| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Plants, Soils, and Biometeorology Dep., Utah State Univ., Logan, UT 84322-4820
b USDA-ARS Forage and Range Research Lab., Utah State Univ., Logan, UT 84322-6300
c The Scotts Company, Gervais, OR 97026
d The Scotts Company, Marysville, OH 43041
* Corresponding author (paul.johnson{at}usu.edu)
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Improvement of P. pratensis through intentional breeding is fairly recent compared with most agriculturally important crops. Most early turfgrass breeding efforts (pre 1970s) used collections from populations under various forms of management, by selecting plants showing desirable stress tolerance, growth habit, and appearance. (Bashaw and Funk, 1987; Huff, 2003). Although apomixis simplifies seed increase of P. pratensis varieties and provides for uniformity in turf settings, this asexual reproduction complicates traditional breeding processes by reducing hybridization and genetic recombination opportunities (Bashaw and Funk, 1987; Huff, 2003). Poa pratensis also displays complex forms of polyploidy that may obscure trait segregation, inheritance, and expression leading to complicated inheritance and trait expression during sexual reproduction (Wendel, 2000). Matzk et al. (2005) have recently reported a model for the control of asexual seed formation in P. pratensis involving five genes.
The complex evolution and taxonomy of this genus is largely due to extensive hybridization and introgression among Poa species (Bor, 1952; Clausen, 1961). Some of the hybrids with P. pratensis reported include P. alpina L., P. arachnifera, P. arctica R. Br., P. compressa, P. longifolia Trin., P. nemoralis L., P. nervosa (Hook.) Vasey, P. palustris, P. reflexa Vasey & Scribn, P. trivialis, and P. secunda (Knobloch, 1968; Welsh et al., 1987). In addition, apomixis has allowed many Poa species to overcome consequences that interspecific hybridization may have on sexual reproduction—most importantly sterility of progeny (Clausen, 1961).
Hybridization among Poa species and introgression of genes may be facilitated by the sharing of genomes that species have in common. Poa pratensis likely contains the genomes from up to four progenitor species, and these are shared with other allopolyploid Poa species (Patterson et al., 2005). Because of the similarities among the species and environments in which they are found, many opportunities appear to exist for gene flow among populations of Poa (Johnson and Riordan, 1999) and similar to situations in Agrostis (Wipff and Fricker, 2001; Watrud et al., 2004).
Most data on the topic of gene flow are on cultivated crop plants including canola, sunflower (Helianthus annuus L.), and corn (Zea mays L.) (Stewart, 2004), while a more limited amount of information exists for perennial grasses. Modeling of potential gene flow of wind-pollinated grasses through pollen movement has indicated extensive gene flow in some locations and relatively little in others (Meagher et al., 2003). Field hybridization studies of Agrostis stolonifera L. have shown gene flow distances of up to 298 m from a pollen source of 286 plants and extrapolated gene flow through modeling to 1066 m in one study (Wipff and Fricker, 2001). Gene flow was reported from an A. stolonifera seed field of approximately 162 ha to a sentinel plant of Agrostis at a distance of 21 km (Watrud et al., 2004). Gene flow in perennial ryegrass (Lolium perenne L.) was measured at less than 2% at 144 m in a downwind direction from 268 plants as a pollen source (Cunliffe et al., 2004) and less than 1% in tall fescue at 150 m with no gene flow observed at 200 m using 49 plants as a pollen source (Wang et al., 2004).
A number of turfgrasses have been the focus of transgenic breeding, including tall fescue (Festuca arundinacea Schreb.), perennial ryegrass, bermudagrass [Cynodon dactylon (L.) Pers.], and creeping bentgrass (Zilinskas and Wang, 2004) because of strong interest in the turfgrass industry for herbicide resistance, pest resistance, and environmental stress tolerance traits (Johnson and Riordan, 1999; Ostermeyer, 2004). Because large amounts of Kentucky bluegrass are used around the world, a transgenic Poa may be petitioned for a future release. In order for such a petition to be evaluated properly, risk assessment questions, including the potential for introgression of genes from P. pratensis into other species, must be considered to establish an understanding of how such varieties may impact local and regional ecosystems.
