Published online 5 June 2006
Published in Agron J 98:1060-1064 (2006)
DOI: 10.2134/agronj2005.0104
© 2006 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
Notes and Unique Phenomena
Pollination Competition Effects on Gene-Flow Estimation: Using Regular vs. Male-Sterile Bait Plants
Junming Wanga,
Xiusheng Yangb,*,
Yi Lic and
Phillip F. Elliottd
a Dep. of Plant and Environmental Sciences, New Mexico State Univ., MSC3Q, Box 30003, Corner of Knox and College Street, Las Cruces, NM 88003-8003
b Dep. of Natural Resources Management and Engineering, Univ. of Connecticut, 1376 Storrs Road, Storrs, CT 06269-4087
c Dep. of Plant Science, Univ. of Connecticut, 1376 Storrs Road, Storrs, CT 06269-4067
d Dep. of Biology, Eastern Connecticut State Univ., Willimantic, CT 06226
* Corresponding author (xiusheng.yang{at}uconn.edu)
Received for publication April 9, 2005.
 |
ABSTRACT
|
|---|
Pollen-mediated gene flow from transgenic crops is a concern of the scientific community as well as the general public. Although a common practice, the use of male-sterile bait plants in field trials to demonstrate rates of gene transfer has been questioned due to the lack of pollination competition. However, little direct evidence has been published. Field experiments of male-sterile and male-fertile corn bait plants were conducted in 2001 and 2002, respectively, to evaluate the effects of pollination competition on gene-flow assessment. Male-sterile bait plants exhibited a significantly higher rate of outcrossing than male-fertile plants. The results obtained from this study suggest that actual gene flow from transgenic plants to their wild-type cultivars or relatives is likely to be lower than estimates reported in previous studies using male-sterile bait plants, and that male-fertile that is, normal, bait plants should be used in future studies attempting to estimate gene flow.
Abbreviations: ANOVA, analysis of variance • GLM, general linear model • GMC, genetically modified crop • Ln, natural logarithm • Max, maximum • Min, minimum • NS, not significant • Std, standard deviation
|
INTRODUCTION
|
|---|
THE rapid development and commercialization of genetically modified crops (GMCs) have caused worldwide concern about the potential consequences of their use (Persley and Siedow, 1999; Grogan and Long, 2000). The introduction of GMCs has potentially profound effects on food safety, biodiversity, and the environment. Transgenic corn (Zea mays L.) was among the first four pest-resistant crops to flow from the industrial R&D pipeline to commercial production. The area planted with Bt corn (a transgenic hybrid) is estimated to be more than one-third of the total acreage of corn fields in the USA (Grogan and Long, 2000); consequently, among GMCs, transgenic corn may have the highest potential to negatively affect the environment (Altieri, 1998). Not surprisingly, maize has been used extensively in gene-flow trials; and these studies have been comprehensively reviewed by Emberlin et al. (1999).
In past gene-flow studies, researchers have often used male-sterile plants as the bait plants because it simplified the experimental procedures (Timmons et al., 1995, 1996; Squire et al., 1999; Thompson et al., 1999). In these experiments, quantifying gene flow was straightforward, because all pollen came from source plants, and all seeds were resulted from source pollen. In contrast, when using male-fertile regular bait plants, pollen from source plants must be distinguished from bait plant pollen, and seeds resulted from pollination with source plants must be distinguished from seeds resulted from pollination with bait plants. Riegaer and Preston (1999) and Thompson et al. (1999) questioned the results of experiments using male-sterile bait plants, pointing out that gene flow might have been overestimated due to experimental designs that eliminated the normal pollination competition between source and bait plants.
Although the validity of gene-flow estimates from male-sterile bait plant experiments has been questioned, little direct evidence has been published to support the hypothesis that gene flow may be consistently overestimated by removing the effects of pollination competition through the use of male-sterile bait plants. The objective of this study was to quantify the pollination competition effects on gene-flow assessment.
