1Department of Plant and Microbial Biology, University of California, Berkeley CA 94720
2USDA-ARS Natl Small Grains Germplasm Research Center, P.O. Box 307, Aberdeen, ID 83210
Introduction
In the context of classical breeding, the application of genetic engineering methods for commercially important barley cultivars could play an increasingly important role in solving some fundamental challenges. This could include elevating yields by improving agronomic traits, such as enhancing pest, stress and herbicide resistance. Improvements could also be made in the quality of the crop, including its food and feed characteristics. Genetic engineering methods can also be used to create in barley and wheat new value-added products, which are presently being made from other raw materials, e.g. industrial enzymes, plastics, neutraceuticals. The advent of molecular technologies provides an important adjunct to classical breeding that widens the genepool available for breeders to accomplish these tasks.
The first publication describing transgenic plants expressing engineered foreign genes occurred in 1984 (for review, Gasser and Fraley 1989). Despite the significance of this achievement, significant problems arose in the practical application of biotechnological methods to the study of crop species and to their manipulation for agronomic improvement, particularly in cereals. Over time and with considerable effort, some challenges have been ameliorated; however, significant inefficiencies remain in being able to introduce and stably express just the gene(s) of interest into commercially important cultivars that perform like the parent. These include:
- Lack of reproducible, efficient transformation systems for commercial cultivars
- Minimization of somatic mutation and stable epigenetic changes during transformation
- Removal of vector sequences and selectable/screenable genes
- Transgene and transgene expression instability
These deficiencies substantially increase the time needed to achieve successful transformation and impede the application of these methods to practical germplasm enhancement programs. Therefore much work in our laboratories has focused on minimizing these problems to facilitate efficient and effective new cultivar development. Although the work has focused on barley and wheat, many of the principles and strategies will have broad application for other cereals and perhaps dicotyledonous crops
Results and Discussion
Lack of Reproducible, Efficient Transformation Systems for Commercial Germplasm
The limitations of sexually compatible germplasm resources for barley and wheat emphasize the potential importance of applying molecular technologies to the task of improvement. Knowledge of the physiology underlying particular characteristics and of the genes involved in the expression of these pathways will enable geneticists to identify new sources of allelic variability. The existence of an efficient, non-mutagenic technology for genetic transformation is a pre-requisite if such variability is to be readily accessible to breeders for cultivar development.
The utility of genetic engineering approaches is dependent upon being able to generate large numbers of independently transformed, fertile, green plants that maintain the important characteristics of the starting germplasm. The first demonstrations of stable transformation of barley that resulted in fertile, stably transformed barley plants was in 1994 (Wan and Lemaux 1994). These studies utilized a two-rowed malting cultivar Golden Promise that was amenable to in vitro culture but is currently not commercially important; the DNA introduction method was microparticle bombardment. Primary explants that required little or no preculturing prior to bombardment were utilized, i.e.immature embryos (IEs) and 7-9-day-old callus. These explants were utilized because of the loss of regenerability and increased albinism in in vitro cultured tissue that is maintained over long periods of time.
Using a standard embryogenic callus induction method, Wan and Lemaux (1994) developed a reproducible transformation method with Golden Promise that gave an "effective transformation frequency" (# fertile, regenerable independent lines/total # explants) of approximately 4%. The direct use of this method with commercially important cultivars was problematic. Frequencies of callus induction were often low; target tissues negatively impacted by bombardment; and regenerability of in vitro-cultured tissue lost quickly during selection. The barley germplasm that forms the foundation for North American breeding programs, many in six-rowed backgrounds, is different from Golden Promise in virtually every aspect important to commercial production and utilization. Therefore, utilization of Golden Promise as a source of transgene-derived traits will require multiple cycles of backcrossing and selection to identify plants that are identical to the commercials variety with the exception of the new value-added trait since the selected allelic combinations that are critical to commercial success of these varieties would have to be preserved.