Although intraspecific and interspecific hybrids involving P. pratensis have been described, quantitative assessments of relative hybridization frequencies under field conditions are lacking. Moreover, the impact of apomixis on pollen-mediated gene flow is not fully understood. Our objective was to quantify intra- and interspecific pollen-mediated gene flow from a P. pratensis genotype to other Poa species by testing hybridization potential in field conditions using herbicide tolerance as a selectable trait.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
Weather data was recorded at the plot area on 15-min intervals from April through July of 2002 and 2003 with a WatchDog 900ET weather station (Spectrum Technologies, Plainfield, IL). Measurements included wind direction, wind speed, and wind gusts, relative humidity, temperature, and solar radiation.
In 2002 and 2003, all plants were monitored weekly for inflorescence emergence and anthesis from May through early July and at least twice weekly during anthesis of BR99-1033. Seed from the receptor plot was harvested from each plant when mature but before excessive shattering. This occurred from mid-June to early July. Seed from P. annua and P. supina Schrad. was harvested throughout June and early July because of their indeterminate flowering. Seed from BR99-1033 was harvested in early July of both years. All seed was air dried and cleaned by hand.
Seedling Evaluation
We screened seedlings from the receptor plants for resistance to glyphosate. Seed from the receptor plants were sowed into greenhouse flats containing moist potting mix composed of 50% peat and 50% perlite (by volume), covered lightly, watered, and stratified for at least 2 wk at 4 to 6°C. Control flats were also seeded and evaluated, which included seedlings of glyphosate susceptible P. pratensis ‘Coventry’ and seedlings from BR99-1033. After stratification, the flats of seed were moved to a warm greenhouse 22°C/18°C for germination and growth. When seedlings were at the 1 to 2 leaf stage, they were counted individually, if under 100 seedlings per flat. If more than 100 seedlings germinated per flat, a grid was usually used to count seedlings in a random 10% of the flat and a total number of seedlings in the flat was then estimated. The seedlings were sprayed with a 1% (v/v) glyphosate solution at the 2 leaf stage, and a second spray, at the same concentration, was made 14 d later with a complete kill of susceptible seedlings obtained within 4 wk of the initial spray. Surviving seedlings were transplanted to pots and grown for verification. Flats were monitored every 2 to 3 d for additional germination after the seedling counts. Additional seedling counts and herbicide sprays were conducted if new seedlings were identified.
Hybrid Confirmation
The presence of transgenic DNA in putative hybrids was verified by PCR amplification of the cauliflower mosaic virus 35S-promoter sequences and Agrobacterium tumefaciens nopaline synthase-terminator (NOS) sequences, using the 35S-1//35S-2 and NOS-1//NOS-3 primer combinations (Table S1) described by Lin et al. (2001). The identity of the 195 bp 35S.1//35S.2 and 180 bp NOS.1//NOS.3 amplicons from the BR99-1033 genotype was verified by DNA sequencing. Routine screening with this PCR assay was performed by agarose gel electrophoresis using 100-bp ladder size standards, transgenic BR99-1033 positive control, nontransgenic Coventry negative control, and water negative control as references for each set of assays.
To eliminate possible seed contaminations from BR99-1033, the maternal parent identity was verified by sequencing polymorphic chloroplast DNA regions of the ndhF gene and/or trnK-rps16 intergenic spacer. The ndhF gene was amplified and directly sequenced using primers (Table S1) described by Olmstead and Sweere (1994). The trnK-rps16 intergenic spacer was amplified and directly sequenced using the primers (Table S1) described by (Kress et al., 2005).
