP synthase has been used to produce glyphosate resistant transgenic petunia plants (Schulz et al. 1990). The herbicidal effect of imidazolinones and the sultonylureas is in both cases due to inhibition of the enzyme acetolactate synthase (ALS), which is the first enzyme in the biosynthesis of the branched chain amino acids. Mutated forms of the enzyme, conferring herbicide resistance, are known (Schulz et al. 1990).
A gene coding for the winter flounder antifreeze protein was introduced into corn protoplasts by electroporation, and expression subsequently observed in the protoplasts (Cutler 1991). Expression was reported also in tobacco plants transformed with a gene construct encoding an antifreeze protein of ocean pout (Kenward 1992). A small proportion of antifreeze protein (10–20 ng in 120 ug of protein analysed) was sufficient to displace the survival curve for transgenic plantlets by approx. 1°C in the direction of increased freeze tolerance. Antifreeze protein has been inserted also into potato, but cold tolerance of the plants not tested (Garcia et al. 1992).
With the exception of cell wall proteins, wood components are products of reactions catalyzed by many enzymes, and genetic manipulation is therefore a matter of manipulation of genes which encode key enzymes in various biosynthetic pathways (Whetton & Sederoff 1991). The pathway leading to lignin formation is well characterized for many species, and many projects aimed at lignin modification are underway. General strategies involved are:
Reduction of lignin production. Enzyme regulation of the aromatic amino acid, phenyl-propanoid and monolignol pathways is under study in many laboratories, and enzymes within these pathways are logical targets for genetic engineering. Useful constructs for lignin reduction include those leading to overexpression of gene products intended to scavenge extracellular free radicals, and those causing underexpressing of gene products intended to polymerize monolignols (Dean & Eriksson 1992). Many studies concern genes encoding the enzymes cinnamyl alcohol dehydrogenase (CAD), the last step in the synthesis of monolignol precursors, and phenylalanine ammonia lyase (PAL) (Harry, Strauss & Sederoff 1991, Whetton & Sederoff 1991, Feuillet et al. 1992, O'Malley et al. 1992). These enzymes are encoded by single genes in Pinus, and by multigene families in angiosperms (Harry, Strauss & Sederoff 1991). The cDNA corresponding to eucalypt CAD has been cloned (Feuillet et al. 1992). Underexpression of lignin associated enzymes using antisense technology has been accomplished with CAD and a lignin-associated peroxidase (Dean & Eriksson 1992). Other target enzymes include:
Coumarate CoA ligase (CCL), an enzyme which catalyzes a reaction from the middle of the lignin biosynthetic pathway. This enzyme has been purified from loblolly pine and the gene is being characterized (Voo et al. 1992).
Modification of lignin composition. This includes the modification of softwood lignin to make it more readily hydrolyzed, by incorporation of the hardwood genes that produce syringyl lignin. In the simplest case, genes for only two enzymes would be needed; but many gymnosperm enzymes do not use angiosperm substrates readily, so that as many as seven hardwood enzymes might be required (Whetton & Sederoff 1991). Work is proceeding to introduce an angiosperm methyl transferase into a gymnosperm to give greater quantities of sinapyl alcohol (Dean & Eriksson 1992).
The use of xylem specific promoters will be important in the alteration of lignin quantity and content, since interference with lignification in non-vascular tissues, which might play a role in resistance to pests, should be avoided (Whetton & Sederoff 1991).
By contrast with the lignin biosynthetic pathway, those for cellulose, hemicellulose and extractives remain poorly understood (Whetton & Sederoff 1991). The generally poor understanding of the genetic basis of most wood properties is a major constraint to genetic engineering of these traits.
Cytoplasmic male sterility (CMS) has been reported in over 140 higher plant species. The commercial use of CMS lines as female parents was made possible by the discovery of a specific, dominant, nuclear restorer of fertility genes (Leaver 1992). Nuclear male sterility has been engineered by targeting expression of ribonucleases such as barnase, from Bacillus amyloliquefaciens, specifically to the tapetum of immature anthers, leading to tapetal degeneration and the arrest of microspore development (Leemans 1992). Genes which can restore fertility have been constructed by linking the tapetal promoter to barstar, the intracellular inhibitor of barnase. This hybrid system has been successfully introduced in oilseed rape, cauliflower, chicory, lettuce, tomato, cotton and corn (Leemans 1992).
