| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2Department of Biological Sciences ML6, University of Cincinnati, Cincinnati, Ohio 45221-0006; 3Oak Ridge Institute for Science and Education, Science/Engineering Education Division, Oak Ridge, Tennessee 37831-0117; 4Mathematical Sciences, University of Cincinnati, Cincinnati, Ohio 45221-0025; and 5Ecological Exposure Research Division, National Exposure Research Laboratory, U.S. Environmental Protection Agency, Cincinnati, Ohio 45268
Received for publication July 21, 1998. Accepted for publication February 2, 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
8-km intervals along a 320-km transect; (2) among ramets (N = 220; 117 genotypes detected) from five study sites separated by 50–3000 m; and (3) even among ramets within each study site [N = 44 per site; from 13 to 34 genotypes detected per site (270 m2)] exceeds that previously found in those more geographically widespread allozyme surveys. Among the 260 ramets analyzed here, the mean number of bands scored per individual was 48.61 (SD = 2.80). Mean genetic similarity among ramets collected along the 320-km transect was 0.91, which was within the range of mean genetic similarity within the five study sites (range: 0.89–0.95). Among the five study sites, 61% of the samples analyzed appeared to be clonal ramets, with up to 12 clones detected for 44 ramets sampled within a site. Clones grew intermingled and ranged up to 39 m in extent. Permutation tests of genetic similarity revealed significant genetic differentiation between each of the five study sites. Consistent with the previous allozyme studies, T. latifolia was characterized by extremely low genetic variation relative to levels of polymorphism detected at VNTR loci in other plant species. Estimated heterozygosity among ramets along the 320-km transect ranged from 0.11 to 0.13, while that within the five study sites ranged from 0.05 to 0.12. Estimates of Fst (0.32–0.41) also indicated considerable genetic subdivision among these stands. Significantly higher genetic diversity was detected at the two study sites that chemistry and toxicity data indicate to be the most severely impacted by pollutants. Although this correlation does not establish cause and effect, the results of this study indicate that the analysis of genetic diversity at VNTR loci may be a useful tool for monitoring anthropogenic-induced changes in the genetic structure of natural populations of plants.
Key Words: asexual reproduction • cattails • clonal growth • DNA fingerprinting • population genetics • Typhaceae • variable-number-tandem-repeat (VNTR) analysis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). Not surprisingly, given its ability to grow across a variety of environmental conditions, numerous morphological and physiological differences have been documented between populations (McNaughton, 1966, 1967
; Suda, Sharitz, and Straney, 1977
). Species with the above characteristics generally harbor relatively high levels of genetic diversity (Nevo, 1988
). However, surveys of allozyme loci in T. latifolia populations have revealed an almost complete lack of genetic variation. Mashburn, Sharitz, and Smith (1978)
found no genetic variability in ten allozyme systems across 74 populations throughout the southeastern United States. Surveying the same ten allozyme systems in 30 T. latifolia stands in the eastern part of the United States, Sharitz et al. (1980)
detected polymorphism at just a single locus in only one population.
Whether this lack of allozyme polymorphism is characteristic of just those loci surveyed or reflects a general paucity of genetic diversity in T. latifolia can be more thoroughly investigated by using additional molecular techniques that assess genetic variation at other, potentially more variable regions of the genome. Here we use synthetic DNA probes that detect tandem repeats of short "core" sequences of genomic DNA at multiple loci simultaneously to examine population genetic diversity in T. latifolia. Such probes have detected relatively high levels of genetic variation in a variety of other plant species (Rogstad, 1993, 1994
; Weising et al., 1995
; Busemeyer et al., 1997
; Kumar and Rogstad, 1998
). Since the variation detected by these probes is primarily due to differences among alleles in the number of tandem repeats of a DNA core sequence, the loci surveyed by these probes have been termed variable-number-tandem-repeat (VNTR) loci. VNTR loci have been increasingly used in population genetic studies across a wide variety of species because they have been demonstrated to be among the most variable genetic markers known (Rogstad, 1996
).