Data Analysis
Hybrid frequency data were analyzed in comparison to linear, quadratic, logarithmic, and exponential decay function models that mimic gene or pollen flow to describe the relationship of percentage of hybrids with distance from the pollen source. These analyses were done using SigmaPlot 8.0. Wind summaries were made using WindRose version 2005-02-03 (Enviroware, Agrate Brianza, Italy).
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Weather Summary
Wind patterns were variable and distinctly different in each year of the study. In 2002, winds started predominantly from the SE then switched to the SW, then NE, then finally returned to the SW. Very little wind came from the W or NW directions overall. In 2003, wind direction was dominated by NW, S, and SE directions with little from the SW or NE directions. In 2002, average wind speeds were less than 2 m s–1 50% of the time and 15% were above 4 m s–1. In 2003, wind speeds were less than 2 m s–1 55% of the time and 16% above 4 m s–1. Wind gusts in 2002 were 62.4% above 4 m s–1 and 12.7% above 8 m s–1. In 2003, 63.4% of the wind gusts were above 4 m s–1 and 17.6% above 8 m s–1. All of these wind data were summarized between 3 and 10 am, the typical period of anthesis during the day.
Relative humidity during anthesis in both years was approximately 80 to 90% at 3am, 75 to 95% at 0600 h, then dropping to 35 to 70% by 1000 h. These humidity conditions are similar to those experienced in the bluegrass seed growing regions of Idaho, eastern Oregon, and eastern Washington.
Seedling Screening Results
Overall, very low levels of inter- or intraspecific hybridization were detected in this experiment during the 2 yr of the study. When seedlings from all entries and all the receptor plots are pooled together in both years, a hybrid frequency of 0.048% occurred (Table 2). Hybrids were not observed with P. fendleriana, also a female plant, likely because of temporal separation. Hybrid frequency for female plants of P. arachnifera was higher than the other species at 3.4%, while the other interspecific hybrid frequencies ranged from 0.0 to 0.196% (Table 2). Hybrids occurred with P. interior, P. pratensis x P. secunda, and three P. pratensis entries [‘Kenblue’, ‘Rutgers’, and P. pratensis subsp. angustifolia (L.) Dumort.]. All P. pratensis Coventry seedlings (susceptible control seedlings) died and all seedlings from BR99-1033 (resistant control seedlings) survived the glyphosate screening (Table 2).
|
The chloroplast genome of diploid P. trivialis and tetraploid P. annua are distinct from P. pratensis (Soreng, 1990; Gillespie and Boles, 2001; Patterson et al., 2005) and in examination of two nuclear genes CDO504 and TRX, only one P. pratensis TRX sequence showed any affinity to corresponding sequences of P. trivialis. This P. pratensis TRX sequence displayed significantly more similarity to corresponding P. secunda and P. arida sequences, than it did to P. trivialis (Patterson et al., 2005). Poa annua TRX sequences showed far less affinity with corresponding sequences of P. pratensis (Patterson et al., 2005).
Even among the P. pratensis receptor plants, gene flow was low. Highly apomictic Poa would be expected to experience very low gene flow from another parent, explaining the lack of hybrids from the apomictic ‘Fairfax’ and Coventry varieties. Apomixis in many Poa species may significantly reduce the occurrence of hybrids because of the reduced frequency of sexually produced progeny. The more sexual P. pratensis Rutgers line, Kenblue, and ‘Midnight’ produced hybrids but still at low frequency.
The pollen donor plants (BR99-1033) appeared highly apomictic on the basis of FISSR profiles and uniform morphological characteristics of all 500 plants in the field. Seedlings from BR99-1033 were not evaluated for morphological characteristics, but a very high frequency of apomixis is suggested since all BR99-1033 seedlings survived the herbicide screen (Table 2). Glyphosate resistance in this variety is conferred by a single gene insert which would be expected to be heterozygous and show segregation if genetic recombination had occurred.