Application of the published transformation method (Wan and Lemaux 1994) to commercially important North American varieties, e.g. Galena, Harrington, Moravian III and Morex, resulted in the identification of independently transformed lines, which were either nonregenerable or yielded only albino plants (Wan, Y., Cho, M.-J. and Jiang, W. unpublished results). Changing the level of selection agent or shortening the time of selection led to the regeneration of green plants, but none were transformed (Wan, Y. and Lemaux, P.G., unpublished results).
One insight into a potential mitigator of genotype-dependence came from a report in which the effects of DNA bombardment parameters were examined relative to their effects on the callus-induction frequency of certain target tissues. Koprek et al.(1996), utilizing six two-rowed spring cultivars, studied callus response of IEs following bombardment with both a Bio-Rad PDS-1000 and a particle inflow gun (Finer et al. 1992). The callus response of cultivars, other than Golden Promise, was found to be severely reduced by bombardment at pressures around 1100 psi with the Bio-Rad machine. Using the particle inflow gun and less destructive conditions, stably transformed plants were recovered from the previously recalcitrant cultivars, Corniche, Salome and Femina, despite the fact that numbers of transiently expressing GUS foci were very low following use of the particle inflow gun. Scanning electron microscopic analysis of scutella showed that bombardment with the Bio-Rad gun caused extensive damage to the IEs of Salome, a recalcitrant variety (Figure 1A), while the particle inflow gun caused much less physical impact.
In the development of a transformation technology for the two-rowed commercial variety, Galena, damage to IEs of microparticles of this variety was found to be significantly more than to IEs of Golden Promise, as judged by SEM. These results suggest that IEs from different cultivars have differing capacities to withstand microparticle bombardment and this parameter must be optimized to achieve success. An optimization protocol for the bombardment of Galena, using the generation of embryogenic callus as a guide, bombardment of IEs from Galena at 600 psi at a distance of 11 cm was optimal (Figure 1B, T. Koprek, unpublished results).
Figure 1A. Scanning Electron Micrograph of Salome Immature Embryo Bombarded at 1100 psi
Figure 1B. Bar Graph of Effects of Bombardment on Immature Embryos of Galena
Another hurdle to successful transformation of commercial varieties had to do with the numbers of totipotent cells found in the cultures. In general, callus-maintenance medium containing auxin and no cytokinin was used for the long-term tissue culture periods needed during transformation. In barley this led to the generation of tissues with low regenerability. The use of medium, containing an auxin, 2,4-D, and a low level of a cytokinin, BAP, and higher levels of copper for callus initiation, stimulated the formation and maintenance of green, regenerative tissues with long-term regenerability characteristics (Cho et al., 1998). The use of this method of initiating and maintaining callus caused the small number of green, totipotent cells in the scutellum to proliferate, and, by transferring these tissues to a medium with increasing cytokinin, it was possible to convert the tissue from an embryogenic state to a state more closely resembling a shoot meristem culture (Cho et al., 1998; Lemaux et al., 1999). This tissue, termed green regenerative tissue, could give rise to multiple shoots over long periods. The judicious use of media with differing auxin to cytokinin ratios improved callus quality and significantly enhanced duration of regenerability of tissues of the barley varieties, Galena (Jiang et al. 1998), Harrington, and Morex (Cho and Jiang, unpublished results) and the wheat varieties, Yecora Rojo, Karl and Anza (Cho and Kim, unpublished results). Levels of cupric sulfate in the range of 5-1000 times the level in the culture media was also found to have positive effects on the regenerability of the wheat and barley cultures.
The transformation of recalcitrant varieties using green regenerative tissue has certain advantages over IEs. First, the use of this tissue does not require a constant source of IEs from plants grown under controlled growth conditions. Second, green regenerative tissues do not require special care during bombardment. Third, these tissues incur negligible losses of regenerability compared to the IE-derived callus during the 2 to 3 months needed for selection of transformed tissue. Fourth, albinism problems rarely occur except following certain selection conditions. Fifth, it is possible that plants deriving from green tissues incur less somaclonal variation than those deriving from culture under auxin alone (Zhang et al., 1999).
Potential for somatic mutation and heritable changes introduced during in vitro culture.
