Crop Science 43:4-12 (2003)
© 2003 Crop Science Society of America
CROP BREEDING, GENETICS & CYTOLOGY
Inheritance and Expression of Transgenes in Barley
Phil Bregitzer*,a,b and
Dennis Tonksa,b
a United States Dep. of Agriculture, Agricultural Research Service, P.O. Box 307, Aberdeen, ID 83210, USA
b Univ. of Idaho Research and Extension Center, P.O. Box AA, Aberdeen, ID 83210, USA
* Corresponding author (pbregit{at}uidaho.edu)
 |
ABSTRACT
|
|---|
Empirical assessments of transgene inheritance and phenotypic expression will assist in the development of efficient breeding strategies for transgenic germplasm, and guide research into the improvement of transformation techniques. The inheritance of a barley yellow dwarf virus (BYDV) coat protein gene and bar, and the expression of bar as measured by resistance to glufosinate-ammonium damage, was studied in the T1 and T3 generations of barley (Hordeum vulgare L.) populations derived from seven independent transformation events. Most populations deviated from Mendelian inheritance patterns, and several showed evidence of transgene silencing. To further study transgene behavior, several transgenic lines were crossed to a diverse set of nontransgenic cultivars and breeding lines to produce single cross- and backcross-derived populations. In these populations, the inheritance of glufosinate-ammonium resistance generally fit Mendelian expectations for single, dominant loci. Quantitative measurements of glufosinate-ammonium resistance showed heritable variability for glufosinate-ammonium resistance both among and within individual transformation events, but no variability could be attributed to the different genetic backgrounds of the nontransgenic parents. It is concluded that, although transgenic parents such as these can be used in a breeding program, transformation systems that result in greater stability of transgene behavior are desirable.
Abbreviations: bp, base pair • BYDV, barley yellow dwarf virus • CP, coat protein • PAT, phosphinothricin acetyltransferase • PCR, polymerase chain reaction • R:S, resistant:susceptible
|
INTRODUCTION
|
|---|
TRANSGENIC LOCI introduced into higher plant species frequently display unpredictable patterns of inheritance and expression. Transgene instabilities—either of inheritance or expression—complicate the identification, selection, and use of transgenic lines. The development of efficient breeding strategies, and of improved transformation systems, should be based on knowledge acquired from empirical assessments of the inheritance and expression patterns of transgenic loci.
Studies of transgene silencing and inheritance have shown that unpredictable behavior is common. Observations include irregular patterns of silencing and reactivation, and variability of transgene expression among "clonal" progeny of a given transformation event (Müller et al., 1996; Zhang et al., 1996; Kumpatla and Hall, 1998; Pawlowski et al., 1998). Variable transgene expression among clonal cell lines of mice (Mus musculus) has been reported and compared with the position effect variegation phenomenon described for yeast (Saccharomyces cerevisiae) and Drosophila (Kioussis and Festenstein, 1997); it is reasonable to assume that a similar phenomenon could occur in transgenic plants and contribute to irregular transgene expression. Inheritance patterns of transgenes also exhibit irregular behavior, and frequently deviate from Mendelian expectations. For example, Srivastava et al. (1996) observed the loss of transgenes in R3 wheat plants following loss of expression in prior generations, and poor transmission of integrated transgenes to progeny in some transgenic lines has been observed also in maize (Zea mays L.) (Spencer et al., 1992; Register et al., 1994) and barley (Cho et al., 1999).
We are interested in developing breeding methodologies that will result in the efficient use of transgenic barley plants as part of a routine breeding program. To properly develop routine protocols for identifying useful transgenic parents, and properly handle recombinant populations, we are seeking answers to several basic questions: What are the basic characteristics of transgenic barley plants; in particular, how do they perform agronomically and with respect to stability of transgene expression and inheritance? A previous study examined the agronomic performance of the progeny of transgenic barley (Bregitzer et al., 1998), and for this report we examined the inheritance of a BYDV coat protein gene and bar, and the expression of bar as measured by resistance to glufosinate-ammonium damage, in the progeny of seven independent transformation events through the T3 generation. From these populations, several plants from two independent transformation events were selected and used as parents to cross to a variety of elite barley germplasm; single cross- and backcross-derived progeny populations were assayed for their inheritance and expression patterns based on the expression of resistance to glufosinate-ammonium damage. The implications of the results of these assays for the analysis, selection, and use of transgenic parents in a breeding program are discussed.
|
MATERIALS AND METHODS
|
|---|
General
Transgenic barley lines were derived from seven independent transformation events which were produced and described by Wan and Lemaux (1994). All lines were derived from phosphinothricin acetyltransferase (PAT)-positive callus lines which were shown by Southern hybridizations to have incorporated bar and pRsa1BYDVcp (described below) into high molecular weight genomic DNA. The particular T0 plants from which the materials in this study were derived were shown to be transgenic based on Southern hybridization, PAT assays, and/or by phenotypic assays for glufosinate-ammonium resistance (data not shown).