Measured hybridization rates using our methods may have been lower than actual hybridization rates. The glyphosate resistant gene construct was inserted into one location in the pollen donor plant polyploid genome and may be heterozygous for the trait; therefore, producing pollen with and without the herbicide resistance gene via meiosis. Poa pratensis appears to have up to four, and possibly more, parental genomes based on work by Patterson et al. (2005). However, these data represent an accurate measurement of transgene movement through hybridization into other populations.
Hybrid Confirmation
The NOS and 35S DNA PCR amplification products were consistently detected and clearly visible in the BR99-1033 transgenic pollen donor parents and glyphosate-tolerant progeny (putative hybrids) obtained from nontransgenic receptor plants (Table 2). Faint amplification products have been observed in the negative controls, but these false-positive tests were eliminated in subsequent testing. One of the glyphosate tolerant P. arachnifera progeny died before DNA was sampled; however, the other two transgenic P. arachnifera progeny were weak and difficult to maintain in our greenhouse environment. Otherwise, these results confirmed the presence of transgenic DNA in all putative hybrids (Table 2).
With one noted exception, the maternally inherited chloroplast DNA ndhF and/or rps16-trnK sequences of all putative (transgenic) interspecific hybrids analyzed were identical to the nontransgenic maternal receptor genotypes (Table 2) and different from the paternal BR99-1033 genotype. The chloroplast DNA of one glyphosate-tolerant transgenic seedling, allegedly grown from P. palustris PI387934 seed, contained the chloroplast ndhF-1 and rps16-trnK-3 alleles characteristic of the transgenic BR99-1033 genotype. However, the sequences of the seedling were different from the P. palustris receptor sequences. The chloroplast DNA of P. palustris is quite clearly different from P. pratensis (Patterson et al., 2005); thus, we concluded this transgenic seedling was a contaminant from BR99-1033 seed. The chloroplast DNA of two transgenic P. arachnifera offspring contained the AY589107 ndhF and DQ389141 rps16-trnK sequences, which are unique to the P. arachnifera receptor plants (i.e., different from all other Poa species tested). As mentioned above, the third transgenic P. arachnifera offspring died before DNA was collected. The chloroplast DNA of all seven transgenic P. interior offspring contained the ndhF-2 allele shared only by P. interior, P. palustris, and P. nemoralis and the DQ389137 rps16-trnK sequence found only in the P. interior accession (i.e., different from all other Poa species tested). The chloroplast DNA sequences of 21 transgenic P. secunda x P. pratensis PI 578808 offspring and 11 transgenic P. secunda x P. pratensis PI 578818 offspring contained the ndhF-2 allele shared only by the respective maternal receptor parents and the P. secunda PI 578851 genotypes. The PCR or sequencing reactions failed for six transgenic P. secunda x P. pratensis PI 578808 offspring and one transgenic P. secunda x P. pratensis PI 578818 offspring, but otherwise all putative interspecific hybrids were confirmed (Table 2). These data also confirm the maternal P. secunda lineage of the P. secunda x P. pratensis hybrid accessions (PI 578808 and PI 578818) and distinguish these genotypes from several other P. secunda sequences (AY589111 and AY589112).
All P. pratensis varieties tested shared the same chloroplast ndhF-1 DNA sequence. Thus, the ndhF marker could not distinguish the paternal (BR99-1033) and maternal parents of the intraspecific hybrids. The marker does, however, effectively confirm the maternal P. pratensis identity of the intraspecific hybrids (i.e., they are not interspecific hybrid seed contaminations). The chloroplast rps16-trnK DNA sequences from most of the 112 intraspecific hybrids (P. pratensis) were identical to the nontransgenic maternal receptor accessions but different from the transgenic BR99-1033 genotype (Table 2).