|
Method
|
|---|
Experimental Crops and Meteorology
Field experiments were conducted in 2001 and 2002 at the University of Connecticut Agronomy Research Farm using male-sterile and male-fertile bait plants to test the hypothesis that gene flow is overestimated by studies using male-sterile bait plants. In each experiment, source plants were grown in circular areas with a diameter of 16 m, surrounded by bait plants grown in the same density (Fig. 1
). Male-sterile corn was used as bait plants in 2001, whereas male-fertile plants were used as bait plants in 2002. In 2001, sweet corn (Jackpot F-1, Hoffman Seeds, Inc., Landisville, PA) was grown as the source crop at a density 53400 plants ha–1, surrounded by receptor plants of the same cultivar of corn grown in the same density. The receptor plants were manually detasseled just before the pollination seasons. In 2002, 8464Wx waxy mutant corn with yellow-color kernel was planted as source plants and 8419W regular white corn as the bait plants (Garst Seeds Company, Slater, IA), with a uniform density of 71200 plants ha–1. The corn cultivars used in the experiments were all hybrids and not related to each other.
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 1. Schematic illustration of the experimental setup (stars represent the sampling locations for the pollen deposition flux).
|
|
In each year, the same experiment was duplicated at two nearby fields subject to the same meteorological and environmental conditions. The first field was planted in the middle or late May, while the second field was planted about 4 wk later to avoid cross-plot pollination. For both years and fields, plants started shedding about 60 d after planting and the plants started silking about 3 d after shedding started. Each pollination period lasted about 2 wk. The plants grew to an average height of 1.5 m and ears were at 0.7-m height in male-sterile experiment; plants in the male-fertile experiment grew to an average of 2.9 m and ears were at 1.8-m height.
An automated weather station at the experimental site recorded 30-min averages and variations in solar radiation, precipitation, temperature, relative humidity, and wind speed. The weather conditions during the two growing seasons were characterized by the descriptive statistics (mean, standard deviation, and 30-min mean minimum and maximum) of the above mentioned parameters.
Pollen Deposition Measurement
Microscope slides (2.5 by 7.5 cm) with silicon grease (Surveillance Data, Inc., Plymouth Meeting, PA) were used to measure pollen depositions at silk height in the receptor field along two sampling lines (Fig. 1). In 2001, pollen was sampled every 10 m in the receptor field, while in 2002 the sampled distances were revised at 1, 4, 7, 10, 15, 25, 35, 45, and 60 m, respectively, because our results from 2001 indicated that pollen deposition decreased exponentially with distance and after 50 m the pollen deposition variation with distance was very small.
Pollen sampling was conducted throughout each pollination season. Each sampling period for the collectors was every 1.5 or 3 h during daytime (0700–1930 h). No samples were collected at night or on rainy days because little, if any, pollen was released during those time periods.
Laboratory Analysis and Deposition Calculation
A microscope with power magnification of 3.5x and binocular eyepiece of 15x (B-35-83, American Optical Co., Instrument Division, Buffalo, NY) was used to measure the pollen parameters and visually count the pollen collected on the slides. For each cultivar, the average pollen grain size was estimated from a sample of 10 randomly selected pollen grains. One-way analysis of variance (ANOVA) was used to compare average pollen diameters for the three different corn cultivars (sweet, 8464Wx, and 8419W).
Total deposition of pollen grains was estimated from random samples (n = 8) of circular areas (d = 5.3 mm) drawn from each slide, following the standard procedures of cluster sampling (Scheaffer et al., 1990). The sampling error ranged from 12 to 34%, depending on the deposition density. Corn pollen has corn kernel shape, and can be easily distinguished from other pollen and particles. For distinguishing the pollen grains produced by the source plants from those by the bait plants in the male-fertile experiment, the classical method in Brink and MacGillivray (1924) was used by applying iodine solution to stain the samples before counting under the microscope. Common cornstarch is approximately 73% amylopectin and 27% amylose, whereas waxy starch is composed entirely of amylopectin, which is in the branched molecular form. Pollen grains from the regular bait plants with ordinary cornstarch stain blue with 2% potassium iodide solution, whereas the pollen grains from the waxy source plants with waxy cornstarch stain a reddish brown (Brown and Darrah, 2002).
The average pollen count per unit sampling area of selected clusters for each slide per unit time was regarded as the deposition flux density during each sampling period at the given location (grains cm–2 s–1). The collection efficiency of the slide traps was assumed to be 100%, according to Raynor et al. (1970), and Aylor and Ferrendino (1989).
The grand total deposition flux (grains cm–2) at each location was obtained by integrating the deposition flux density over the whole silking season. The integration excluded the periods of time in which it rained, when little pollen was released and the pollen was soaked, saturated, and broken with water.