This report refers to transgenic families; each family was comprised of progeny from a single transformation event. Each family was further subdivided into one to four T0–derived lines, each comprised of all progeny of a particular T0 plant. In several cases, additional within-line selections were made, resulting in several T3– or T4–derived sublines for these selected lines, as detailed below.
Six families contained the coat protein gene from a PAV (Rochow, 1979) isolate of the BYDV driven by the cauliflower mosaic virus (CaMV) 35S promoter, and the bar and uidA genes driven by the maize ubiquitin promoter and first intron encoding, respectively, resistance to the herbicide glufosinate-ammonium and ß-glucuronidase activity. An additional family, GP717B-9, differed in that it did not contain the coat protein gene. The agronomic performance of families derived from each of the transformation events has been described (Bregitzer et al., 1998).
Inheritance and Expression of Transgenes from the T1 to the T3
Each T1 line used in these studies was produced by self-pollination of individual T0 plants, as described in Wan and Lemaux (1994). For each transgenic line, 18-30 T1 plants were grown in the greenhouse during the winter season of 1993–1994 and harvested in bulk to produce T2 seed that was used for field studies in 1994; T3 populations were produced by bulking equal numbers of seeds from each of 30–40 T2 plants grown under irrigation at Aberdeen and Burley, ID (Bregitzer et al., 1998).
Progeny from each line were grown in Speedling 128-cell flats (Speedling, Inc., Sun City, FL, USA) in a potting mix of sand:peat:vermiculite (3:2:2) supplemented with commercially available slow-release fertilizer. All plants were grown in greenhouses with natural light. Temperatures ranged from 
10 to 25°C. The expression of bar was assayed by applying an herbicide solution [0.5 g kg-1 (0.05%) a.i. glufosinate-ammonium; ammonium-DL-homoalanin-4-ylmethylphosphinate; Riedel-deHaën, Germany] to sections of two leaves of seedlings at the three- to five-leaf stage, and visually assessing herbicide damage (resistant or susceptible). Each line was characterized by scoring progeny in two separate tests. Resistant plants were selected at random for generation advance, and transplanted to larger pots containing the same potting mix as described above. Several of these selected plants were subsequently used as parents for various crosses (described below).
Polymerase chain reaction (PCR) assays were used to detect the coat protein (CP) and bar genes in selected T1 and T3 plants. Seedling leaf tissue samples were collected from all progeny grown for the second test of herbicide resistance, from a negative control (nontransgenic ‘Golden Promise’), and from positive controls (two different transgenic Golden Promise lines known to be homozygous for both CP and bar). DNA was extracted by the following steps: small (50-mg) samples of seedling tissues were ground in 1.5 µL microcentrifuge tubes suspended in liquid N; on ice, 300 µL of a CTAB-based extraction buffer was added to each sample and mixed vigorously with the ground tissue, followed by the addition of 300 µL of chloroform, and incubation at 65°C for 30 to 45 min; at room temperature, 225 µL of the aqueous phase was removed to a clean microcentrifuge tube and DNA was precipitated with an equal volume of isopropanol, pelleted, and rinsed with 100 µL of 70% ethanol; and finally, pellets were dried and redissolved in 30 µL of 10:0.01 Tris-EDTA. Two microliters of each DNA sample, plus a 200-ng 
-HindIII standard, was loaded onto a 1% agarose gel and electrophoresed for 3 h at 80 V, stained with ethidium bromide, and examined under UV light to determine relative DNA concentration and quality. Samples in which DNA appeared to be excessively sheared or extremely dilute were discarded, and reextracted from remnant plant tissues stored at -20°C. All DNA samples found to be of suitable quality were further quantified via Hoechst 33258-dye based fluorometric measurements (using calf thymus DNA at 100 ng µL-1 as a concentration standard), and subsequently diluted to 10 ng µL-1 with 10:0.01 Tris-EDTA.