We had three principal objectives in investigating genetic variation at VNTR loci in T. latifolia. First, since previous studies have indicated extremely low levels of genetic diversity at allozyme loci in T. latifolia, we were interested in determining whether this was also the case for VNTR loci. Little information exists concerning variation at VNTR loci in plant species known to have almost no genetic variation at allozyme loci (but see Alberte et al., 1994
). Unlike allozymes, most VNTR loci are thought to be noncoding, selectively neutral markers (Jeffreys et al., 1988
) so there is no a priori reason to expect VNTR variation in T. latifolia to parallel that of allozyme loci. By comparing variability at allozyme and VNTR loci we hope to achieve greater insight into the evolutionary forces acting upon genetic diversity in T. latifolia. Moreover, obtaining a more complete picture of the overall level of genetic variation in T. latifolia is essential for understanding the extent to which an underlying genetic basis is responsible for its ability to grow across such a broad spectrum of environmental conditions.
Second, T. latifolia is known to propagate vegetatively, via rhizomatous growth, as well as by seed, but studies of the patterns and extent of clonality in natural stands have not been previously possible because of the lack of genetic variation at allozyme loci. Genetic variation at VNTR loci has been used to establish the presence of clonal growth in natural populations of a variety of other plant species (Rogstad, Nybom, and Schaal, 1991
; Rogstad, Wolff, and Schaal, 1991
; Antonius and Nybom, 1994
; Kumar and Rogstad, 1998
). Through intensive local sampling of T. latifolia within several different stands we may be able to assess, from variation at VNTR loci, the prevalence of clonal spread. Both theoretical work and empirical studies indicate that extensive clonal growth tends to reduce genetic variability, at least within populations (Les, 1991
; Sipes and Wolf, 1997
). However, before the degree to which vegetative propagation contributes to the lack of genetic diversity at allozyme loci or other regions of the genome can be determined, the extent of clonal growth in natural populations of T. latifolia must first be established.
Third, we were interested in the potential effect that differences in the extent of pollution in an area may have had on the genetic structure of natural stands of T. latifolia. Many chemical contaminants have been demonstrated to induce genetic mutations, at least in some organisms (Fishbein, Flamm, and Falk, 1970
). If environmental contaminants are inducing mutations in T. latifolia, then genetic variation at VNTR loci may be expected to increase with increasing exposure levels. Furthermore, the presence of chemical pollutants in the environment may alter selective pressures on T. latifolia. Such altered selection may lead to changes in allele frequencies so that significant differences in genetic structure occur between T. latifolia populations occupying areas differing in the degree of contamination by pollutants.
For example, reduced levels of genetic diversity have been found to be correlated with increased levels of chemical pollution in several animal species (Gillespie and Guttman, 1989
; Guttman, 1994
). The loss of genetic variation within a population may reduce its potential resiliency to future evolutionary challenges. Alternatively, if exposure to pollutants increases mutation rates, the genetic load of impacted populations should increase since most mutations at functional loci will be deleterious. Given that pollutants may alter genetic diversity within populations, monitoring the genetic structure of natural populations in areas contaminated by anthropogenic pollutants can provide information about the genetic response of populations to such environmental stress. However, few studies to date have evaluated the impact of chemical contaminants, other than heavy metals, on the genetic structure of natural plant populations (but see Keane, Smith, and Rogstad, 1998
).
To our knowledge, this is the first study to investigate genetic diversity in T. latifolia other than at allozyme loci. Specifically, we analyzed genetic variation at VNTR loci in natural stands of T. latifolia to examine the following issues: (1) is VNTR diversity higher than allozyme diversity in T. latifolia? (2) does variation at VNTR loci still indicate a reduced amount of genetic diversity in T. latifolia relative to other plant species? (3) are VNTR markers informative about possible clonal propagation in T. latifolia stands? and (4) are environmental differences in the levels of anthropogenic contaminants associated with significant differences in local genetic structure in T. latifolia?