Outcrossing Ratio
Seed set can be used as an index of outcrossing in population structure. After seed shelling, at the location of each deposition sampling point, four surrounding plants were sampled. In the male-sterile experiment, the number of kernels counted on each sampled ear was regarded as the measure of seed set, because the male-sterile bait plants could not have contributed pollen to cause fertilization of ovules. Ten ear samples were also sampled in the center of the source to estimate the average kernel number of a fully filled ear. In 2002, fertilization was made by pollen from both 8464Wx and 8419W plants. Because yellow endosperm allele is dominant over white endosperm allele, kernels on the 8419W plants were white if the ovules were pollinated by 8419W, and yellow if pollinated by 8464Wx. The seed set from the source plants, therefore, was detected and counted by color.
Outcrossing ratios were calculated for the two bait plants from the measurements of seed set. For the male-sterile bait experiments, the outcrossing ratio for each ear sample was estimated by dividing the seed set value of the sample by the average seed set value of the source plants. Mean value over the four ear samples was reported as the outcrossing ratio for each sampling location. For the male-fertile experiment, the value of the outcrossed kernel number divided by the total kernel number for each ear sample was obtained and the average of the resulting values at each sampling point was regarded as the outcrossing ratio at the corresponding location.
|
Data Analysis
|
|---|
Fertilization Ratio Index
A fertilization ratio index (I) was defined and used to quantify the pollination competition effects on gene-flow estimation. Mathematically, the fertilization ratio index is expressed as:
 | [1] |
where x is the outcrossing ratio at each sampling location (% outcrossing), and F is the corresponding grand total deposition flux of source pollen grains at silk height (grains cm–2). The so-defined fertilization ratio index represents the efficiency of fertilization at a given location for a unit deposition flux of pollen from the source plants. Because the deposition flux was determined at the silk height of the bait plants, the fertilization ratio index should be independent of plant canopy structure and density, and provides a measure to assess the pollination competition effects on gene-flow estimation.
Pollination Competition Effects on Gene-Flow Estimation
The ANOVA and regression analysis were used to determine if the fertilization ratio index and its spatial distribution were significantly different between the two different bait plants. For ANOVA, the fertilization ratio index was classified into five groups corresponding to the distance from the source, for example, Group 1: <10 m; Group 2: 10 to 20 m; Group 3: 20 to 30 m; Group 4: 30 to 50 m; and Group 5: 50 to 100 m. Comparisons were made across factors of cultivar, distance, and sampling line by using general linear model (GLM). Before ANOVA, the Bartlett's test of equal variance for the fertilization ratio index was conducted on each factor (Ott, 1993). Because the Bartlett's test revealed that the variance of the fertilization ratio index was significantly different among different factors (Bartlett's statistic = 81.34, P < 0.001), the variable was transformed to its natural logarithmic form in the analysis. The variance of the transformed fertilization ratio index among different factors was not significantly different (Bartlett's statistic = 23.95, P = 0.35).
Linear regression analysis was employed to determine if the fertilization ratio index varied as a function of distance, and if the variation was different between the male-sterile and male-fertile bait plants. F-tests were used to evaluate the statistical difference between the resulting slopes of the relationships for male-sterile or male-fertile bait plants. All data analyses were done with the commercial statistical software Minitab Version 13.3 (MINITAB, 2000).
|
Results
|
|---|
Pollen Diameter and Weather
Pollen sizes among the three corn cultivars used in this study did not differ significantly (one-way ANOVA F = 0.84, P > 0.4). The average diameter of the sampled pollen grains was 82.2 µm for sweet corn, 82.9 µm for 8464Wx, and 84.8 µm for 8419W, respectively. The weather conditions of the two growing seasons were similar (Table 1).
Pollination Competition Effects
Table 2 summarizes the results of ANOVA on the means of the transformed fertilization ratio index with respect to factors of cultivar, distance, and sampling line. The results indicated that the fertilization ratio index was significantly affected by cultivar, distance from the source, and the interaction between cultivar and distance. No significant effects were detected for any of the remaining interaction terms (Table 2).