Plants derived from in vitro culture frequently accumulate heritable genetic changes, termed somaclonal variation (SCV). In barley these can manifest themselves as moderate to severe negative alterations in critical agronomic and biochemical characteristics. The elements of the in vitro environment and the transformation process, which induce these changes, are poorly understood, but are likely related to the establishment of in vitro-cultured tissue, introduction of DNA and subsequent selection of transformed tissue. SCV likely will hinder the success of transformation efforts causing reductions in the regeneration of fertile, green plants. The relationship of SCV to this gradual loss of totipotency is likely caused by the accumulation of genomic alterations in cultured cells (for review, Kaeppler and Phillips 1993) which interfere with the physiological processes necessary for redifferentiation and viability. SCV can also reduce the fidelity of the resulting plants, interfering with their use as parents in a breeding program. Bregitzer and Poulson (1996) noted decreased field performance in plants from six genotypes regenerated after short periods of in vitro culture. For example, yield of Golden Promise plants regenerated from in vitro culture was reduced to 91% of control plants on average. Although backcrossing can often minimize the latter problem, this requires considerable additional time and effort and would be problematic if the induced mutations were closely linked to the transgene.
During the first few weeks after the establishment of in vitro cultured barley tissue, regenerability typically declines drastically and widespread genomic alterations can be observed, either directly in cultured cells or in regenerated plants. Reports include observations of heritable alterations in the methylation patterns of microspore-derived plants (Devaux et al. 1993) and alterations in the methylation patterns of plants derived from different in vitro culturing methods (Zhang et al., 1999). The alterations in methylation patterns observed by Zhang et al. are positively correlated with time in culture, negatively correlated with regeneration potential of the cultures and negatively correlated with agronomic performance in the field.
The field performance of transgenic Golden Promise plants are even more dramatically impacted than those having resulted from tissue culture alone. In comparison to the 9% yield loss of the Golden Promise plants from in vitro culture alone, the transgenic plants on average gave only 50% the yield of control plants in the fourth generation, indicating that the transformation process caused more severe impact than the tissue culture process alone. Consistent with this are the observations of Choi et al. (1999) in which the chromosomal state of transgenic plants was determined. Whereas no barley plants taken through tissue culture alone on the same medium and for the same period of time had a change in ploidy or aneuploidy, 46% of the transgenic lines gave rise to aneuploid or tetraploid plants (Table 1).
Table 1. Ploidy Variation in Transgenic and Nontransgenic Barley Plants
Culture |
Chromosome number |
# plants |
% abnormal ploidy | ||
medium |
14 |
aneuploid/28 |
analyzed |
(# abnormal/total) | |
Nontransgenic |
DM |
6 |
0 |
6 |
0% |
D |
22 |
1 (tetraploid) |
23 |
4.3% | |
DC+ |
34 |
0 |
34 |
0% | |
DBC1 |
10 |
0 |
10 |
0% | |
DBC2 |
19 |
0 |
19 |
0% | |
Transgenic |
DC |
32 |
27 |
59 |
46% |
Three-month-old calli from each callus-induction medium except for DC medium were transferred onto FHG regeneration medium (Cho et al., 1997).
+An intermediate culturing step on DBC2 between the callus induction (DC) and regeneration (FHG) step was applied for 1 month.
This is another indication that the transformation procedure per se causes additional stresses that result in impact on genomic integrity, reducing the performance of transgenic plants. Efforts are currently underway to dissect the transformation process in order to identify and ameliorate the major factors causing genomic instability and mutation.
Removal of vector and selectable marker sequences
The use of a maize transposable element system has been pursued in barley and wheat (McElroy et al., 1997; Koprek, unpublished results). This system utilizes an immobilized, trans-active Activator (Ac) transposase gene linked to a negative selectable marker codA and an unlinked cis-responsive Dissosiation (Ds) element carrying the selectable marker gene, bar, as an insert. In vitro transient assays demonstrated that frequent excisions of Ds occur when the non-autonomous Ds element was transiently introduced into stably transformed Ac transposase-expressing plant tissue (McElroy et al., 1997). Stably transformed Ac-positive plants expressing transposase either under transcriptional control of the putative Ac promoter or the promoter and first intron from the maize ubiquitin gene were crossed with Ds-bar-containing plants. Low somatic and germinal transposition frequencies were observed in F1 plants (Koprek, unpublished results); higher frequencies were observed in F2 progeny, derived from individual selfed F1 plants. Further analyses of F2 plants showed that both linked and unlinked resinsertions occurred. Transposed Ds elements of plants, lacking the Ac transposase gene, could be reactivated by reintroducing the transposase gene transposase through crossing, making this transposon system useful for gene delivery.