Projects aimed at genetically engineering sterile trees of Pinus, Eucalyptus and Populus are underway (Strauss et al. 1992 unpub.). Two approaches are being attempted in this work:
Isolation as cDNA clones of genes expressed only during early stages of differentiation of male and female reproductive buds. Reproductive tissue specific genes from herbaceous species are also being used to screen pine cDNA libraries.
Isolation of tree homologues to floral homeotic genes identified in Arabidopsis and Antirrhinum. Such genes and their promoters should allow the construction of novel genes which cause sterility when inserted into trees.
The most common form of self-incompatibility in angiosperms is the single locus multi-allelic gametophytic system (S-locus). Some gene transfer work has been done, but the use of genetic engineering to transfer naturally occurring self-incompatibility systems or parts thereof to functional heterologous combinations remains a distant prospect (Olesen et al. 1992). The sensitivity of the S-locus to its genetic background is likely to be a constraint. Even less is known about the genetic control of cross incompatibility. A change in activity of adenylate cyclase has been reported in association with compatible vs incompatible pollinations in interspecific crosses in poplars, but the molecular basis of the phenomenon remains obscure (Olesen et al. 1992).
The expression of antisense RNA to ACC synthase in tomato fruits inhibits ethylene production, and ripening and softening of the fruits, an effect which can be reversed by treatment with exogenous ethylene (Theologis 1992). It has been suggested that this may offer a general method for preventing senescence in a variety of fruits and vegetables.
Enhanced rooting of cuttings of Eucalyptus grandis, E. dunnii, E. nitens (Haigh 1992), E. gunnii, Salix alba, Allocasuarina verticillata and some fruit tree species (Chandler et al. 1993) has been achieved by transformation with the natural root promoting gene from Agrobacterium rhizogenes. This can be achieved simply by dipping the bottom of the cutting into a culture of the bacterium. Although certain strains of the bacterium were found to be specific to individual eucalypt clones, one strain was suitable for a wide variety of clones (Haigh 1992). Plants are chimaeric, aerial growth is non-transgenic, and there is no risk therefore of release of bacterial genes through pollen or seed dispersal.
Symbiotic studies have targeted protein formation and gene expression specific to infected tissues, during the process of signalling between the symbionts, during the early stages of nodule formation, and in the assimilation of fixed nitrogen during later stages of nodule ontology (Graham 1992). Studies underway for Phaseolus, soybean, alfalfa and other crops have resulted in the identification of some proteins and the characterisation of a few genes. Attempts to select for enhanced levels of enzymes involved, e.g. PEP carboxylase, glutamine synthase and leghaemoglobin, however, did not result in improved nitrogen fixation in alfalfa (Graham 1992).
For forest tree species, studies are underway of protein synthesis and gene expression during mycorrhizal infection in Eucalyptus (Martin et al. 1992, Tagu et al. 1992), and of the regulation of gene expression during symbiosis in Casuarina (Fana et al. 1992). Recombinant DNA techniques are providing a valuable tool for the study of these processes, but commercially useful manipulation by genetic engineering remains a distant prospect.
Plants have a variety of mechanisms for coping with water stress, generally under the control of multiple genes. A common approach to their study has been the isolation of proteins expressed specifically during water stress, and the isolation and sequencing of corresponding cDNA clones (Mullet et al. 1992). Some apparent homology of dessication related genes among distantly related plants has been demonstrated (Iturriaga et al. 1992, Cairney et al. 1992). cDNAs encoding dessication proteins in the resurrection plant Craterostigma plantagineum were used for Agrobacterium mediated transformation of tobacco, but transgenic plants expressing these proteins displayed no phenotypic or growth differences, and no improvement in dessication tolerance (Iturriaga et al. 1992). In another study, cDNA clones of water-deficit genes from the arid woody shrub Atriplex canescens were isolated, and transient expression achieved in transgenic pine tissues (Cairney et al. 1992). The expression of betaine aldehyde dehydrogenase, catalysing the last step in the synthesis of betaine, accumulated during salt or water stress, has been under study in sugar beet (Hanson 1990).
Genes encoding heat shock proteins, which perhaps act as “molecular chaperones” in preventing or repairing cellular damage, have been clones for pea and Arabidopsis (Helm et al. 1992). In this work, antisense transformation of Arabidopsis to reduce the level of HSP21 by 50% resulted in no phenotypic changes. Transformants overexpressing HSP21 showed detrimental effects - greatly reduced size and early flowering, illustrating that manipulation of HSP expression with the goal of increasing thermotolerance is a complex problem.
