Journal of Heredity Advance Access originally published online on September 19, 2006
Journal of Heredity 2006 97(5):451-455; doi:10.1093/jhered/esl022
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Mating Patterns of Black Oak Quercus velutina (Fagaceae) in a Missouri Oak-Hickory Forest
From the Laboratoire d'Ecologie, Systématique et Evolution, Bât 360, Université Paris 11 Sud, 91405 Orsay Cedex, France (Fernández-Manjarrés); the National Human Genome Research Institute, National Institute of Health, Bethesda, MD 20892 (Idol); and the Department of Ecology and Evolutionary Biology, The University of California at Los Angeles, CA 90095-1606 (Sork)
Address correspondence to J. F. Fernández-Manjarrés at the address above, or e-mail: juan.fernandez{at}ese.u-psud.fr.
Wind-pollinated forest trees usually have high outcrossing rates, but allogamy does not necessarily translate into high pollen movement. The goal of this study was to determine the outcrossing rates, pollen pool genetic structure, and the size of the effective pollination neighborhood in a population of black oak, Quercus velutina, in a Missouri oak-hickory forest. Based on 6 allozyme loci, 12 maternal trees, and 439 progenies sampled along a transect of 1300 m, we found complete outcrossing (tm = 1.000, P < 0.001) and small amounts of biparental inbreeding. Using a TwoGener analysis of the pollen gene pool, we found significant structure across maternal plants (
FT = 0.078, P < 0.001), which when corrected for adult inbreeding translates into FT = 0.066 that corresponds to an effective number of pollen donors of 7.5 individuals. Assuming a bivariate normal distribution and an adult density of 16.25 trees ha–1, we estimated that the effective pollination neighborhood area had a radius of 41.9 m. Even assuming that our estimates may be conservative, these findings join a growing body of evidence that suggest that the local neighborhood of wind-pollinated forest tree populations may be relatively small creating opportunities for local selection and genetic drift.
Wind-pollinated trees typically have high levels of genetic variation within populations, low levels of differentiation among populations, and high outcrossing rates, which are all traits that indicate high levels of gene flow (Hamrick and Godt 1996). Despite these patterns, if the movement of a large portion of pollen is local, these same populations could be genetically subdivided creating local genetic structure (Smouse and Sork 2004). Recent evidence in oaks suggests that local pollen flow occurs within moderate distances, originating from neighboring trees, including relatives. For example, an analysis of pollen pool structure in forest populations of Quercus alba from the Missouri Ozarks indicates that the effective mean pollen dispersal area could have a radius as small as 16 m (Smouse et al. 2001). Results in Pinus echinata indicate high pollen pool structure and localized pollen flow (Dyer and Sork 2001), possible due to vegetative density (RJ Dyer and VL Sork, in preparation). At the same time, oaks in sparse landscapes have also shown both significant pollen pool structure despite high outcrossing rates (Sork et al. 2002; Fernández-Manjarrés and Sork 2005). Thus, locally heterogeneous pollen pools are not simply attributable to high forest cover.
Here, we examine the mating patterns of Quercus velutina Lam. in a secondary oak-hickory forest in southern Missouri, USA, using the mixed mating system model as implemented in the software MLTR (Ritland 2002) and the pollen pool structure model, TwoGener (Austerlitz and Smouse 2001; Smouse et al. 2001). MLTR is a widely used method of estimating the outcrossing rates of populations, but increasingly, we are seeing that tree species often show such high outcrossing rates that this parameter alone is not an informative indicator of gene flow patterns. The TwoGener model provides additional information about effective pollen movement contributing to the genetic neighborhood (sensu Wright 1943), without the need of high genetic resolution (Smouse and Sork 2004). In this paper, we present a case study of pollen movement for a forest oak population based on a previously generated data set (Idol 1991) to determine the size of the pollination neighborhood. We asked the following questions: 1) What is the population outcrossing rate? 2) Does the pollen pool have significant genetic structure? 3) What are the effective number of pollen donors and the size of the effective pollination neighborhood?
 |
Materials and Methods |
|---|
 Top Materials and Methods Results and DiscussionReferences |
|---|
Study Species
Quercus velutina Lam. (Fagaceae) is a common tree found in eastern and central North America growing on dry, sandy, or rocky ridges and upper slopes (Fowells 1965). It is commonly associated with hickories (Carya spp.) and many other oak species, such as post oak (Quercus stellata Wangenh.), scarlet oak (Quercus coccinea Muenchh.), southern red oak (Quercus falcata Michx.), blackjack oak (Quercus marilandica Muenchh.), chestnut oak (Quercus prinus L.), white oak (Q. alba L.), and northern red oak (Quercus rubra L.). Acorns mature in 2 years, and seeds need a light covering of leaves for germination. Seedlings cannot survive under a dense understory.