One possible mechanism for separating the transgene(s) of choice from selectable marker genes and vector sequences utilizes transposable elements like the maize system, Ac/Ds. In this strategy the gene of interest is placed inside the recognition sites for the transposition enzyme; the selection gene remains outside the recognition sites. In the presence of the transposase, the gene of interest can be moved away from the remainder of the introduced DNA allowing the desired transgene to be physically separated from the remainder of the introduced DNA during subsequent generations.
The availability of this system also provides a mechanism to generate multiple insertions of the transgene into unlinked locations. Only two plants must be generated in order to generate hundreds of transformants with the transgene of interest scattered throughout the genome (Figure 2). It is possible to generate small numbers of transformants, even with the most recalcitrant of cultivars. If one transformant carries the Ds recognition sites along with the transgene of interest and the other transformant carries and expresses the transposase gene product, it will be possible to cross the two plants to activate the Ds-transgene element to move around the genome.
Figure 2. Schematic of Creation of Multiple Independent Insertions from Two Transgenic Plants
Transgene and transgene expression instability.
One limitation of current methods of transformation is that insertions occur randomly and the location of the insertion might not be optimal for gene expression. This has been noted in plants derived from direct DNA introduction methods, but also occurs with Agrobacterium-mediated methods despite the stated preference for insertion into transcriptionally active regions (Czernilofsky et al. 1986; Koncz et al. 1989). Another problem, more common with direct DNA methods, is that the introduced DNA is often present in multiple, tandemly arrayed copies. Having multiple copies of the introduced genes, closely linked in the genome, can lead both to gene inactivation and genetic instability (Flavell 1994; Matzke and Matzke 1995). Present methods also lead to transformed events that retain the selectable marker gene and vector sequences in the plants that express the desired value-added gene.
Stable physical transmission and expression of transgenes are critical for efficient application of transformation technologies to plant breeding. However, as currently practiced, the stability of the transgene itself or its expression is often unstable (for review, Finnegan and McElroy 1994). This instability in the transgene or its expression has been well-documented in barley. In the original report of stable transformation, 69% (24/35) of the T0 plants, representing 21 independent lines, did not show Mendelian inheritance of transgene expression in T1 progeny (Wan and Lemaux 1994). It is likely that the variation in stability among different transgenic events is related at least in part to the site of transgene integration. The influence of the site could be mediated by the fact that different genomic methylation patterns are present in embryogenic versus mature plant tissue, leading to a situation where a transgene, integrated into a site not methylated in the embryo, is subsequently methylated in a mature tissue. Another mediating factor could be the fact that demethylation occurs during in vitro culture (Phillips et al. 1994). This could lead to a situation in which a transgene integrates into a site that is demethylated in the T0 plant but becomes remethylated in subsequent generations. The propensity for and speed with which any one site will undergo demethylation and remethylation could be different in different transgenic lines, leading to variation in the propensity and speed with which an integrated transgene undergoes silencing due to methylation.
It has been previously noted that insertions of Agrobacterium sequences occur frequently in transcriptionally active regions (Koncz et al. 1989); this has also been suggested for the insertion of Ac and Ds elements (Doring et al. 1989). Therefore, it was of interest to determine whether this propensity of Ds elements would have an effect on transgene expression stability in barley. Preliminary results indicate that it does. Comparisons were made among the progeny of several F2 plants, one group of which had a single copy of Ds-bar that had been introduced via direct DNA introduction methods. The other group of plants also had a single copy of Ds-bar, but these copies resulted from the movement of the element via transposase. Comparisons of the stability of expression of bar in the two sets of plants showed a dramatic improvement in transgene expression stability. None of the plants (0%) that resulted from transposition showed transgene expression silencing, while 25% of the nontransposed group showed evidence of silencing in the subsequent generation (Table 2).
