Apomictic Grass Research Articles

Genetic Transformation of Apomictic Grasses: Progress and Constraints.

The available methods for plant transformation and expansion beyond its limits remain especially critical for crop improvement. For grass species, this is even more critical, mainly due to drawbacks in in vitro regeneration. Despite the existence of many protocols in grasses to achieve genetic transformation through Agrobacterium or biolistic gene delivery, their efficiencies are genotype-dependent and still very low due to the recalcitrance of these species to in vitro regeneration. Many plant transformation facilities for cereals and other important crops may be found around the world in universities and enterprises, but this is not the case for apomictic species, many of which are C4 grasses. Moreover, apomixis (asexual reproduction by seeds) represents an additional constraint for breeding. However, the transformation of an apomictic clone is an attractive strategy, as the transgene is immediately fixed in a highly adapted genetic background, capable of large-scale clonal propagation. With the exception of some species like Brachiaria brizantha which is planted in approximately 100 M ha in Brazil, apomixis is almost non-present in economically important crops. However, as it is sometimes present in their wild relatives, the main goal is to transfer this trait to crops to fix heterosis. Until now this has been a difficult task, mainly because many aspects of apomixis are unknown. Over the last few years, many candidate genes have been identified and attempts have been made to characterize them functionally in Arabidopsis and rice. However, functional analysis in true apomictic species lags far behind, mainly due to the complexity of its genomes, of the trait itself, and the lack of efficient genetic transformation protocols. In this study, we review the current status of the in vitro culture and genetic transformation methods focusing on apomictic grasses, and the prospects for the application of new tools assayed in other related species, with two aims: to pave the way for discovering the molecular pathways involved in apomixis and to develop new capacities for breeding purposes because many of these grasses are important forage or biofuel resources.

Differential Methylation Patterns in Apomictic vs. Sexual Genotypes of the Diplosporous Grass Eragrostis curvula.

DNA methylation is an epigenetic mechanism by which a methyl group is added to a cytosine or an adenine. When located in a gene/regulatory sequence it may repress or de-repress genes, depending on the context and species. Eragrostis curvula is an apomictic grass in which facultative genotypes increases the frequency of sexual pistils triggered by epigenetic mechanisms. The aim of the present study was to look for correlations between the reproductive mode and specific methylated genes or genomic regions. To do so, plants with contrasting reproductive modes were investigated through MCSeEd (Methylation Context Sensitive Enzyme ddRad) showing higher levels of DNA methylation in apomictic genotypes. Moreover, an increased proportion of differentially methylated positions over the regulatory regions were observed, suggesting its possible role in regulation of gene expression. Interestingly, the methylation pathway was also found to be self-regulated since two of the main genes (ROS1 and ROS4), involved in de-methylation, were found differentially methylated between genotypes with different reproductive behavior. Moreover, this work allowed us to detect several genes regulated by methylation that were previously found as differentially expressed in the comparisons between apomictic and sexual genotypes, linking DNA methylation to differences in reproductive mode.

Recurrent phenotypic selection and recurrent selection based on combining ability in tetraploid bahiagrass

AbstractPaspalum notatum Flüggé is an apomictic grass used for forage and turf. The recent generation of a sexual synthetic tetraploid population of the species provides the unique advantage to exploit heterosis by accumulating favorable alleles through recurrent selection cycles. The objective was to evaluate recurrent phenotypic selection (RPS) and recurrent selection based on combining ability (RSCA) as breeding methods. The best sexual genotypes in terms of warm‐season growth from a sexual synthetic tetraploid population were selected by RPS and RSCA. The selected genotypes were polycrossed and two new sexual populations were created, which were crossed with superior apomictic genotypes. Both methods allowed obtaining families that were evaluated for summer, fall, and spring growth. Breeding values and heterosis were also calculated. Summer growth for RPS families was greater than for RSCA families (33.1 and 29.08 g plant−1, respectively); however, fall, spring, and accumulated growth were similar. The breeding value for RPS sexual parents was greater than for RSCA sexual parents for summer growth, whereas those for fall and spring growth were similar. A greater level of families with heterosis (80%) was observed for RPS than for RSCA (30%) for summer growth, whereas the levels for fall and spring growth were similar. These results indicate that both breeding methods were successful in improving tetraploid P. notatum. They increased the mean value of the progenies and produced families superior for forage yield by accumulating favorable alleles.