Study Site
The study site was located in an oak-hickory forest in Tyson Research Center, an ecological reserve owned and managed by Washington University of St Louis. This 2000-acre reserve is located in Eureka, St Louis Co., MO, at the northeastern end of the Ozark Plateau (38°31'N, 90°33'W). Oak-hickory forest covers approximately 75% of the research area. Three dominant species of oak inhabit this area: Q. rubra, Q. alba, and Q. velutina. Hence, black oak is common in the reserve but is not the dominant species. The largest trees in the area range in age from 120–160 years and show spatial aggregation (Hampe 1984) as shown in Figure 1. Study trees are located within continuous forest along a ridge top accessed by a dirt road, which approximates a transect of 1.3 km. Study trees were separated by distances ranging from 10 to 1320 m.
|
Sampling
During the fall of 1987, acorns from 12 adult black oak trees were collected and transported to the laboratory and float tested for viability. Viable acorns were placed in cold storage (4 °C) for at least 90 days for stratification. After 90 days, at least 40 acorns per tree were planted in a greenhouse for germination. A total of 439 acorns germinated and grew to the seedling stage at which point they were sampled. Final per tree array sizes were 20, 36, 37, 41, 40, 37, 38, 35, 52, 32, and 37, respectively. The year 1987 was the second largest masting year relative to a 12-year study of crop production (VL Sork, unpublished data).
Genetic Markers
Leaf tissue samples were prepared for electrophoresis by crushing to a dry powder with liquid nitrogen in a mortar and pestle. Powdered leaf tissue was ground using a phosphate grinding buffer (Mitton et al. 1977) with the addition of 10% PVP-40 (Manos and Fairbrothers 1987). Starch gels were run following standard methods (Gottlieb 1981). Buffer system 1 from Soltis et al. (1983) was used to detect shikimate dehydrogenase (Sdh, EC 1.1.1.44
[EC]
), phosphoglucomutase (Pgm, EC 5.4.2.2
[EC]
), and malate dehydrogenase (Mdh, EC 1.1.1.37
[EC]
). Buffer system 6 was used for peroxidase (Per, EC 1.11.1.7
[EC]
), phosphoglucoisomerase (Pgi, EC 5.3.1.9
[EC]
), and diaphorase (Dia, EC 1.6.99). For detailed description of laboratory procedures, see Sork et al. (1993).
Genetic Analyses
Population and family estimates of single locus (ts) and multilocus (tm) outcrossing rates were obtained using maximum likelihood procedures (Ritland 2002). The algorithm used in the estimation of the mating system parameters was the expectation maximization routine that bounds outcrossing rates between 0 and 1. The program was seeded with initial values of outcrossing rate t = 0.90, parental inbreeding F = 0.1, and paternity correlation rp = 0.1. Paternity correlation is based on the sibling pair model (Ritland 1989) that estimates the proportion of full sibs within family arrays. Family (i.e., individual maternal trees) outcrossing rates were estimated fixing the pollen allele frequency to the global population estimate to increase the precision of the estimated rate, thus assuming an overall pollen pool. All confidence intervals (CIs) were obtained from 1000 sorted bootstrap values.
The pollen structure was analyzed with the TwoGener procedure that analyzes the genetic structure of pollen pools sampled by individual trees relative to the global pollen pool (Smouse et al. 2001). Briefly, the paternal contribution of each seed (i.e., pollen multilocus haplotype) is deduced by subtracting the maternal gamete contribution from the diploid genotype of each seed in a locus-by-locus analysis. Next, the population of deduced haplotypes is analyzed using a variance components analysis, and the proportion of variation attributed to the differences among mothers is summarized in the parameter FT. Computations were performed using software of R Dyer (available upon request to R Dyer from rjdyer{at}vcu.edu).
Because inbreeding is significant in the adult population and it can inflate the observed values of pollen genetic structure, we corrected the estimate of 'FT using the formula described in Austerlitz and Smouse (2001):
 | (1) |
FT is the pollen structure parameter due only to limited pollen dispersal and F is the adult inbreeding coefficient.