Polymerase chain reaction assays were conducted using primers specific to sequences within the coding region of each gene (for CP: CP-F, 5'-AGAACAAGTTCGGCCAGTGGT-3' and CP-R, 5'-CCATTAATCGCTTCGGCTCT-3'; for bar: BAR5F, 5'-CATCGAGACAAGCACGGTCAACTTC-3' and BAR1R, 5'-ATATCCGAGCGCCTCGTGCATGCG-3'). Thermal cycler conditions were for detection of CP were: one cycle, 94°C 1 min; 35 cycles, 94°C 30 s–55°C 15 s–72°C 30 s. Thermal cycler conditions for detection of bar were: seven cycles, 95°C 1 min–65°C 30 s–72°C 2 min; 16 cycles, 95°C 1 min–60°C 30 s–72°C 2 min; 16 cycles, 95°C 1 min–55°C 30 s–72°C 2 min; 1 cycle, 72°C 7 min. Twenty-five µL reactions were set up according to directions supplied by the manufacturer's of the DNA polymerase (Amplitaq DNA polymerase was used for CP reactions, supplied by Perkin-Elmer, Branchburg, NJ, USA; DyNAzyme EXT DNA polymerase was used for bar reactions, supplied by MJ Research, Inc., Waltham, MA, USA). Approximately 25-ng template DNA was used for each CP assay, and 100-ng template DNA was used for each bar assay. Twenty µL of each reaction was loaded onto a 2% agarose gel and electrophoresed for 3 to 4 h at 80 V. For each CP and bar assay, each DNA sample was assayed at least twice, and data was recorded only for assays in which amplified products were of the expected size [CP = 416 base pair (bp); bar = 343 bp], the positive controls produced products of the expected size, and the negative control produced no product. In cases where PCR assay results were inconclusive, DNA was reextracted from remnant tissues and additional assays were conducted as necessary to produce interpretable results.
Inheritance and Expression of Transgenes in the Progeny Derived from Crosses and Backcrosses
Two transgenic families which had relatively good transmission of transgenes through the T3 generation, and which showed relatively good stability of bar expression (measured as resistance to 0.5 g kg-1 glufosinate-ammonium) were selected for further study. Several plants homozygous for bar (based on progeny tests for glufosinate-ammonium resistance and PCR assays) were selected at random from T3 progeny of GP724B-4-9 (designated T3 #30, 31, 32, and 37) and T4 progeny of GP717B-59-11 (T4 #19). These plants were self-pollinated to produce T3– and T4–derived sublines, and were used as parents in crosses to nontransgenic breeding lines and cultivars. F1 progeny of these crosses were then backcrossed to these same breeding lines and cultivars. The recurrent (nontransgenic) parent was used as the female for most crosses and for all backcrosses.
Progeny derived from these crosses were assayed in the F2, F3, BC1F1, and BC1F2 generations for glufosinate-ammonium damage based on visual observations; scoring was qualitative (either resistant or susceptible). For the development of backcross populations, putatively nontransgenic BC1F1 plants were eliminated based on glufosinate-ammonium susceptibility; the remaining BC1F1 plants were hemizygous for bar and therefore the expected segregation patterns of the unselected BC1F2 populations were the same as for the F2 populations (3:1 resistant:susceptible). Resistance or susceptibility to glufosinate-ammonium was determined by applying a commercial formulation (Finale herbicide, AgrEvo, Montvale, NJ, USA) to greenhouse-grown seedlings at the three- to five-leaf stage, as an aqueous solution at a final concentration of 0.5 g kg-1 (0.05% a.i.) using a hand sprayer equipped with a flat-fan nozzle. The herbicide solution was applied from two different directions to uniformly and thoroughly wet the leaves. Except as noted in the results, these tests were not replicated.
Additional tests were conducted to quantify the level of herbicide resistance in populations consisting of single cross- and backcross-derived plants (in the F5 and BC1F4 generations, respectively), transgenic parent sublines (in the T7 generation), and nontransgenic (recurrent) parents. These transgenic single cross- and backcross-derived populations had been characterized as homozygous for bar, based on progeny tests for qualitatively-scored resistance to glufosinate-ammonium as described above. Resistance to glufosinate-ammonium damage was quantified based on the dry weights of seedlings following a defined period of growth after application of glufosinate-ammonium solutions from 0 g kg-1 to 10.0 g kg-1 (0 to 1%) a.i. Herbicide was applied at the three- to five-leaf stage at a rate equivalent to 635 L ha-1 using a specialized spray cabinet capable of accurate and repeatable applications. This rate approximated the rate applied in previous experiments using the hand sprayer. One week after treatment with various rates of glufosinate-ammonium, plants were harvested at the soil level and dried for 24 h at 37°C, and weighed. Data were collected from three replicates; each replicate included eight seedlings of each progeny population, transgenic parent subline, or nontransgenic parent. Data were analyzed via analysis of variance, using the mean dry weight of the eight seedlings as the experimental unit, with the PROC GLM procedure within the statistical analysis software provided by the SAS Institute (1999). The option lsmeans/pdiff was invoked to detect significant differences between transgenic populations and the controls at the various treatment levels.