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). In this study, species identification was based on the presence of wide leaves (6–15 cm) and terminal staminate spikes contiguous with the lower pistillate inflorescences (Voss, 1972
). The herbaceous shoots of cattails may or may not develop an apical flowering spike and produce linear vertical leaves that can exceed 2 m in height. Pollen is wind dispersed, but individuals are apparently self-compatible as well (Dickerman and Wetzel, 1985
), although seed production via agamospermy has not been ruled out. Seeds are also dispersed by wind, but established stands are thought to be maintained predominantly by vegetative growth (clonal production of underground rhizomes that give rise to new shoots at intervals) rather than by seedling recruitment (McNaughton, 1968
). The typical size or number of clones at a particular site was previously unknown due to a lack of morphological, allozyme, or other markers for distinguishing clones. In this study, we examined genetic diversity at VNTR loci in T. latifolia on three different geographical scales. First, we collected T. latifolia leaves along a 320-km transect (hereafter termed the LDT for "long-distance transect") between Louisville, Kentucky (38°14.8' N; 85°45.7' W), and Circleville, Ohio (39°36.2' N; 82°56.8' W), in July 1996 to estimate genetic diversity at VNTR loci in T. latifolia over a relatively large spatial scale. The transect was equally divided into 40 intervals of 8-km in length, and the first T. latifolia population that was accessible in each interval was sampled. If no population could be located within a given 8-km interval, two populations separated by at least 1.6 km were sampled in the succeeding interval. We hypothesized that individual vegetative clones would not span such distances, and sampling a single ramet at each site would thus allow us to examine genetic diversity over 320 km of individuals established from seeds.
Second, using a more intensive local sampling scheme, we also collected T. latifolia leaves for VNTR analysis from five separate sites located on the Wurtsmith Air Force Base (WAFB) near Oscoda, Michigan (44°25.8' N; 83°19.6' W; Fig. 1a). These five sites were separated by 0.5–3 km. Each of the sites contained large, nearly monospecific stands of T. latifolia, which allowed us to use the same sampling regime at each site. At each site, a total of 45 leaves were collected from separate shoots. Leaves were collected at 3-m intervals along each of three parallel 45-m linear transects (15 shoots-leaves per transect). Transects were separated from each other by 3 m. Sample locations were numbered sequentially along the transects. The five sites enabled us to investigate genetic variation in cattail stands separated by 50 to 3000 m as well as within 45 x 6 m plots. For both the LDT and the local transects, leaf samples were placed in individually labeled plastic bags, stored on ice, and transferred to a -70°C freezer within 7 d.
|
In August 1995, T. latifolia leaves were collected at four sites from plants growing along the edges of a large wetland located several hundred metres south of an abandoned landfill and fire training area on the WAFB (Fig. 1b). The intervening woodlands between the wetland and the landfill/fire training area contain numerous groundwater seeps that drain into the northern end of the wetland. This groundwater then flows southward through the wetland, which empties into a tributary of the Au Sable river. Previous analysis of the groundwater from each of these seeps by the NCIBRD detected one or more of six organic compounds classified as "hazardous substances" by the U.S. Environmental Protection Agency (EPA; Sittig, 1991
), as well as elevated levels of iron in the groundwater from several adjacent seeps in this area. None of these organic compounds (or any other substances classified by the U.S. EPA as hazardous) was detected at any of the other groundwater seeps surrounding the wetland. One of the sites (site A) at which T. latifolia leaves were sampled is located in the area where groundwater from the adjacent seeps contaminated with hazardous organic compounds and high levels of iron enters the wetland. A second site (B) was located 50 m north (upstream) of site A in an area where seep groundwater in which no hazardous organic compounds and only trace amounts of iron have been detected drain into the wetland. A third site (C) was located 50 m south (downstream) of site A and any of the groundwater seeps. The final site (D) was located 0.75 km away from the contaminated groundwater seeps near the area where the water from the wetland drains into the tributary of the Au Sable river. During August 1996, a local sampling transect was also conducted at a pond (site E) on the WAFB that is located 3 km east of the wetland sites A through D (Fig. 1b). We chose to sample T. latifolia at site E because groundwater monitoring had detected that a plume of several hazardous organic compounds had moved through and beyond the area surrounding the pond (M. Henry, personal communication, NCIBRD).