The estimate of FT can be used to further calculate the derivative measure of the average distance of pollination 
and the effective neighborhood pollination area (Aep). If we assume a bivariate normal distribution for pollen flow, the estimate of pollen structure FT is inversely related to the variance in pollen flow 
2 and reproductive adult density d, 
Given that FT is known from genetic data and d from ecological observations, it follows that the variance in pollen flow can be computed as 
. Next, 2 is used to estimate the effective number of pollen donors as Nep = 4
2d and the effective neighborhood area as Aep = 42 (Austerlitz and Smouse 2001).
|
Results and Discussion |
|---|
|
TopMaterials and MethodsResults and DiscussionReferences |
|---|
The allozyme markers yielded an average number of alleles per locus of 2.5, with 2 alleles for Pgi, Dia, and Pgm and 3 alleles for each of the other loci. Pooled offspring genetic diversity averaged across loci consisted of observed heterozygosity Ho = 0.439 (SD ± 0.248), expected heterozygosity He = 0.383 (SD ± 0.135), and excess of heterozygosity f = –0.121 (SD ± 0.353). The average pollen genetic diversity estimated from allele frequencies from MLTR output (Table 1) was He = 0.33 (SD ± 0.14). The genetic resolution, as measured by paternity exclusion probability based on these same pollen allele frequencies was 0.639. In practice, this marker resolution allowed for an average unambiguous identification of paternal alleles for 68.9% of cases all loci combined. At the within-locus level, resolution varied being the lowest for Pgi (34.3%); intermediate for Mdh, Per, and Skh (59.1%, 64.5%, and 66.6%, respectively); and high for Pgm and Dia (89.1% and 99.8%, respectively).
|
Mating System
The multilocus global population outcrossing rate was not significantly different from complete outcrossing (Table 2). In contrast, the minimum variance single-locus outcrossing rate was significantly lower than unity ts = 0.978 (95% CI = 0.979, 0.993). With respect to individual tree outcrossing rates, multilocus family rates were not significantly different from complete cross-pollination for all trees (results not shown). The estimate of biparental inbreeding for the total population (tm – ts) = 0.022 (95% CI = 0.004, 0.014) was significantly greater than zero, indicating the presence of a low proportion of consanguineous matings. The correlation of paternity for outcrossed progeny was significant (rp = 0.418, 95% CI = 0.263, 0.511), indicating that certain adults in the population performed better than others. When we translated paternity correlation values into the effective number of pollen donors (Nep = 2.392), it yields a very low estimate of the effective number of trees siring most of the seeds (Table 2). This very low estimate may reflect a bias in MLTR in estimating this parameter as indicated in other studies (see Hardy et al. 2004). So, whereas estimates of outcrossing rate and biparental inbreeding seem very robust in MLTR, estimates of Nep might be best done with other approaches. Adult inbreeding, as estimated from the progeny array data, appears to be nonexistent (Table 2).
|
Pollen Genetic Structure
The variance structure analysis performed on the pollen pool of Q. velutina at Tyson reserve was
'FT=0.078 (P < 0.001), indicating significant differences in the distribution of the pollen pool from tree to tree (Table 2). In an analysis of the adult population, Idol (1991) found significant inbreeding in the adult trees with fis = 0.17. We are not sure the reason for this high value of fis, although it could be due to the history of the population, which included selective logging in the past 100 years. Nevertheless, to take into account for this characteristic of the adult trees, we corrected the pollen structure parameter using Equation 1 and found a slightly smaller value of 0.066.
Pollination Neighborhood and Number of Pollen Donors
The density of adult trees with diameter at breast height >20 cm was d = 16.25 trees ha–1 (Hampe 1984). Adult tree density is not even across the reserve; thus, we should expect that the variance around individual trees in adult density will increase the estimated variance in pollen flow (2 = 1/[8dFT] = 0.0371). If we assume that all adults produce pollen, then the average distance of pollen movement = (2/2)1/2 = 24.1 m. The estimate of effective number of fathers is Nep = 42d = 7.58. Finally, the effective neighborhood area of the pollinators is Aep = 42 = 0.466 ha or a 41.9-m radius circle around maternal trees. These assumptions may underestimate these parameters, but the estimates provide values for comparison with other studies.