View this table:
[in this window]
[in a new window]
|
Table 2. Analysis of variance (ANOVA) results on the logarithmically transformed fertilization ratio index (% fertilization/[grains cm–2]) with respect to factors of cultivar, distance, and sampling line.
|
|
The mean fertilization ratio index of the male-sterile bait plants was much higher than that of the male-fertile bait plants. Averaged over the experiment plots, sampling lines, and all the sampling distances, the mean fertilization ratio index for the male-sterile bait plants was 0.89 in percent outcrossing per unit deposition flux of source pollen, whereas that for the mail-fertile bait plants was only about 0.05, a 17-fold difference. The variation of the fertilization ratio index with distance from the source was also significant, but only for the male-fertile bait plants. This can be clearly shown in Fig. 2
in which the relationship between the fertilization ratio index and sampling distance was plotted respectively for the two bait plants. For the male-fertile bait plants, the fertilization ratio index declined linearly in the logarithmic scale as a function of distance (slope = –0.008, F = 4.34, P < 0.05). For the male sterile bait plants, distance had no effect on the fertilization ratio index (F = 1.30, P > 0.26).
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 2. Variation of fertilization ratio index with distance for male-sterile (a) and male-fertile (b) bait plants.
|
|
In sum, it was found that there existed a significant difference in the fertilization ratio index between the male-sterile and male-fertile bait plants. The difference was in two aspects. First, the magnitude of the outcrossing for a unit deposition flux of source pollen grains was much higher for the male-sterile than for the male-fertile bait plants. Second, the fertilization ratio index for the male-fertile bait plants declined with distance from the source.
|
Discussion
|
|---|
The major finding in this study is that rates of outcrossing were significantly higher in plots with male-sterile bait plants than in those where male-fertile bait plants were used. This result is consistent with that of previous studies and supports the argument of Riegaer and Preston (1999) and Thompson et al. (1999) that gene flow is likely to be overestimated by studies using male-sterile bait plants. Previously published gene-flow experiments using male-sterile oilseed rape (Brassica napus L.) bait plants (e.g., Timmons et al., 1996; Thompson et al., 1999) obtained much higher gene-flow levels than the experiments using normal oil rape plants (e.g., Scheffler et al., 1993; Paul et al., 1995). Thompson et al. (1999) obtained 49% gene-flow level at 100 m from source using male-sterile bait plants, while the gene-flow level in Scheffler et al. (1993) was only 0.1% at 10 m using male-fertile bait plants.
We believe that pollination competition between source and bait plant is the most likely explanation for the dramatic differences observed between rates of outcrossing between male-sterile and male-fertile bait plants. With the presence of pollen from both source and bait plants, an ovule was fertilized by the first deposited pollen grain. The availability of pollen from the local bait plants thus significantly reduced the fertilization effectiveness of the distant source pollen. This competition effect became more and more significant with distance from the source, as it took longer and longer time for the source pollen to travel. In the absence of bait source pollen, the a priori expectation would be that the fertilization ratio index, standardized to unit pollen grains per square centimeter, should show no response to increasing distance, although the total pollen deposition declines with distance.
It probably should be noted that the original purpose of the experiments was to provide gene-flow data for model development and validation instead of testing pollination competition effects. That was why we used sweet corn, which had mixed kernel color, as the experimental crop in 2001. In 2002, in an attempt to include the effects of pollination competition in our quantitative model of gene flow, we chose to use the 8464Wx waxy mutant and 8419W regular corn to be our source and bait plants, respectively, to take the advantages of their different colors in pollen and seeds. The use of different cultivars of corn in different experimental years, different planting densities and plant height in the experiments may have introduced potential sources of error and it encourages caution in the conclusions drawn. However, because of the manner in which the fertilization ratio index was defined, plant cultivar, height, or density could not systematically bias the results.
The results of this paper suggest that published estimates of gene flow derived from studies using male-sterile plants overestimate gene flow and that pollination competition is a significant factor influencing spatial patterns of outcrossing. In a more applied context, our results suggest that pollination competition should be taken into account when attempting to estimate the gene-flow risk posed by genetically modified crops.
|
ACKNOWLEDGMENTS
|
|---|
The authors thank the farm crew of the University of Connecticut Agronomy Research Farm for their technical assistance. Financial support for the research was partially provided by the University of Connecticut Research Foundation and the Storrs Agricultural Experiment Station.
|
REFERENCES
|
|---|
- Altieri, M.A. 1998. The environmental risks of transgenic crops: An agroecological assessment. Dep. of Environ. Sci., Policy, and Management, Univ. of California, Berkeley.