At four of the five sites at which we collected T. latifolia, sediment and/or water samples collected between 1995 and 1997 had been analyzed by the NCIBRD for evidence of contamination, and these data are presented in Table 1. These data indicate that site A was impacted by pollutants to a greater extent than any of the other three sites for which sediment and/or water chemistry data were available. Unlike the other three sites, no evidence of organic pollutants has been detected at site B, which was located upstream from any known point sources of pollution and was immediately adjacent to the pollution-impacted site A.
|
). By the standards of these tests, only the water sample from site A and the sediment sample from site E were considered to be chronically toxic to the test organisms (water toxicity tests: Ceriodaphnia dubia and Daphnia magna; sediment toxicity tests: Hyalella azteca). None of the other water or sediment samples exhibited evidence of toxicity in these tests. These results are consistent with the chemistry data for these samples. The water sample from site A was the only water sample tested in which organic pollutants had been detected. Among the sediment samples tested, the highest concentrations of organic pollutants were detected in the sediments collected at site E.
Genomic DNA was obtained from leaf samples by grinding 1 g of leaf tissue in liquid nitrogen and incubating the powdered tissue for 4 d at 68°C in 12 mL of 2x hexadecyltrimethylammonium bromide with 1% 2-mercaptoethanol. Following incubation, DNA was extracted and gel purified as described in Rogstad (1993)
. We determined that long-term incubation improves DNA yield at the gel purification step. Genomic DNA was digested with a fivefold excess of the endonuclease HaeIII according to manufacturer's (New England Biolabs, Beverly, Massachusetts) instructions, and the resulting DNA fragments were size separated by electrophoresis (3 µg DNA/lane) in submerged 1.2% agarose equivalent gels (with Synergel; Diversified Biotech, Boston, Massachusetts). Gels were run at 4°C for 45 h at 60 V until the bromophenol blue marker dye migrated to the end of the 24-cm gel. In addition to the digested DNA sample, each lane contained HindIII cut lambda DNA (25 ng) as a molecular mass marker to verify that all lanes on a gel ran uniformly. Following electrophoresis DNA fragments were transferred to a nylon membrane (Biodyne A, GibcoBRL, Gaithersburg, Maryland) by Southern blotting.
The blotted nylon membranes were then successively hybridized with three different radiolabeled, polymerase chain reaction-synthetic tandem repeat probes (PCR-STR) produced according to Rogstad (1993)
. The core sequences of the three PCR-STR multilocus probes were GACA, GATA, and GCAC, respectively. The thermocycling parameters used to generate the GACA PCR-STR probe were: 12 cycles at 98°C for 1 min, 49°C for 1 min 10 s, and 68°C for 2 min. For the GATA PCR-STR probe the thermocycling parameters were: 14 cycles at 98°C for 1 min 10 s, 51°C for 1 min, and 70°C for 1 min 30 s. The GCAC PCR-STR probe was produced using 14 cycles of the same thermocycling parameters used to generate the GACA PCR-STR probe. Membrane prehybridization and hybridization were both conducted overnight at 60°C and, along with the wash procedure, followed the methods of Westneat et al. (1988)
with the following exceptions: (1) the prehybridization solution contained 0.65% (w/v) bovine serum albumin; (2) the prehybridization solution for the GCAC probings contained 0.237 mol/L NaPHO4; and (3) membranes were washed twice for 1 h at 60°C in 2x sodium chloride sodium citrate (SSC) following the GACA and GATA probings or 1.8x SSC following the GCAC probings. After washing, filters were exposed to autoradiograph film (Biomax MS-1, Kodak, Rochester, New York) with an intensifying screen (Biomax MS, Kodak, Rochester, New York) at -70°C. Two to three exposures of different time lengths (1–5 d) were made of each probed filter in order to better visualize bands of varying intensity. Following probings, membranes were stripped for 2.5 min with 0.1% (v/v) sodium dodecyl sulfate heated to 95°C.
The leaves collected at 40 sites along the LDT were split into two groups of 20, and the DNA samples from each group of 20 was electrophoresed on a separate gel. Samples collected from successive stands along the LDT were assigned to different gels (LDT1 and LDT2). Thus, the distribution of distances (range 16 – 304 km) between the populations from which the individuals were collected was similar for the samples electrophoresed on both gels. On each gel, every fragment with a unique migration distance (a population band; bands were considered different when migration distance differences exceeded 1 mm) was numbered sequentially and all individuals were manually scored for the presence or absence of these population bands detected with each of the three PCR-STR probes.