Temporal and spatial expression of genes involved in DNA methylation during reproductive development of sexual and apomictic Eragrostis curvula

Recent reports in model plant species have highlighted a role for DNA methylation pathways in the regulation of the somatic-to-reproductive transition in the ovule, suggesting that apomixis (asexual reproduction through seeds) likely relies on RdDM downregulation. Our aim was therefore to explore this hypothesis by characterizing genes involved in DNA methylation in the apomictic grass Eragrostis curvula. We explored floral transcriptomes to identify homologs of three candidate genes, for which mutations in Arabidopsis and maize mimic apomixis (AtAGO9/ZmAGO104, AtCMT3/ZmDMT102/ZmDMT105, and AtDDM1/ZmCHR106), and compared both their spatial and temporal expression patterns during reproduction in sexual and apomictic genotypes. Quantitative expression analyses revealed contrasting expression patterns for the three genes in apomictic vs sexual plants. In situ hybridization corroborated these results for two candidates, EcAGO104 and EcDMT102, and revealed an unexpected ectopic pattern for the AGO gene during germ line differentiation in apomicts. Although our data partially support previous results obtained in sexual plant models, they suggest that rather than an RdDM breakdown in the ovule, altered localization of AtAGO9/ZmAGO104 expression is required for achieving diplospory in E. curvula. The differences in the RdDM machinery acquired during plant evolution might have promoted the emergence of the numerous apomictic paths observed in plants.

Marker-assisted screening of breeding populations of an apomictic grass Cenchrus ciliaris L. segregating for the mode of reproduction

Abstract Cenchrus ciliaris L. is an apomictic forage grass grown in pastures and rangelands of the semi-arid tropics. It reproduces predominantly through apomixis; rarely, obligate sexual plants have also been reported. Absence of sexual reproduction limits the possibility of genetic improvement through hybridization. This study reports on hybridization of an obligate sexual, self-incompatible buffelgrass with pollen from apomictic plants towards development of an F2 population segregating for mode of reproduction and use of sequence characterized amplified region (SCAR) markers for screening the population. The segregation ratio of 3:1 (facultative: apomictic) was observed in the F1 generation, whereas it was 1:2:1 (apomictic: facultative: sexual) in the F2 generation. A number of obligate sexual F2 progenies with desirable agronomic traits were obtained. The SCAR markers were able to screen out apomictic plants from sexual ones, but failed to discriminate between facultative and sexual. Marker-assisted screening could be useful for introgression of desirable trait(s) in the apomictic genotype through hybridization.

Haploid embryo production in rice and maize induced by PsASGR-BBML transgenes.

The PsASGR - BBML transgene, derived from a wild apomictic grass species, can induce parthenogenesis, embryo formation without fertilization, in rice and maize, leading to the formation of haploid plants. The ability to engineer apomictic crop plants using genes identified from naturally occurring apomicts will depend on the ability of those genes to function in crop plants. The PsASGR-BBML transgene, derived from the apomictic species Pennisetum squamulatum, promotes parthenogenesis in sexual pearl millet, a member of the same genus, leading to the formation of haploid embryos. This study determined that the PsASGR-BBML transgene can induce haploid embryo development in two major monocot crops, maize and rice. Transgene variations tested included two different promoters and the use of both genomic and cDNA PsASGR-BBML-derived sequences. Haploid plants were recovered from mature caryopses (seed) of rice and maize lines at variable rates. The PsASGR-BBML transgenes failed to induce measurable haploid seed development in the model genetic plant system Arabidopsis thaliana. Complexity of embryo development, as documented in transgenic rice lines, identifies the need for further characterization of the PsASGR-BBML gene.