|
RESULTS
|
|---|
Expression of Herbicide Resistance in T1 and T3 Progeny
Application of 0.5 g kg-1 glufosinate-ammonium to sections of nontransgenic Golden Promise leaves resulted in necrosis of the treated leaf sections within 4 to 6 d. Early symptoms of damage were a very slight chlorosis of the treated tissue, followed by a loss of turgor, complete chlorosis, and eventually total desiccation of at least part of the treated tissue (Fig. 1a). Transgenic plants expressing herbicide resistance, when grown under appropriate environmental conditions, ranged from no detectable symptoms to a slight chlorosis (Fig. 1c). The tests described in this report were conducted across a period of several years; during that time it became obvious that stress, such as high temperatures (>30°C) typical of summer greenhouse conditions, resulted in reduced, variable expression of glufosinate-ammonium resistance; such data were discarded. Tests conducted under appropriate conditions, however, were found to be highly reliable, as evidenced by the cosegregation of resistance and positive bar assays, and/or by the results of numerous progeny tests (see below). Furthermore, particularly for lines with relatively low levels of bar-expression, we have found that lower rates of herbicide application, coupled with daily observation of the development of damage symptoms, are necessary to properly score segregating populations. By carefully controlling the timing and rate of glufosinate-ammonium application, greenhouse conditions, and by making multiple and careful observations of treated plants, we have been able to routinely, and reliably, identify plants expressing bar.
View larger version (96K):
[in this window]
[in a new window]
|
Fig. 1. Typical plant reactions to brush application of 0.05% a.i. glufosinate-ammonium to leaf sections of ‘Golden Promise’ (top), GP724B-4-9 T3 #30 (middle), and GP724B-4-9 T3 #31 (bottom). Each pair of leaves is from an individual plant; glufosinate was applied to a 3-cm section of each leaf directly below the punched hole.
|
|
Segregation for herbicide resistance deviated from Mendelian expectations for the segregation of single loci in the majority of transgenic families. No evidence was observed that any of these families had multiple transgenic loci. The segregation patterns of lines within families showed no significant differences, and therefore the data were pooled across lines and presented as family means in Table 1. The level of herbicide resistance appeared to vary among lines, but the nature of these assays did not allow for accurate quantification. Quantification of resistance was examined in subsequent experiments (see later in this report).
View this table:
[in this window]
[in a new window]
|
Table 1. Transgene inheritance and expression in transgenic lines from seven transgenic T0-derived families, each from an independent insertion event.
|
|
T1 progeny showed the expected 3:1 resistant:susceptible (R:S) patterns in three of seven families; none of the families showed the expected 9:7 R:S segregation patterns in the T3 progeny. These data for segregation for herbicide resistance among T1 progeny were consistent with preliminary data reported for these families (Wan and Lemaux, 1994). In all but one case where aberrant segregation was noted, fewer resistant progeny were detected than expected. The exception was for T3 progeny of GP717B-59 which showed 3:1 (vs. the expected 9:7). Data from Southern analyses (not shown; courtesy of P.G. Lemaux and R. Williams, 1995, personal communication), T1 and T3 transgene inheritance patterns (see below), and from F2 and BC1F1 herbicide resistance segregation patterns (see later in this report) indicated that there was only one transgenic locus in this family. Additional herbicide assays of T3 progeny of GP717B-59 were conducted (included in the data presented in Table 1) to verify that sampling error was not the cause of these results. The regeneration of tetraploid transgenic barley plants has been observed to be common (Choi et al., 2000), and cytogenetic abnormality was considered as a potential source of unexpected segregation patterns, but limited cytogenetic analyses (courtesy of P.G. Lemaux and R. Williams-Carrier, 1998, personal communication) detected only diploid plants. These analyses were conducted also on the other studied lines; all were observed to be diploid.
Segregation of Transgenes in T1 and T3 Progeny
Polymerase chain reaction analyses showed identical segregation patterns for the inheritance of two transgenes, bar and the BYDV CP gene, indicating (as expected) tight linkage. The data on transgene inheritance therefore are presented as single ratios representing both genes in Table 1. The majority of families showed fewer transgenic progeny than expected, indicating poor transgene transmission to progeny. Two of the families tested (GP717B-189 and GP724B-4) showed expected patterns of transgene inheritance for both generations tested, and GP717B-33 showed expected patterns in the T1 generation only. Note that for GP724B-4, the assumption was made that susceptible plants that died prior to tissue sampling were transgene negative. Note also the unexpected preponderance of transgene-positive progeny in the family GP717B-59 in both the T1 and T3 generations, which is consistent with the data for herbicide resistance.