Potential Sources of Bias of Pollen Structure Estimates
Barring adult inbreeding effects that have been already taken into account, potential bias in FT can appear if intermaternal distances are smaller than 5 (Austerlitz and Smouse 2001). In our study, the average separation of the sampled trees was 618 m, which is much greater than 5 = 121 m. Second, one of the properties of FT is that the variance is minimized if the number of offspring per tree is in the order of 
(Smouse et al. 2001). In our case, 
seeds, which we fulfilled. Thus, the possible FT bias due to insufficient progeny sampling or unequal spatial maternal distribution can be ignored, and our estimates of Nep and Aep can be regarded as coherent with a bivariate normal model, although the normality assumption of pollen dispersal curves cannot be tested with the available information. (For discussion of other pollen dispersal curves, see Austerlitz et al. 2004).
One advantage of the TwoGener method is that it is relatively robust to genetic marker resolution (Smouse et al. 2001). The resolution in this paper (64% of paternity exclusion probability) based on the simulations of Smouse et al. (2001) might result in a high variance around the mean, but the mean should be relatively unbiased. The number of maternal trees, however, may be limited for a TwoGener type study, and future studies should consider a larger sample of trees. Irwin et al. (2003) recommend increasing the number of maternal arrays at the expense of several offspring per maternal tree if a compromise is necessary. In our case, the large intertree spacing helps mitigate this problem. Moreover, the pollen structure observed is within the range of results found for continuous forests of oaks for 54 maternal trees in a nearby forest population of Missouri oaks (Smouse et al. 2001).
Mating Patterns in Black Oak
Mating system and pollen structure analyses of Q. velutina at the Tyson reserve has revealed an interesting result in this oak species population. Despite the high outcrossing rate and the small proportion of biparental inbreeding (tm – ts = 0.022), the effective number of pollen donors and the estimated neighborhood size attributable to pollen movement are not very large. In fact, an estimate of 7.5 effective pollen donors is comparable with the estimates for other oaks (Smouse et al. 2001; Sork et al. 2002; Austerlitz et al. 2004) and is not particularly small for this parameter. For example, Irwin et al. (2003) estimated Nep to be 2–3 individuals for a given season in Albizia julibrissin, an insect-pollinated tree. Our result is typical for the genus Quercus and can be explained by the fact that the effective number of pollen donors is much smaller than an observed count of number of fathers that might be determined through paternity analysis (Smouse and Sork 2004). The difference in the 2 measures is that Nep is an estimate of the relative contribution of pollen donors to the progeny gene pool, which can be biased by a few individuals, usually quite proximate, whereas the actual census number of fathers will include many pollen donors, including many that fertilized only one progeny (Smouse and Sork 2004). For example, Streiff et al. (1999) observed numerous pollen sources for their oak populations based on paternity analysis, with the most successful identified pollen donors being close to the seed source and many pollen sources being beyond their study plot. If their data were analyzed with a structure approach, the effective number of pollen donors would likely be similar to our estimates (Smouse et al. 2001). So, what our findings illustrate is that mating patterns can be highly outcrossed but outcrossed predominantly by a few individuals. The fact that the biparental inbreeding is low indicates that those individuals siring the offspring are weakly related to the maternal plant.
The findings of this study indicate that the mating patterns of Q. velutina have very low inbreeding but a relatively small neighborhood area through pollen movement, which means that the population is locally subdivided where natural selection can have an impact. Our estimates of neighborhood size might increase slightly if we were to use a different dispersion curve (see Austerlitz et al. 2004), and the values of Nep might increase if we were to include more years of observation (see Nakanishi et al. 2005). Nevertheless, our results join a growing pool of studies which indicate that local population subdivision is possible even in highly outcrossing tree species.
|
Acknowledgments |
|---|
The authors thank F. Austerlitz, P. E. Smouse, and R. Dyer for help with analysis and comments on the manuscript and for the input of 2 anonymous reviewers. J.F.F-M. was supported with a Graduate School dissertation grant from the University of Missouri-St Louis while writing the first draft of this manuscript. J.I. and V.L.S. were funded by the Missouri Department of Conservation and V.L.S. also by National Science Foundation DEB 0072909.
|
Footnotes |
|---|
Corresponding Editor: James Hamrick
Received December 12, 2005
Accepted August 3, 2006
|
References |
|---|
|
TopMaterials and MethodsResults and DiscussionReferences |
|---|
-
Austerlitz F, Dick CW, Dutech C, Klein EK, Oddou-Muratorio S, Smouse PE, Sork VL. (2004) Using genetic markers to estimate the pollen dispersal curve. Mol Ecol 13:937–954.[CrossRef][Medline]
Austerlitz F and Smouse PE. (2001) Two-generation analysis of pollen flow across a landscape. II. Relation between ft, pollen dispersal and interfemale distance. Genetics 157:851–857.