The LOSS OF APOMEIOSIS (LOA) locus in Hieracium praealtum can function independently of the associated large‐scale repetitive chromosomal structure

Apomixis or asexual seed formation in Hieracium praealtum (Asteraceae) is controlled by two independent dominant loci. One of these, the LOSS OF APOMEIOSIS (LOA) locus, controls apomixis initiation, mitotic embryo sac formation (apospory) and suppression of the sexual pathway. The LOA locus is found near the end of a hemizygous chromosome surrounded by extensive repeats extending along the chromosome arm. Similar apomixis-carrying chromosome structures have been found in some apomictic grasses, suggesting that the extensive repetitive sequences may be functionally relevant to apomixis. Fluorescence in situ hybridization (FISH) was used to examine chromosomes of apomeiosis deletion mutants and rare recombinants in the critical LOA region arising from a cross between sexual Hieracium pilosella and apomictic H.praealtum. The combined analyses of aposporous and nonaposporous recombinant progeny and chromosomal karyotypes were used to determine that the functional LOA locus can be genetically separated from the very extensive repeat regions found on the LOA-carrying chromosome. The large-scale repetitive sequences associated with the LOA locus in H.praealtum are not essential for apospory or suppression of sexual megasporogenesis (female meiosis).

Mulato II (Brachiaria sp.)

Mulato II is a semi-erect perennial apomictic grass that can grow up to 9 ft tall. Brachiariagrasses, including Mulato II, are tropical warm season forages native to Africa and are the most widely grown forages in tropical South America. This revised 4-page fact sheet describes the morphology, region of adaptation and growing season in Florida, recommended production practices, and the results of research into the performance of heifers grazing Mulato. Written by J. Vendramini, B. Sellers, L.E. Sollenberger, and M. Silveira and published by the UF Department of Agronomy, March 2011. SS AGR 303/AG310: Mulato II (Brachiaria sp.) (ufl.edu)

Apomixis and ploidy barrier suppress pollen-mediated gene flow in field grown transgenic turf and forage grass (Paspalum notatum Flüggé)

Bahiagrass (Paspalum notatum Flüggé) is the predominant forage grass in the southeastern US. The commercially important bahiagrass cultivar 'Argentine' is preferred for genetic transformation over sexual diploid cytotypes, since it produces uniform seed progeny through apomixis. Pseudogamous apomictic seed production in Argentine bahiagrass may contribute to transgene confinement. It is characterized by embryo development which is independent of fertilization of the egg cell, but requires fertilization with compatible pollen to produce the endosperm. Pollen-mediated gene transfer from transgenic, glufosinate-resistant apomictic bahiagrass as pollen donor at close proximity (0.5-3.5 m) with non-transgenic sexual or apomictic bahiagrass cultivars as pollen receptors was evaluated under field conditions. Hybridization frequency was evaluated by glufosinate herbicide resistance in >23,300 seedlings derived from open-pollinated (OP) pollen receptor plants. Average gene transfer between transgenic apomictic, tetraploid and sexual diploid bahiagrass was 0.03%. Herbicide-resistant hybrids confirmed by immuno-chromatographic detection of the PAT protein displayed a single copy bar gene identical to the pollen parent. Hybrids resulting from diploid pollen receptors were confirmed as triploids or aneu-triploids with significantly reduced vigor and seed set as compared to the parents. Transmission of transgenes to sexual bahiagrass is severely restricted by the ploidy difference between tetraploid apomicts and diploid sexual bahiagrass. Average gene transfer between transgenic apomictic tetraploid and non-transgenic, apomictic tetraploid bahiagrass was 0.17%, confirming a very low frequency of amphimixis in apomictic bahiagrass cultivars. While not providing complete transgene containment, gene transfer between transgenic apomictic and non-transgenic bahiagrass occurs at a much lower frequency than reported for other cross-pollinating or facultative apomictic grasses.

The effect of genotype and culture medium on somatic embryogenesis and plant regeneration from mature embryos of fourteen apomictic cultivars of buffel grass (Cenchrus ciliaris L.)

AbstractBuffel grass (Cenchrus ciliaris L.) is an important apomictic grass used as forage for ruminant livestock. Biotechnological methods provide opportunities for producing new germplasm. Mature embryos of fourteen buffel grass apomictic cultivars (2n = 4x = 36) were used to induce embryogenic callus formation using a basal medium supplemented with 3% sucrose and with the testing of five concentrations of 2,4‐dichlorophenoxyacetic acid (2,4‐D) and four concentrations of 6‐benzylaminopurine (BAP). The effects of cultivar and culture medium on callus induction and plant regeneration were evaluated. Significant differences were observed among the fourteen cultivars and the five concentrations of 2,4‐D (P < 0·01). Values for embryogenic callus production varied from 0 to 86·7. Most cultivars showed the highest level of embryogenic callus production on the medium with the concentration of 3 mg L−1 2,4‐D. The addition of different BAP concentrations in combination with 2,4‐D in the medium inhibited embryogenic callus growth and did not permit plant regeneration. The data clearly demonstrated that the genotype and concentrations of 2,4‐D had significant effects both on the frequency of embryonic callus formation from mature embryos and on the subsequent efficiency of plant regeneration of apomictic cultivars of buffel grass. Cultivars Biloela and Nunbank showed the greatest efficiency in in vitro culture response.