Comparisons of segregation patterns for transgene expression vs. transgene inheritance suggest failures of transgene inheritance, and not transgene silencing, as the primary reason for the loss of herbicide resistance from this transgenic population. In most families, herbicide-susceptible/transgene-positive progeny were a small fraction of the herbicide susceptible progeny. The most notable exceptions were seen in the families GP717B-33 and GP717B-189, in which numerous transgene-positive but herbicide-susceptible progeny were detected (Table 1).
Failure of transgenic gametes to compete with nontransgenic gametes is a potential explanation for the recovery of fewer than expected transgenic progeny. Zhang et al. (1996) showed that transmission of transgenes through male gametes was lower than through female gametes. It is possible also that the PCR assays failed to detect transgenic sequences in some plants; however, Southern hybridizations on DNA of selected plants using probes specific to bar (data not shown; courtesy of P.G. Lemaux and R. Williams-Carrier, 1998, personal communication) showed perfect correlation to the PCR data. Furthermore, these tests were based not just on the detection of bar (which can be difficult to detect via PCR) but on detection of the BYDV CP gene. We have found PCR assays for the CP gene to be very reliable. Instances where PCR assays of resistant plants failed to detect transgenic sequences were rare.
Transgene Behavior in Progeny following Crossing and Backcrossing
Two transgenic sublines from GP717B-4-9, which showed the expected patterns of transgene segregation, and one transgenic subline from GP717B-59-11, which was extremely resistant to glufosinate-ammonium damage based on visual assessments, were selected for further study. Several sublines from each of these lines were used as the transgenic parents of single cross and backcross populations derived from crosses to a variety of elite cultivars and breeding lines used (Table 2). Crosses and backcrosses were made also to nontransgenic Golden Promise. Segregation for herbicide resistance among progeny of these parents generally fit Mendelian expectations for a single, dominant locus (Table 3). The preponderance of data showed segregation for herbicide resistance among F2 and BC1F1 progeny to be, respectively, 3:1 R:S and 1:1 R:S (note that transgenic parent of the BC1F1 progeny was the hemizygous F1). The exception was the unexpectedly high numbers of resistant segregants among the F2 progeny from the one cross involving GP724B-4-9 T3 #37. Note that, despite the statistical significance (based on chi-square analyses) of this aberrant segregation ratio, only 27 F2 progeny were tested; sampling error may be invoked as a possible explanation for this case.
View this table:
[in this window]
[in a new window]
|
Table 3. Segregation for glufosinate-ammonium resistance in F2 and BC1F1 progeny of crosses of transgenic barley with nontransgenic barley.
|
|
To further examine segregation patterns for glufosinate-ammonium resistance, additional progeny tests were conducted on each of the cross- and backcross-derived populations. For each population, F3 and BC1F2 progeny derived from randomly selected, resistant F2 and BC1F1 plants (12 for each generation) were treated with 0.05% glufosinate-ammonium. No differences in segregation patterns were noted for different populations derived from the same transgenic parent, and data from the various crosses to a particular parent were therefore pooled (Table 4). In general, expectations were met both for the number of homozygous and heterozygous individuals, and for the segregation patterns of heterozygous plants. The exceptions, again, were several instances of more resistant segregants than expected. The very large deviations from expectations among progeny of crosses to subline GP724B-4-9 T3 #32 are inconsistent with the normal patterns of inheritance observed among F2 and BC1F1 progeny of this subline, and an adequate explanation for this observation remains elusive.
Quantification of Herbicide Resistance
Further tests were conducted on progeny derived from crosses to the sublines GP717B-59-11 T3 #19, GP724B-4-9 #30, and GP724B-4-9 #31. Data on the dry weight of seedlings, expressed as a percentage of the dry weight of untreated control plants, following a 1-wk period of growth after herbicide treatment are shown in Tables 5 and 6. Symptoms of herbicide damage generally were evident within several days, and after 1 wk many of the nontransgenic controls sprayed with the 0.5 g kg-1 (0.05%) glufosinate-ammonium began to show signs of recovery and regrowth. Linear relationships did not exist for some of the treated groups (for instance, the transgenic parents did not exhibit reductions in dry weights except at high levels of herbicide treatment), so analysis of variance was used instead of linear regression.