Specific expression of genes, e.g. CuZn-superoxide dismutase (SOD), in response to exposure to sulphur dioxide, oxides of nitrogen (Karpinski et al. 1992), and ozone (Wegener-Strake et al. 1992) has been under study for some tree species.
Recombinant DNA technology is thus providing a valuable new tool for the study of the molecular control of these and other complex traits, but commercially important manipulation through genetic engineering remains a distant prospect. Existing technology does not permit the ready manipulation of complex pathways (De Waele 1992).
Species for which transformed plants have been produced were listed recently by Birch (1993):
Table 2. Species for which transgenic plants reported (Birch 1993)
|Year||Crops||Vegetables||Ornamental/ Medicinal||Fruit/ Trees||Pastures|
|1986||tomato, cabbage, cucumber||lotus||lucerne|
|1987||sunflower, Brassica spp., cotton, flax/linseed||lettuce, carrot||Arabidopsis||popular||white clover|
|1988||soybean, maize, rice, moth bean||potato, celery, cauliflower, asparagus, eggplant||walnut||Styosanthes, orchard grass|
|1990||kale, mustard||pea, muskmelon, capsicum||geranium, licorice, duboisia||Lisianthus, foxglove||tamarillo, strawberry, grape, citrus, papaya|
|1991||Vicia||chrysanthemum, carnation, rose, datura||pepino, pear, kiwifruit|
|1992||sugarcane, wheat, oats, oilseed rape, safflower||Phaseolus, tomatillo, sweet potato||Dendroboim, belladonna, poppy||craneberry, spruce, apricot||fescue, medic|
At least 400 field trials have been established to 1993, involving genes to improve insect resistance, viral and fungal resistance, weed control, product flavour or other post-harvest qualities. It has been predicted that the first products from transgenic plants are likely to be widely available to consumers in 3–5 years (Birch 1993).
While it is true that tropical crops such as cassava, sweet potato, yam, pulses, banana and plantain have received less attention than others (van Montagu 1992b), tropical crops are nevertheless now the subject of quite active research programmes aimed at genetic engineering of important traits, for example:
Herbicide tolerance, insect resistance and increase of seed nutritional value in crops such as peanut, potato, tomato, grapevine, rice and eggplant in Brazil (Mansur et al. 1992a). Glyphosate resistant potato plants have been obtained.
Virus resistance in rice (rice tungro viruses and rice yellow mottle virus) and cassava (African cassava mosaic geminivirus and cassava common potexvirus) (Beachy et al. 1992).
Insect resistance and improved nutritional value (gene modification to increase essential amino acids) in sweet potato (de la Riva et al. 1992).
Several other projects involve transformation with marker genes (e.g. Gama et al. 1992, Magioli et al. 1992, Mansur et al. 1992b).
Of particular significance is the inclusion of cassava - a crop grown almost entirely by resource-poor farmers in developing countries (Roca & Thro 1992).
Useful or potentially useful genes introduced into forest tree species include:
B.t. genes - inserted into poplar clones (Harry, Strauss & Sederoff 1991, Strauss et al. 1991, Leple et al. 1992).
Proteinase inhibitor genes, e.g. the potato proteinase inhibitor II gene inserted into hybrid poplar (Avila et al. 1992) and into elite clones of Betula pendula (Keinonen-Mettala et al. 1992); and a soybean serine proteinase inhibitor and a rice cysteine proteinase inhibitor inserted into poplar (Leple et al. 1992). In the latter study, transformed plants expressing the proteinase inhibitor genes at a high level are currently being assessed for tolerance to two coleopterous pests, Chrysomela populi and C. tremulae.
Glyphosate resistance gene - inserted into poplar (Whetton & Sederoff 1991).
Antisense petunia chalcone synthase gene - introduced into Juglans with a view to improving rooting through manipulation of phenolics. Transgenic plants are under test (Pastuglia et al. 1992).
Transformation with maker genes, generally GUS or antibiotic resistance, has been reported frequently: e.g. for poplar (Chun 1992, Han et al. 1992), black locust (Han et al. 1992), Eucalyptus grandis (Levee et al. 1992), Liriodendron tulipifera (Wilde et al. 1992), Betula pendula (Keinonen- Mettala et al. 1992), Picea abies (Yibrah et al. 1992), Picea glauca (Charest et al. 1992, Ellis, McCabe et al. 1992, Bommineni et al. 1992), P. mariana (Charest et al. 1992, Bommineni et al. 1992), P. rubrens (Charest et al. 1992) and Larix x eurolepis (Charest et al. 1992). In some cases, only transient expression has been observed.