Polymorphism and Genotype-specific Markers for Dichanthium Identified by Random Amplified Polymorphic DNA

One hundred decamer arbitrary primers were tested for PCR based amplification of seven genotypes (IG2208-S-1, IG2177, IG2180, IG2178, IG2165-S-1-1, IG2165-1 and Local-1) of an apomictic grass, Dichanthium annulatum, with the aim of screening polymorphic primers and genotype-specific markers. Out of 100 decamer primers tested, 42 produced no amplification or smeared non scorable bands, 12 amplified only single band and 46 yielded more than one polymorphic bands. Thirty-two primers out of 46 selected showed high level of polymorphism, producing 3–15 reproducible bands each for the seven Dichanthium genotypes examined. Among the total of 307 amplified fragments 222 were polymorphic, 53 bands were unique to the genotypes and 32 were monomorphic. Thus, with selected primers sufficient polymorphism could be detected to allow identification of individual genotypes. Genetic similarities of RAPD profiles generated were estimated via a coefficient of DICE and then the data were processed by cluster analysis (UPGMA). The maximum similarities between two genotypes (IG2180 and IG2178) was 58% and these two made a cluster with genotype IG2177 having similarity of only 54%. It clearly corroborated existence of high levels of polymorphism in this grass though being apomictic in nature. Primers like OPE-16, OPG-02, OPG-18, OPH-05, OPH-09, OPH-16, OPI-07 and OPF-06 found most informative as they produced specific bands pertaining to five out of seven genotypes. Polymerase chain reaction (PCR) offers a substantially simple, rapid and reliable method for identification of large number of Dichanthium genotypes once enough number of reproducible and suitable primers is screened.

Identification of differentially expressed cDNA sequences in ovaries of sexual and apomictic plants of Brachiaria brizantha.

The isolation of genes associated with apomixis would improve understanding of the molecular mechanism of this mode of reproduction in plants as well as open the possibility of transfer of apomixis to sexual plants, enabling cloning of crops through seeds. Brachiaria brizantha is a highly apomictic grass species with 274 tetraploid apomicts accessions and only one diploid sexual. In this study we have compared gene expression in ovaries at megasporogenesis and megagametogenesis of sexual and apomictic accessions of B. brizantha by differential display (DD-PCR), with 60 primer combinations. Specificity of 65 cloned fragments, checked by reverse northern blot analysis, showed that 11 clones were differentially expressed, 6 in apomictic ovaries, 2 in sexual and 3 in apomictic and sexual, but at different stages. Of the 6 sequences isolated that were preferentially expressed in the apomictic accession: one sequence was from ovaries at megasporogenesis stage; three were from megagametogenesis stage; two were from both stages. Of the two sequences isolated from the sexual accessions, one showed expression in ovaries at megagametogenesis, while the other sequence was shown to be specific to both stages. Three sequences were from megasporogenesis stage in apomicts but were also detected at megagametogenesis in sexual plants. Sequence analysis showed that 5 of the 11 clones had no apparent homologues in the protein database. Some of the clones identified as apomictic-specific shared homology with known genes enabling their functional annotation. The relationships of these functions to the generation of the apomictic trait are discussed.

Construction of BAC libraries from two apomictic grasses to study the microcolinearity of their apospory-specific genomic regions.

We have constructed bacterial artificial chromosome (BAC) libraries from two grass species that reproduce by apospory, a form of gametophytic apomixis. The library of an apomictic polyhaploid genotype (line MS228-20, with a 2C genome size of approximately 4,500 Mbp) derived from a cross between the obligate apomict, Pennisetum squamulatum, and pearl millet ( P. glaucum) comprises 118,272 clones with an average insert size of 82 kb. The library of buffelgrass ( Cenchrus ciliaris, apomictic line B-12-9, with a 2C genome size of approximately 3,000 Mbp) contains 68,736 clones with an average insert size of 109 kb. Based on the genome sizes of these two lines and correcting for the number for false-positive and organellar clones, library coverages were found to be 3.7 and 4.8 haploid genome equivalents for MS 228-20 and B12-9, respectively. Both libraries were screened by hybridization with six SCARs (sequence-characterized amplified regions), whose tight linkage in a single apospory-specific genomic region had been previously demonstrated in both species. Analysis of these BAC clones indicated that some of the SCAR markers are actually amplifying duplicated regions linked in coupling in both genomes and that restriction enzyme mapping will be necessary to sort out the duplications.