View this table:
[in this window]
[in a new window]
|
Table 5. Expression of glufosinate-ammonium resistance in nontransgenic parents, transgenic parents, and progeny of crosses with transgenic parent sublines derived from transgenic line GP724B-4-9, based on dry weight of plants treated with various rates of glufosinate-ammonium.
|
|
View this table:
[in this window]
[in a new window]
|
Table 6. Expression of glufosinate-ammonium resistance in nontransgenic parents, transgenic parents, and progeny of crosses with transgenic parent sublines derived from transgenic line GP717B-59-11, based on dry weight of plants treated with various rates of glufosinate-ammonium.
|
|
The nontransgenic parents were clearly damaged by herbicide treatment, and at the highest treatment levels (5.0 and 10.0 g kg-1 [0.5 and 1%] glufosinate-ammonium) most plants rapidly ceased growth, exhibited severe damage, and were dead or nearly dead 1 wk after treatment. No differences in herbicide resistance were observed among the nontransgenic breeding lines and cultivars; all appeared equally susceptible. The differential control responses at the 0.5 g kg-1 (0.05%) treatment level (Table 5 vs. Table 6) cannot be directly compared. Note that different cultivars comprise the nontransgenic controls in these two data sets, and these data were collected from experiments conducted at different times in the greenhouse. We have observed environmentally-induced variability in herbicide damage; such variability is the likely source of the noted differences in the responses of the two groups of nontransgenic controls.
The transgenic parents exhibited obvious resistance to herbicide treatments and showed no statistically significant damage (P = 0.05) even at the highest treatment level, although there was a trend toward lower dry weights and increased damage as herbicide concentration increased.
Differences were observed in the relative level of herbicide resistance between progeny of the two transgenic lines. Transgenic progeny derived from crosses to GP717B-59-11 subline showed the highest levels of resistance, and were similar in their reactions to line GP717B-59-11; this is consistent with earlier, visual observations of a high level of resistance in this line. The progeny of crosses to the GP724B-4-9 sublines were relatively less resistant. Interestingly, and unexpectedly, the progeny of the two different sublines within GP724B-4-9 exhibited differences in herbicide resistance that were obvious from both visual observations and analyses of dry weights. No differences were detected among crosses to the same transgenic parent, nor between the two different populations (F2 and BC1F1). The data are therefore presented in Tables 5 and 6 as progeny means for each progeny population (progeny of line GP717B-59-11, and the sublines GP724B-4-9 T3 #30 and GP724B-4-9 T3 #31).
Progeny derived from crosses to subline GP724B-4-9 T3 #30 exhibited resistance very similar to that of subline GP724B-4-9 T3 #30, although there was a significant decrease in dry weight of progeny lines at the 10.0 g kg-1 treatment level not seen for the transgenic parent. Note, however, that the numerical differences between progeny and transgenic parent performance for this treatment are slight, and that the performance of subline GP724B-4-9 T3 #30 was based three data points (three reps) vs. 24 data points (three reps x four crosses x two generations) for the transgenic progeny. Additional replication of T3 #30 likely would show statistically significant damage at the 10.0 g kg-1 treatment level.
It seemed likely that the herbicide resistance differences between progeny populations of the two sublines, GP724B-4-9 T3 #30 and GP724B-4-9 T3 #31, would trace to differences between their two transgenic parent sublines. Unfortunately, such differences were unanticipated and, because of limited quantities of subline GP724B-4-9 T3 #31 seed, it was deemed sufficient to include only subline GP724B-4-9 T3 #30 as the transgenic parent check in these experiments. Following these results, however, subline GP724B-4-9 T3 #31 seed was increased and direct comparisons of the two transgenic parents were made. Both dry weight measurements (data not shown) and visual observation (Fig. 1) clearly showed that subline GP724B-4-9 T3 #30 was indeed more resistant to herbicide treatments than was subline GP724B-4-9 T3 #31, and that the resistance profile (based on dry weight measurements) of subline GP724B-4-9 T3 #31 was very similar to that of its progeny.
|
DISCUSSION
|
|---|
The majority of the transgenic families showed some level of transgene instability, with failure to transmit transgenes to progeny more common than transgene silencing. Furthermore, variability in transgene expression occurred among progeny sublines of line GP724B-4-9, both of which are derived from the same transformation event, a phenomenon that has been observed in other populations of transgenic barley and oats (S. Zhang, 1997–2002, personal communication; Bregitzer, 1997–2002, unpublished data; see also the studies referenced in the introduction). Further studies of these sublines have been conducted; both appear to have identical transgenic loci based on restriction analyses, but differential methylation has been detected in the ubiquitin promoter complex used to drive bar expression in these sublines (L. Meng, S. Zhang, P. Bregitzer, and P.G. Lemaux, 2002, unpublished data).