Table 2. bar Gene Expression Instability in Transposed and Nontransposed Loci
Efforts are currently underway to confirm this effect in multiple sets of plants and to determine the nature of the stabilized insertion sites. Such a finding will prove to be a significant advantage to cereal transformation efforts for a number of reasons.
References
1. Bregitzer P and Poulson M (1995) Agronomic performance of barley lines derived from tissue culture. Crop Sci 35: 1144-48.
2. Bregitzer P, Halbert SE and Lemaux PG (1998) SCV in the progeny of transgenic barley. Theor. Appl. Genet.96:421-425.
3. Cho M-J, Jiang W and Lemaux PG (1998) Transformation of recalcitrant genotypes: improvement in regenerability and decreased albinism, Plant Science 138:229-244.
4. Choi HW, Lemaux PG and Cho M-J (1999) Increased Cytological Instability in Transgenic versus Nontransgenic Barley (Hordeum vulgare L.) Plants. Crop Science, accepted.
5. Czernilofsky AP, Hain R, Herrera-Estrella L, Goyvaerts E, Baker BJ and Schell J (1986) Fate of selectable marker DNA integrated into the genome of Nicotiana tabacum. DNA 5: 101-113.
6. Devaux P, Kilian A and Kleinhofs A (1993) Anther culture and Hordeum bulbosum-derived barley doubled haploids: mutations and methylation. Mol Gen Genet 241: 674-79.
7. Doring HP (1989) Tagging genes with maize transposable elements. An overview. Maydica 34:73-88.
8. Finer JJ, Vain P, Jones MW and McMullen MD (1992) Development of the particle inflow gun for DNA delivery to plant cells. Plant Cell Rep 11: 323-328.
9. Finnegan J and McElroy D (1994) Transgene inactivation: plants fight back! Bio/Technology 12: 883-888.
10. Flavell RB (1994) Inactivation of gene expression in plants as a consequence of specific sequence duplication. Proc Natl Acad Sci USA 91: 3490-3496.
11. Gasser CS and Fraley RT (1989) Genetically engineering plants for crop imrpovement. Science 244:1293-1299.
12. Jiang W, Cho M-J and Lemaux PG (1998) Improved callus quality and prolonged regenerability in model and recalcitrant barley (Hordeum vulgare L.) genotypes. Plant Biotechnology, 15:63-69.
13. Kaeppler SM and Phillips RL (1993) Tissue culture-induced DNA methylation variation in maize. Proc Natl Acad Sci USA 90: 8773-8776.
14. Koncz C, Martini N, Mayerhofer R, Koncz-Kalman Z, Körber H, Redei GP and Schell J (1989) High-frequency T-DNA-mediated gene tagging in plants. Proc Natl Acad Sci USA 86: 8467-8471.
15. Koprek T, Haensch R, Nerlich A, Mendel RR and Schulze J (1996) Fertile transgenic barley of different cultivars obtained by adjustment of bombardment conditions to tissue response. Plant Sci 119: 79-91.
16. Lemaux PG, Cho M-J, Zhang S and Bregitzer P (1999) Transgenic Cereals: Hordeum vulgare L. (barley). In: Molecular Improvement of Cereal Crops (IK Vasil ed.) Kluwer Academic Publ, UK, pp 255-316.
17. Matzke MA and Matzke AJM (1995) How and why do plants inactivate homologous (trans)genes? Plant Physiol 107: 679-85.
18. McElroy D, Louwerse JD, McElroy SM and Lemaux PG (1997) Development of a simple transient assay for Ac-Ds activity in cells of intact barley tissue. Plant J 11: 157-165.
19. Zhang S, Zhang S, Cho M-J, Bregitzer P and Lemaux PG (1999) Comparative analysis of genomic DNA methylation status and field performance of plants derived from embryogenic calli and shoot meristematic cultures. In: Plant Biotechnology and In Vitro Biology in the 21st Century (A. Altman ed.) Kluwer Academic Publishers, The Netherlands, pp. 263-267.