An endophyte of the tropical forage grass Brachiaria brizantha: isolating, identifying, and characterizing the fungus, and determining its antimycotic properties.

Brachiaria, predominantly an African genus, contains species, such as B. brizantha, an apomictic C4 grass, that are commercially important forage grasses in tropical America, where they now cover about 55 million hectares. From B. brizantha accession CIAT 6780, we isolated an endophytic fungus that may be economically significant. The fungus was identified as Acremonium implicatum (J. Gilman & E.V. Abott). 18S rDNA and ITS rDNA sequences were used to characterize isolates of the endophyte, and showed that they belonged to the Acremonium genus, being close to A. strictum and A. kiliense. Using the random amplified polymorphic DNA (RAPD) technique, involving arbitrary primers of 10 bases, we showed that the isolates were highly similar to each other. Antiserum produced from a monoconidial culture of A. implicatum isolated from B. brizantha 6780, differentiated the isolates consistently in line with the DNA data. When we compared endophyte-free with endophyte-infected B. brizantha CIAT 6780 plants, both artificially inoculated with the pathogenic Drechslera fungus, we found that the endophyte-infected plants had fewer and smaller lesions than did the endophyte-free plants. Sporulation of Drechslera sp. on artificially inoculated leaf sheath tissues was also much less on tissue infected with the endophyte.

Genetic variability and phytogeography ofMiscanthus sinensisvar.condensatus, an apomictic grass, based on RAPD fingerprints

DNA fingerprinting using random amplified polymorphic DNA (RAPD) markers was employed to investigate the genetic variation within and among populations of Miscanthus Anderss. sinensis var. condensatus (Hack.) Makino, an apomictic grass distributed along the coasts of Taiwan and Ryukyu Islands. A total of 250 plants from three Taiwanese populations (Southeast Coast, Orchid Islet, and Green Islet) and two populations from Ryukyu (Ishigaki and Amami-O-Shima Islets) were sampled. The amplified products of 40 random primers showed monomorphic banding patterns within all populations as well as among the three populations from Taiwan. Low genetic variation (with only two polymorphic loci), but significant differentiation, was detected between populations from Taiwan and Ryukyu (ΦCT= 0.864) and between populations (ΦST= 1.0) from Ishigaki and Amami-O-Shima Islets. In contrast, a high level of variation was exhibited in the outcrossing Miscanthus sinensis var. glaber (Nakai) Li. In addition to apomictic reproduction, low genetic variation across populations of M. sinensis var. condensatus may be a result of high salinity acting as a selective agent. With the cost of reduced genetic heterogeneity, apomixis may have provided a mechanism for avoiding the transmission of endophytic fungi. The phytogeographic pattern of M. sinensis var. condensatus, as reflected by the RAPD data, likely represents isolation between Taiwan and Ryukyu since the mid-Pleistocene.Key words: apomixis, Miscanthus sinensis var. condensatus, phytogeography, population differentiation, RAPD, system of mating.

Genetic variability and phytogeography of <i>Miscanthus sinensis</i> var. <i>condensatus</i>, an apomictic grass, based on RAPD fingerprints