The origin of the observed variability among progeny of the transgenic event GP724B-4-9 is unknown, but it is possible that epistatic gene interactions may occur between transgenic loci and endogenous genes which can affect transgene expression. An erroneous, but common, description of progeny from the same transformation event as "clonal" leads to the misconception that such progeny are genetically identical. Reports in the literature have shown the widespread occurrence of heterozygous, tissue culture-induced genetic changes, such as point mutations and altered methylation patterns (Olhoft and Phillips, 1999). At the affected loci, the altered and wild-type alleles will segregate, creating a heterogeneous population with respect to these unlinked (to the transgene) genetic changes. Interactions between the transgene and these alleles could, in some cases, result in differential transgene expression among plants. For instance, heterogeneity at loci involved in mechanisms such as post transcriptional gene suppression (PTGS) could result in variability among transgenic progeny for the propensity to silence transgenes. Such a possible scenario is especially intriguing given preliminary evidence that the bar locus in GP724B-4-9 may contain an inverted repeat (L. Meng, Zhang, Bregitzer, and Lemaux, 2002, unpublished data); inverted repeats have been shown to efficiently induce post-transcriptional gene suppression; Waterhouse et al., 1998, 2001; Morino et al., 1999).
Although the results of this study do not address the underlying molecular phenomena, the observed inheritance and expression patterns provide important empirical information that will be useful in designing breeding approaches. The observed instabilities in inheritance and expression suggest a limited utility of early generation testing for transgene expression, and a need for screening large populations derived from multiple transgenic events. For instance, if the majority of transgenic events are unstable with respect to inheritance and/or expression, then it is somewhat pointless to expend substantial resources on early generation characterizations of transgene inheritance and/or expression. It would be far more productive to advance many T0–derived transgenic populations in bulk. Transgenic loci that survive (both physically and with respect to expression) several meiotic cycles would be good candidates for further study; highly unstable events would have been naturally removed from the population with minimal effort and expense. Consistent with this idea is our observation that progeny of crosses to advanced generation selections of GP717B-59-11 and GP724B-4-9 generally showed expected patterns of transgene inheritance and parental levels of transgene expression.
Variable expression among "clonal" progeny dictates screening multiple lines and sublines from each transgenic family. For instance, sublines GP724B-4-9 T3 #30 and T3 #31 were not equivalent with respect to herbicide resistance, and adequate characterization of the GP724B-4-9–derived population could not have been based on early generation tests, nor based on tests of a single selection from within the line. What is the optimum number of such selections necessary to efficiently screen for the best transgenic parent? Additional research into the nature and scope of such variability, in combination with classical genetic research, will be necessary to answer this question. However, it is obvious from this and other studies that transgenic parents should not be handled in a breeding program simply as sources of simply inherited, single-gene sources of a particular trait. Relative to nontransgenic sources of qualitatively inherited traits, breeding procedures for transgenic parents may need to be modified by increasing population sizes, delaying testing to later generations, and consideration of the potential for significant heterogeneity for transgene expression within any given transgenic population.
Finally, one can conclude that increased stability of transgene behavior is desirable. Transformation systems have been developed that have significantly lower occurrences of genetic instability (Lemaux et al., 1999) that are manifested through enhanced plant regenerability from cultured barley tissues, and lower incidences of phenotypic abnormalities (Zhang et al., 1999; Bregitzer et al., 2002). It is reasonable to expect reductions in transformation-induced genomic perturbations to be associated with greater transgene stability, given the potential of these perturbations to pleiotropically effect transgene expression. Recently developed transgenic barley populations that have made use of the improved transformation technologies seem to exhibit improved transgene stability, at least with respect to inheritance (P.G. Lemaux, 2000–2002, personal communication).
Another step toward transgene stability has been seen with the delivery of single, transgenic loci to various genomic locations via Ac-transposase mediated Ds element transposition. Transposed transgenic loci have exhibited significant improvements in expression stability (Koprek et al., 2001). Experiments are underway to further characterize the value of this system for transformation-based germplasm development.
|
ACKNOWLEDGMENTS
|
|---|
The authors thank consulting ARS Statistician, Gary Richardson, for his expert assistance; Doctors Peggy G. Lemaux, Shibo Zhang, Myeong-Je Cho, Thomas Koprek, and Ms. Rosalind Williams-Carrier and Mr. Robert Nunez for valuable intellectual and technical contributions to this work; and Doctors Bill Berzonsky and Bob Graybosch for critical reviews of this manuscript.
Received for publication November 6, 2001.
|
REFERENCES
|
|---|
- Bregitzer, P., S.E. Halbert, and P.G. Lemaux. 1998. Somaclonal variation in the progeny of transgenic barley. Theor. Appl. Genet. 96:421–425.
- Bregitzer, P., S. Zhang, M.-J. Cho, and P.G. Lemaux. 2002. Reduced somaclonal variation is associated with culturing highly differentiated, meristematic tissues. Crop Sci. 42:1303–1308.[Abstract/Free Full Text]
- Cho, M.-J., H.W. Choi, B.B. Buchanan, and P.G. Lemaux. 1999. Inheritance of tissue-specific expression of barley hordein promoter-uidA fusions in transgenic barley plants. Theor. Appl. Genet. 98:1253–1262.