Transformation with bacterial genes has been reported for Betula pendula (Aronen & Haggman 1992), Populus hybrids (Olsson et al. 1992, Sundberg et al. 1992), Allocasuarina verticillata (Galiana et al. 1992), Casuarina glauca (Galiana et al. 1992), Acacia mangium (Galiana et al. 1992), Acacia albida (Galiana et al. 1992), Pinus radiata (Bergmann & Stomp 1992), Pinus taeda (Huang & Tauer 1992), Pinus sylvestris (Aronen & Haggman (1992), and Larix decidua (Huang 1991)
An important question is the extent to which the insertion of novel or modified genes disrupts other processes. It has been suggested that stress resistance requires energy (Tal 1983), and that fitness costs involved in the use of resistance genes will play an important role in decisions concerning their use (Burdon & Jarosz 1989). Such costs may vary greatly - low for B.t. toxins, effective at low concentrations, but perhaps higher for genes such as proteinase inhibitors, which depend on high levels of expression for effectiveness (Strauss et al. 1991). Glyphosate or phosphinothricin resistant mutants have been reported to display lower fitness than wild-type plants, the mutation in the active centre amino acids conferring resistance to the inhibitor perhaps also resulting in a loss of catalytic properties (Schulz et al. 1990). Few experimental data are available for transgenic plants at this stage. Virus resistance engineered by the coat protein approach has resulted in satisfactory protection under field conditions, without reduction of yield (Hoekema et al. 1989, Harms 1992). At the end of the second growing season, poplar clones transformed with constructs containing the potato proteinase inhibitor gene, on the other hand, displayed significantly lower diameter than untransformed plants of the same clones (McNabb et al. 1991). In some cases, new genes may have unpredictable effects on metabolism - for example expression of the full length B.t. gene was toxic to tobacco cells (Strauss et al. 1991). This discussion underscores the importance of thorough field testing prior to release of transgenic varieties.
A series of generic patents on gene manipulation technologies may develop as a major hurdle to commercial application of transformation - patents are held on Agrobacterium vectors, particle bombardment, the 35S promoter, and antisense technology at least (Birch 1993).
An important selection criterion for many agricultural crops, insect tolerance has been of low priority in breeding programmes with most industrial forest plantation species. Some attention has been given to this trait in poplar programmes, where volume losses (at 5 years) resulting from insect attack have been estimated at 18% (Solomon 1985). Insect susceptibility certainly has been a factor limiting interest in some Meliaceae as commercial species. Conceivably, genetically engineered resistance may be of value also with some species for which a certain level of insect damage is currently tolerated or ignored. Attack by the Nantucket pine tip moth (Rhyacionia frustrana), for example, has been reported to result in significant reduction in volume in loblolly and other southern pines (Hedden & Hangen 1987). Among non-industrial species, insect resistance is important in Leucaena leucocephala, but generation intervals are short, resistance is available in related species, and hybridisation offers a simple means of introducing resistance.
Minimisation of insect damage through genetic resistance, rather than insecticide application, is consistent with the concept of low input, sustainable, environmentally responsible agriculture (Strauss et al. 1991, Harms 1992). Genetic engineering of insect resistance into forest tree species, however, is potentially more challenging than for annual crops, where genotypes can be replaced with new ones as resistance falters. Insertion of an array of different resistance genes, e.g. proteinase inhibitors and B.t. genes, may be required to ensure stability of resistance for the rotation.
The incidence of viral diseases is one of the most important factors limiting productivity in agricultural crops, and development of resistance is therefore a major objective of many breeding programmes. By contrast, no forest tree improvement program includes virus resistance as a selection criterion. Viruses have been held responsible for growth losses of 30–40% in poplars (Cooper 1992), have been reported in some other hardwoods, but are largely unknown in conifers (Strauss et al. 1991).
Fungal diseases cause significant growth losses in many major forest plantation trees, e.g.: poplars, where a 35% reduction of the growing season in clones sensitive to Melampsora rust has been reported (Ontario Tree Imp. and Forest Biomass Institute Forest Res. 1987); slash and loblolly pines, for which a volume reduction of 18% can result from 50% infection (Harrison & Pienaar 1987); and radiata pine. Resistance to fungal disease is a significant selection criterion in breeding programmes for these species, especially the poplars.