- Aylor, D.E., and F.J. Ferrendino. 1989. Dispersion of spores released from an elevated line source within a wheat canopy. Boundary-Layer Meteorol. 46:251–273.
- Brink, R.A., and J.H. MacGillivray. 1924. Segregation for the waxy character in maize pollen and differential development of the male gametophyte. Am. J. Bot. 11:465–469.[CrossRef][ISI]
- Brown, W.L., and L.L. Darrah. 2002. Origin, adaptation, and types of corn, national corn handbook. Available at www.inform.umd.edu (verified 20 Mar. 2006).
- Emberlin, J., B. Adams-Groom, and J. Tidmarsh. 1999. A report on the dispersal of maize pollen. National Pollen Research Unit, University College, Worcester. Available at www.soilassociation.org/ (verified 20 Mar. 2006). Soil Assoc., Bristol, UK.
- Grogan, J., and C. Long. 2000. The problem with genetic engineering. Organic Gardening 2000 (J/F), p. 42–47.
- MINITAB. 2000. Minitab reference manual, release 13.3. Minitab, State College, PA.
- Ott, R.L. 1993. An introduction to statistical methods and data analysis. Wadsworth, Belmont, CA.
- Paul, E.M., C. Thompson, and J.M. Dunwell. 1995. Gene dispersal from genetically modified oil seed rape in the field. Euphytica 81:283–289.[CrossRef][ISI]
- Persley, G.J., and J.N. Siedow. 1999. Applications of biotechnology to crops: Benefits and risks. Issue Paper 12. CAST, Ames, IA.
- Raynor, G.S., J.V. Hayes, and E.C. Ogden. 1970. Experimental data on dispersion and deposition of timothy and corn pollen from known sources. Brookhaven National Lab., Upton, NY.
- Riegaer, M.A., and C. Preston. 1999. Gene flow from transgenic canola to wild radish—A model system to determine the risks. In P.J.W. Lutman (ed.) Gene flow and agriculture: Relevance for transgenic crops. British Crop Protection Council Symp. Proc. 72, Univ. of Keele, Staffordshire, UK. 12–14 Apr. 1999. BCPC, Alton, Hampshire, UK.
- Scheaffer, R.L., W. Mendenhall, and L. Ott. 1990. Elementary survey sampling. 4th ed. PWS-Kent Publ. Co., Boston, MA.
- Scheffler, J.A., P. Parkinson, and P.J. Dale. 1993. Frequency and distribution of pollen dispersal from transgenic oilseed rape (Brassica napus). Trans. Res. 2:356–364.[CrossRef][ISI]
- Squire, G.R., J.W. Crawford, G. Ramsey, C. Thompson, and J. Bown. 1999. Gene flow at the landscape level. In P.J.W. Lutman (ed.) Gene flow and agriculture: Relevance for transgenic crops. British Crop Protection Council Symp. Proc. 72, Univ. of Keele, Staffordshire, UK. 12–14 Apr. 1999. BCPC, Alton, Hampshire, UK.
- Thompson, C.E., G. Squire, G.R. Mackay, J.E. Bradshaw, J. Crawford, and G. Ramsay. 1999. Regional patterns of gene flow and its consequence for GM oilseed rape. In P.J.W. Lutman (ed.) Gene flow and agriculture: Relevance for transgenic crops. British Crop Protection Council Symp. Proc. 72, Univ. of Keele, Staffordshire, UK. 12–14 Apr. 1999. BCPC, Alton, Hampshire, UK.
- Timmons, A.M., E.T. O'Brien, Y.M. Charters, S.J. Dubbels, and M.J. Wilkinson. 1995. Assessing the risks of wind pollination from fields of genetically modified Brassica napus ssp. Oleifera. Euphytica 85:417–423.[CrossRef][ISI]
- Timmons, A.M., Y.M. Charters, J.W. Crawford, D. Burn, S.E. Scott, N.J. Wilson, A. Roberston, E.T. O'Brien, G.R. Squire, and M.J. Wilkinson. 1996. Risks from transgenic crops. Nature 380:487.[CrossRef][Medline]
TOP
ABSTRACT
INTRODUCTION