Dyer RJ and Sork VL. (2001) Pollen pool heterogeneity in shortleaf pine, Pinus echinata Mill. Mol Ecol 10:859–866.[CrossRef][Medline]
Fernández-M JF and Sork VL. (2005) Mating patterns of a subdivided population of the Andean oak (Quercus humboldtii Bonpl, Fagaceae). J Hered 96:635–643.
Fowells HA. (1965) Silvics of forest trees of the United States. (US Department of Agriculture, Washington, DC).
Gottlieb LD. (1981) Electrophoretic evidence and plant populations. In Reingold J, Harborne JB, Swain T (Eds.). Progress in phytochemistry(Pergamon Press, New York) pp. 1–46.
Hampe CL. (1984) A description of species composition, population structure and spatial patterns in a Missouri oak-hickory forest [Master's thesis]. (University of Missouri, St Louis, MO).
Hamrick JL and Godt MJW. (1996) Effects of life history traits on genetic diversity in plant species. Philos Trans R Soc Lond Ser B Biol Sci 351:1291–1298.
Hardy OJ, Gonzalez-Martinez SC, Colas B, Freville H, Mignot A, Olivieri I. (2004) Fine-scale genetic structure and gene dispersal in Centaurea corymbosa (Asteraceae). II. Correlated paternity within and among sibships. Genetics 168:1601–1614.
Idol J. (1991) Outcrossing rate and genetic structure within a population of Black oak (Quercus velutina Lam.) in a Missouri forest [Master's thesis]. (University of Missouri, St Louis, MO).
Irwin AJ, Hamrick JL, Godt MJW, Smouse PE. (2003) A multiyear estimate of the effective pollen donor pool for Albizia julibrissin.. Heredity 90:187–194.[CrossRef][Web of Science][Medline]
Manos PS and Fairbrothers DE. (1987) Allozyme variation in populations of six northeastern American red oaks (Fagaceae: Quercus subg. Erythrobalanus). Syst Bot 12:365–373.
Mitton JB, Linhart YB, Hamrick JL, Beckman JS. (1977) Observations on the genetic structure and mating system of ponderosa pine in the Colorado front range. Theor Appl Genet 51:101–105.[CrossRef]
Nakanishi A, Tomaru N, Yoshimaru H, Manabe T, Yamamoto S. (2005) Interannual genetic heterogeneity of pollen pools accepted by Quercus salicina individuals. Mol Ecol 14:4469–4478.[Medline]
Ritland K. (1989) Correlated matings in the partial selfer Mimulus guttatus. Evolution 43:848–859.[CrossRef]
Ritland K. (2002) Extensions of models for the estimation of mating systems using n independent loci. Heredity 88:221–228.[CrossRef][Web of Science][Medline]
Smouse P, Dyer RJ, Westfall RD, Sork VL. (2001) Two-generation analysis of pollen flow across a landscape I. Male gamete heterogeneity among females. Evolution 55:260–271.[CrossRef][Web of Science][Medline]
Smouse PE and Sork VL. (2004) Measuring pollen flow in forest trees: an exposition of alternative approaches. For Ecol Manage 197:21–38.[CrossRef]
Soltis DE, Haufler CH, Darrow DC, Gastony GJ. (1983) Starch-gel electrophoresis of ferns: a compilation of grinding buffers, gel and electrode buffers, and staining schedules. Am Fern J 3:9–27.[Medline]
Sork VL, Davis FW, Smouse PE, Apsit VJ, Dyer RJ, Fernandez-M JF, Kuhn B. (2002) Pollen movement in declining populations of California Valley oak, Quercus lobata: where have all the fathers gone? . Mol Ecol 11:1657–1668.[CrossRef][Medline]
Sork VL, Huang S, Wiener E. (1993) Macrogeographic and fine-scale genetic structure in a North American oak species, Quercus rubra L. Ann Sci For 50:261–270.
Streiff R, Ducousso A, Lexer C, Steinkellner H, Gloessl J, Kremer A. (1999) Pollen dispersal inferred from paternity analysis in a mixed oak stand of Quercus robur L. and Q. petraea (Matt.). Liebl. Mol Ecol 8:831–841.[CrossRef]
Wright S. (1943) Isolation by distance. Genetics 28:114–138.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||