Many tree species are quite sensitive to weed competition, and adequate control is essential. Plantation silvicultural systems generally aim at allowing the tree crop to establish control of the site as soon as possible. Commonly this involves some use of herbicides in the early period following planting. Such applications are minimized due to cost and environmental factors. The availability of herbicide resistant plants would permit “over the top”, as opposed to guarded, applications. This is unlikely to be a major economic benefit for species where only one or two applications are involved. Furthermore, environmental legislation in some countries may impede any moves likely to increase the use of herbicides in forested areas. Apart from environmental problems, overuse of herbicides may result in the development of resistance in weed species, as reported for S-triazine herbicides (Hughes 1983). It should be noted, though, that genetic engineering of herbicide tolerance might permit some useful substitution of herbicides - e.g. of the environmentally “friendly” glyphosate for residual herbicides currently applied prior to planting in some programmes. This may be a very important factor in some locations.
Many industrial plantation species are grown as exotics. Due to desirability of silvicultural features, wood properties etc, many are being grown in environments which, climatically, are not matched completely with the natural range. In some cases, adaptation to the new environments has been marred only by poor tolerance to occasional very low temperatures. Examples are the losses of large areas of eucalypts in Florida and southern France during the severe freezes of the mid-1980s. At the provenance level, poor cold tolerance of southern provenance conifers in northern Nordic areas is another example. In the first example, successful use of eucalypts in these areas would require tolerance to temperatures several degrees lower than those to which the species are naturally tolerant. This would be a much larger shift than that reported in preliminary work with cold tolerance genes discussed above.
Several components of wood quality are characterised by high heritability, and wood quality (in particular density) is a selection criterion in many breeding programmes. In many, selection is applied not intensely, but rather in the form of culling of particularly undesirable genotypes. The removal of lignin is an expensive component of pulp production, both economically and in environmental terms, and even modest reductions in the lignin content of wood would be of great value for pulping species such as Eucalyptus grandis. Traditional breeding programmes with these species do not include lignin reduction as an objective, perhaps owing to insufficient variation within species. As noted above though, early results offer hope that significant reductions may be achievable through genetic engineering. Suggestions that 10–15% reductions in lignin content will not result in major structural deficiencies in the plant are supported by work with inhibitors and rubbery wood disease in apple (Whetton & Sederoff 1991, Dean & Eriksson 1992). Nevertheless, proper field testing of structural properties, disease and insect resistance of lignin-reduced trees will be essential prior to deployment. The reduction of lignin is not a useful goal in plantations grown to provide wood for structural purposes or for fuel.
Male sterility is used in many crop species, mainly for the commercial production of hybrids. While male sterility would be of value for the production of some hybrids with forest tree species, other viable approaches to hybrid production are generally available, and this would be a minor applications. More important potential benefits of sterility in forest tree species are:
Increased vegetative growth, resulting from redirection of resources otherwise committed to reproductive development. A review by Strauss (1992, unpublished) suggests potential gains of at least 16% in radiata pine and Douglas fir, and perhaps even higher for some other species.
Prevention of the escape of genes into wild populations. Many plantation programmes involve species which are not exotics, and strategies to prevent the release of novel genes engineered into plantation trees will be important. Such release, for example, might accelerate counter evolution by insects to engineered resistance genes (Strauss et al. 1991).
While effects of sterility on vegetative growth may be complex and difficult to predict, there is no doubt that sterility will greatly facilitate major applications of genetic engineering. In both cases, both male and female sterility will be desirable.
For agricultural crops, the insertion of new genes into established, proven genotypes has been the approach to integration of genetic engineering into traditional improvement programmes most commonly applied or envisaged. For forest tree species, alternatives which can be considered are:
The engineering of new genes into commercially proven clones. These new genotypes are then field tested for stable integration and absence of any pleiotropic effects, and then go into the commercial propagation program. This is essentially the agricultural crop approach. The breeding program underlying a clonal forestry operation such as this could comprise:
In a system such as this, the clonal testing program draws on the best products of the recurrent crossing program, but commercial clones do not re-enter the crossing program (or at least they do not need to). This would be compatible with the requirement for sterility in engineered genotypes. There would be no requirement to re-engineer clones to restore fertility for continued use in breeding. Nevertheless, there are some disadvantages associated with the approach:
Some of the tested superior clones will not be amenable to the transformation and regeneration procedures. There is thus a selection penalty.
There are two consecutive testing phases before clones are available commercially - the clonal test prior to transformation, and then testing to confirm appropriate expression in the transgenic plants. There is thus also a time penalty involved.