The experimental BR99-1033 transgenic variety showed relatively uniform appearance, uniform DNA profiles (results not shown), and did not segregate for glyphosate tolerance. Thus, we deduce that it is fixed for the chloroplast rps16-trnK-3 allele, which was different from all other bluegrasses analyzed except one transgenic Midnight offspring, one transgenic Fairfax offspring, and the Fairfax receptor genotype. The chloroplast rps16-trnK DNA sequences from nine of the other 12 transgenic Midnight offspring contained the rps16-trnK-2 allele characteristic of the Midnight receptor plants. The other two transgenic offspring of Midnight carry the rps16-trnK-1 allele characteristic of the Rutgers receptor genotype. Thus, rps16-trn sequences from 11 of the 12 transgenic Midnight offspring are different from BR99-1033 and most contain the rps16-trnK-2 allele characteristic of the Midnight receptor plants. Likewise, the chloroplast rps16-trnK DNA sequences from 79 of the 80 transgenic Rutgers offspring contain the rps16-trnK-1 allele of the Rutgers receptor genotype. The chloroplast rps16-trnK DNA sequence from one other transgenic Rutgers offspring contained the rps16-trnK-2 allele present in two Midnight receptor genotypes and most of the transgenic Midnight offspring. The chloroplast rps16-trnK DNA sequences from 14 of the 16 transgenic Kenblue offspring and five of the seven transgenic P. pratensis subsp. angustifolia PI 317504 offspring contain the rps16-trnK-4 allele, which is shared only by Kenblue and P. pratensis subsp. angustifolia PI 317504 receptor genotypes. As for the other two transgenic P. pratensis subsp. angustifolia PI 317504 offspring we detected one unique rps16-trnK sequence (DQ389140), albeit similar to other P. pratensis sequences, but not found in any other plant. We did not successfully sequence the other putative hybrid.
With one exception (P. palustris), we effectively confirmed most of the putative hybrids and have reasonable justification to assume correct identity for unconfirmed hybrids. With the single exception of Fairfax, which was not distinguishable from BR99-1033, most putative hybrids within each receptor genotype (Table 2) were properly confirmed.
Distance and Direction of Gene Flow
In 2002, we observed the highest level of gene flow to other Poa entries in the center plot (0-m distance). The number of hybrids detected decreased at the 13-m distance and far fewer at 53 m (Table 3). Four of the six hybrids observed at the 53-m distance were in the NE direction from one plant. Only one hybrid was detected at 53 m from a vector other than to the NE. (Table 3). Results in 2003 were similar, but gene flow was much lower (Table 4). The numbers of seedlings evaluated were decreased because of billbug insect (Sphenophorous spp.) damage on many of the species in late summer 2002 and poor seed set conditions due to hot, dry weather during anthesis in 2003. Hybridization rates in the center plot entries (0 m) was 0.38% for P. pratensis and two hybrids out of 15 seedlings (13.3%) from P. arachnifera. More seedlings were obtained from P. arachnifera in 2003, as the plants were significantly larger because of an additional year growth. Overall, hybridization at 13 and 53 m in 2003 was again lower than in the center plot (Table 4).
|
|
Because so few hybrids were detected, we were unable to effectively model gene flow over distance and direction. When compared with the models used most effectively in other gene flow reports (Wipff and Fricker, 2001; Cunliffe et al., 2004), the fit of our data was poor, with an r2 value of 0.22 in an exponential decay model. Although pollen can move long distances, even hundreds of miles, the majority of gene flow occurs over very short distances (Gleaves, 1973). In Festuca pratensis Huds., a predominantly out crossing species, the ability of intraspecific hybrids to occur at distances from a pollen source were heavily dependent on the density of the potential receptor plants (Rognli et al., 2000). Isolated plants are more likely to hybridize with distant pollen sources than communities of plants (Gleaves, 1973; Rognli et al., 2000). Increasing the number of plants between the pollen source and receptor plots in our experiment may have resulted in more hybrids produced and possibly better gene flow predictions, but the additional plants could further reduce the fit to a exponential decay model through greater pollen competition and pollen flow variability.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication February 13, 2006.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
.