- Choi, H.W., P.G. Lemaux, and M.-J. Cho. 2000. Increased chromosomal variation in transgenic versus nontransgenic barley (Hordeum vulgare L.) plants. Crop Sci. 40:524–533.[Abstract/Free Full Text]
- Kioussis, D., and R. Festenstein. 1997. Locus control regions: Overcoming heterochromatin-induced gene inactivation in mammals. Curr. Opin. Genet. Dev. 7:614–619.[ISI][Medline]
- Koprek, T., S. Rangel, D. McElroy, J.D. Louwerse, R. Williams-Carrier, and P.G. Lemaux. 2001. Transposon-mediated single-copy gene delivery leads to increased transgene expression stability in barley. Plant Physiol. 125:1354–1362.[Abstract/Free Full Text]
- Kumpatla, S.P., and T.C. Hall. 1998. Recurrent onset of epigenetic silencing in rice harboring a multi-copy transgene. Plant J. 14:129–135.
- Lemaux, P.G., M.-J. Cho, S. Zhang, and P. Bregitzer. 1999. Transgenic cereals: Hordeum vulgare L. (barley). p. 255–316. In I.K. Vasil (ed.) Molecular improvement of cereal crops. Kluwer Academic Publ., Dordrecht, The Netherlands.
- Morino, K., O.-A. Olsen, and K. Shimamoto. 1999. Silencing of an aleurone-specific gene in transgenic rice is caused by a rearranged transgene. Plant J. 17:275–285.[ISI][Medline]
- Müller, E., H. Lörz, and S. Lütticke. 1996. Variability of transgene expression in clonal cell lines of wheat. Plant Sci. 114:71–82.
- Olhoft, P.M., and R.L. Phillips. 1999. Genetic and epigenetic instability in tissue culture and regenerated progenies. p. 111–148. In H.R. Lerner (ed.) Plant responses to environmental stresses: From phytohormones to genome reorganization. Marcel Dekker, New York.
- Pawlowski, W., K.A. Torbert, H.W. Rines, and D.A. Somers. 1998. Irregular patterns of transgene silencing in allohexaploid oat. Plant Mol. Biol. 38:597–607.[ISI][Medline]
- Register, J.C., D.J. Peterson, P.J. Bell, W. Paul Bullock, I.J. Evans, B. Frame, A.J. Greenland, N.S. Higgs, I. Jepson, S. Jiao, C.J. Lewnau, J.M. Sillick, and H.M. Wilson. 1994. Structure and function of selectable and nonselectable transgenes in maize after introduction by particle bombardment. Plant Mol. Biol. 25:951–961.[ISI][Medline]
- Rochow, W.F. 1979. Field variants of barley yellow dwarf virus: Detection and fluctuation during 20 years. Phytopathology 68:655–660.
- SAS Institute. 1999. SAS/STAT user's guide. Version 7-1. SAS Inst., Cary, NC.
- Spencer, T.M., J.V. O'Brien, W.G. Start, T.R. Adams, W.J. Gordon-Kamm, and P.G. Lemaux. 1992. Segregation of transgenes in maize. Plant Mol. Biol. 18:201–210.[ISI][Medline]
- Srivastava, V., V. Vasil, and I.K. Vasil. 1996. Molecular characterization of the fate of transgenes in transformed wheat (Triticum aestivum L.). Theor. Appl. Genet. 92:1031–1037.
- Wan, Y., and P.G. Lemaux. 1994. Generation of large numbers of independently transformed barley plants. Plant Physiol. 104:37–48.[Abstract]
- Waterhouse, P.M., M.W. Graham, and M.-B. Wang. 1998. Virus resistance and gene silencing in plants can be induced by simultaneous expression of sense and antisense RNA. Proc. Natl. Acad. Sci. USA 95:13959–13964.[Abstract/Free Full Text]
- Waterhouse, P.M., M.-B. Wang, and T. Lough. 2001. Gene silencing as an adaptive defence against viruses. Nature (London) 411:834–842.[Medline]
- Zhang, S., M.-J. Cho, P. Bregitzer, and P.G. Lemaux. 1999. Comparative analysis of genomic DNA methylation status and field performance of plants derived from embryogenic calli and shoot meristematic cultures. p. 263–267. In A. Altman et al. (ed.) Plant biotechnology and in vitro biology in the 21st century. Kluwer Academic Publ., Dordrecht, The Netherlands.
- Zhang, S., D. Warkentin, B. Sun, H. Zhong, and M.B. Sticklen. 1996. Variation in the inheritance of expression among subclones for unselected (uidA) and selected (bar) transgenes in maize (Zea mays). Theor. Appl. Genet. 92:752–761.
TOP
ABSTRACT
INTRODUCTION