DNA fingerprinting using random amplified polymorphic DNA (RAPD) markers was employed to investigate the genetic variation within and among populations of Miscanthus Anderss. sinensis var. condensatus (Hack.) Makino, an apomictic grass distributed along the coasts of Taiwan and Ryukyu Islands. A total of 250 plants from three Tai- wanese populations (Southeast Coast, Orchid Islet, and Green Islet) and two populations from Ryukyu (Ishigaki and Amami-O-Shima Islets) were sampled. The amplified products of 40 random primers showed monomorphic banding patterns within all populations as well as among the three populations from Taiwan. Low genetic variation (with only two polymorphic loci), but significant differentiation, was detected between populations from Taiwan and Ryukyu (ΦCT = 0.864) and between populations (ΦST = 1.0) from Ishigaki and Amami-O-Shima Islets. In contrast, a high level of variation was exhibited in the outcrossing Miscanthus sinensis var. glaber (Nakai) Li. In addition to apomictic repro- duction, low genetic variation across populations of M. sinensis var. condensatus may be a result of high salinity acting as a selective agent. With the cost of reduced genetic heterogeneity, apomixis may have provided a mechanism for avoiding the transmission of endophytic fungi. The phytogeographic pattern of M. sinensis var. condensatus ,a s re- flected by the RAPD data, likely represents isolation between Taiwan and Ryukyu since the mid-Pleistocene.

Genetic variation in an apomictic grass, heteropogon contortus, in the hawaiian islands

Random amplified polymorphic DNA (RAPD) markers were used to assess genetic variation within and among Hawaiian populations of an apomictic grass, Heterogopon contortus (pili grass). From among 56 individuals sampled from six populations on O'Ahu and Hawai'i, 55 unique genotypes were detected using 33 polymorphic markers. This lack of uniformity among individuals may indicate frequent sexual reproduction in these populations. Analysis of molecular variance (AMOVA) revealed significant variation among populations (30.2%), but higher levels of variation within populations (68.1%). Cluster analysis revealed a high degree of clustering for most populations, but populations from different islands did not cluster together. The presence of among-population differentiation but lack of between-island differentiation may suggest that H. contortus was an early Polynesian introduction to the Hawaiian Islands.

Short Communication: An apospory-specific genomic region is conserved between Buffelgrass (Cenchrus ciliaris L.) and Pennisetum squamulatum Fresen.

Twelve molecular markers linked to pseudogamous apospory, a form of gametophytic apomixis, were previously isolated from Pennisetum squamulatum Fresen. No recombination between these markers was found in a segregating population of 397 individuals (Ozias-Akins et al. 1998, Proc. Natl Acad. Sci. USA, 95, 5127-5132). The objective of the present study was to test if these markers were also linked to the aposporous mode of reproduction in two small segregating populations of Cenchrus ciliaris (= Pennisetum ciliare (L.)Link), another apomictic grass species. Among 12 markers (sequence characterized amplified regions, SCARs), six were scored as dominant markers between aposporous and sexual C. ciliaris genotypes (presence/absence, respectively). Five were always linked to apospory and one showed a low level of recombination in 84 progenies. Restriction fragment length polymorphisms (RFLPs) were observed between sexual and apomictic phenotypes for three of the six remaining SCARs from P. squamulatum when used as probes. No recombination was observed in the F1 progenies. Preliminary data from megabase DNA analysis and sequencing in both species indicate that an apospory-specific genomic region (ASGR) is highly conserved between the two species. Although C. ciliaris has a smaller genome size to P. squamulatum, a higher copy number for markers linked to apospory found in the former may impair the progress of positional cloning of gene(s) for apomixis in this species.

Evolution and adaptedness in a facultatively apomictic grass, Poa pratensis L.

Evolution and adaptedness in a facultatively apomictic grass, Poa pratensis L.

Quantitative analysis of aposporous parthenogenesis inPoa pratensisgenotypes

SUMMARY Several embryological and cytological techniques were developed to determine the degree of aposporous parthenogenesis vs. sexuality in different genotypes of the facultative apomictic grass Poa pratensis L. Fixed embryo sacs were mechanically isolated and individually examined with phase-contrast and interference-contrast light microscopes. The degree of apospory was calculated as the ratio of embryo sacs with an embryo but without endosperm and the total number of embryo sacs studied per plant. The degree of apospory varied between 9 and 76% in nine plants. Photocytometry and flow cytometry were used to determine ploidy levels in embryo and endosperm of individual embryo sacs in order to determine their origin, asposporous vs. sexual. Results based on the calculated degree of apospory agreed with the degree of apomixis deduced from homogeneity scores in field progeny tests. Callose deposition on the megaspore mother cell during megasporogenesis was studied in the same set of genotypes and the percentage of abnormal, micropylar callose deposition closely agreed with results both of the embryo sac analysis and the progeny tests. Combined use of these techniques yields quantitative data on the aposporous tendency of individual genotypes of Poa pratensis.