The engineering of new genes into juvenile clones prior to testing. This is similar to the above, except that the novel genes are introduced to clones prior to testing. Juvenile material, e.g. immature embryos, from the families selected as donors of genotypes for clonal testing is subjected to the transformation procedures. A marker could be used to select for transgenic plants. The transgenic plants are then established in a clonal test, which serves to identify genotypes which are desirable with respect to both traditional traits and the expression of the novel genes. This avoids the disadvantages above - resources are not wasted on clonal testing of genotypes not amenable to transformation and regeneration, and clonal testing and testing for expression of novel traits are conducted concurrently. An additional advantage is that transformation and regeneration are conducted with juvenile tissues, usually much more amenable. A larger number of genotypes must be transformed - but this need not be a problem. It is essential though that the absence of adverse genetic correlations between major commercial traits and competence for transformation and regeneration be confirmed.
Transformation of known good parents, and then use of these in seed orchards. Involving the simple addition of a genetic engineering step to traditional seed orchard technology, this approach is superficially attractive. The same disadvantages outlined for option 1 above apply - some selections will not be amenable to the procedures, and there will be a time penalty for confirmation of appropriate expression. Other disadvantages will be:
The approach is not compatible with the principle of prevention of escape of genes through use of sterility. This is likely to constitute a major obstacle to the integration of genetic engineering into seed orchard programmes.
Not all seed orchard progeny will carry the novel genes, even in the absence of pollen contamination and for completely dominant genes.
Transformation of microspores. Microspores of Brassica napus have been transformed by electroporation (Jardinaud 1992). For tobacco, an in vitro system that allows effective maturation of microspores into pollen grains that can be used to pollinate emasculated flowers in situ has been developed (Alwen et al. 1990). Nevertheless, many technical difficulties remain to be overcome with this approach. Transgenic pollen grains could be used in either of two ways:
In the breeding program. Fertility of progeny would be a requirement, leading to the obstacle discussed above regarding escape of genes into wild populations.
For pollinations to yield genotypes for clonal testing (forward selected families as described above). Fertility of the progeny would not be required, and sterility genes could be included in the constructs.
At this stage, it seems likely that the clonal test will be the most appropriate basis for the integration of genetic engineering into breeding programmes. Transformation is probably most efficiently incorporated prior to clonal testing, and transformation of mature material is therefore not essential.
Work with the genetic engineering of forest tree species is advancing rapidly, and there are likely to be many more reports of transformation of forest tree species with marker genes and the simple genes such as B.t. and glyphosate resistance over next couple of years.
The availability of effective transformation techniques remains an obstacle, but improved techniques are being developed. Regeneration is a difficulty for some tree species, but the problem may be over - rated - the non-competence of mature material is not necessarily an obstacle to effective application of genetic engineering, provided that juvenile material responds satisfactorily. Major traits for which genetic engineering can most realistically be contemplated in the near future include virus resistance, insect resistance and herbicide tolerance. Even so, insertion of one of these genes into a new species would be a substantial undertaking, and insertion of enough genes to confer long term insect resistance in a perennial species (particularly for long rotations) more so. Virus and insect resistance, in particular, are of major significance for crop plants. By contrast, these traits are not among the most important for most forest tree species. Reduction of lignin is a valuable objective for pulp species, and prospects look good. Cold tolerance is a trait of considerable interest, particularly in some eucalypt species. Much remains to be done, though, to establish that sufficient tolerance can be conferred using antifreeze proteins, and to extend the work to tree species. Prevention of the escape of genes into wild populations is likely to become an important concern, and sterility should be an early target of genetic engineering work with forest tree species. The major factor limiting application of genetic engineering in forest tree species is the state of knowledge of molecular control of the traits which are of most interest - those relating to growth, adaptation and stem and wood quality. Genetic engineering of these traits remains a distant prospect. An often overlooked research component is the testing which would be required before a responsible recommendation for large scale deployment of transgenic plants could be made. Such testing could be extensive and prolonged, depending on the species and genes involved.
It is important that genetically engineered genotypes be of high quality with respect to other traits as well. The clonal test is the most logical basis for integration of genetic engineering into traditional tree improvement programmes. For these reasons, genetic engineering is most appropriately conducted with species where breeding programmes are advanced and clonal forestry can be realistically contemplated. Genetic engineering represents, for many crop plants, the best hope for addressing the major priorities of breeding programmes - the acquisition of virus and insect resistance. This applies also for some crops in developing countries, e.g. cassava. By contrast, genetic engineering can do little, at the present time, to address the major priorities of breeding programmes for non-industrial tree species in developing countries.