For the samples collected at the WAFB, we analyzed genetic diversity at VNTR loci from only the first 44 leaves collected along the three transects at each of the five sites. Initially, DNA from the first 22 samples collected at each site were electrophoresed sequentially on a single gel as were DNA from the next 22 samples. For each of these gels, bands were scored using the same procedure as with the LDT gels. After determining the population bands present on each of these ten gels following hybridization with a given probe, a subset of the individuals run on each gel was identified that contained among them all the population bands present on that gel. These ten groups of individuals were then electrophoresed together on additional "master gels" (Fig. 2) so that banding patterns obtained with each of the three probes could be compared across gels and among all 220 samples analyzed from the WAFB.
|
). Redundant examples of bands exhibiting complete linkage, as identified by GELSTATS, were removed and the resulting, finalized data set was used for the analyses presented here. Parameters computed by GELSTATS include the average number of bands per individual, the proportion of bands shared (mean similarity) between individuals (Nei and Li, 1979
; Wetton et al., 1987
), and the probability that two randomly chosen individuals from a population will have identical fragment profiles. All estimates of mean similarity were based on all possible pairwise comparisons of the relevant samples. Since pairwise similarity values are not independent and are not necessarily normally distributed (Lynch, 1990
), GELSTATS employs permutation tests (Good, 1994
) to determine whether differences in similarity between groups are statistically significant. All permutation tests were based on 5000 iterations and generate a P value by which the significance of the difference in average similarity between groups can be assessed (Rogstad and Pelikan, 1996
).
SYSTAT nonmetric multidimensional scaling (MDS; SYSTAT 6.0; Wilkinson, 1990
) was used to graphically depict the genetic similarity (as calculated by GELSTATS) among all 220 samples analyzed from sites A through E. A MDS plot was computed in two dimensions, minimizing the Kruskal stress, using the GELSTATS similarity matrix of all possible pairwise comparisons of the 220 samples. The distances separating sample points in MDS space reflect the genetic similarity among samples.
Assuming that Hardy-Weinberg conditions apply across loci, GELSTATS also estimates allele frequencies from population band frequencies according to Hardy-Weinberg expectations (Stephens et al., 1992
). The percentage of polymorphic loci was calculated according to Kimura and Ohta (1971)
. GELSTATS estimates heterozygosity for a population using the bias-correcting method of Jin and Chakraborty (1993)
and uses a jackknife resampling procedure (Weir, 1990
; modified as described in the GELSTATS documentation with N - 2 observations per resampling) to test whether differences in heterozygosity between populations are statistically significant. A measure of population subdivision, Fst, is estimated by GELSTATS according to the methods of Nei (1973
; based on heterozygosity) and Lynch (1991
; based on similarity). An estimate of the number of migrants exchanged per generation (Nm) between stands was calculated from both the Nei and Lynch estimates of Fst according to Nm = [(1/Fst) - 1]/4 (Wright, 1931
).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Among the 220 T. latifolia ramets analyzed from the five sites sampled on the WAFB (distances between sites ranged from 50 to 3 km), we detected a total of 69 population bands with the three PCR-STR probes. Twenty-six (38%) of these population bands were monomorphic across all five sites, and 52 (75%) of the population bands were present in at least one sample from each of the five sites from which leaves were collected. The total number of population bands detected at each of the five sites did not differ significantly (Table 2; 
2 = 0.79; df = 4; NS). Three population bands were found exclusively among individuals from site A (private alleles; Slatkin, 1985
), and six population bands were found only at site E (Table 2). Over all 220 samples, the mean number of bands scored per sample was 49.34 (SD = 2.21; range 45–55). The mean number of bands scored per sample among all samples within a study plot differed significantly between the five locations where leaves were collected, with site A having the lowest mean number of bands per sample and sites C and E the highest (Table 2).
|
). All comparisons of genetic similarity within sites vs. between sites found genetic similarity to be significantly greater within a site than between sites except for the comparison of genetic similarity within site A with that between sites A, and C, which did not differ significantly (Table 3). All of these results taken together suggest that all five of the sites were genetically differentiated from each other.
|
|
|
2 = 3.94; df = 4; NS: Table 4). The greatest distance separating ramets of the same clone was 39 m. Since none of the samples collected at two different sites ever possessed identical banding patterns, we found no evidence of clonal spread extending over 50 m.
|
2 = 9.80; df = 4; P < 0.05; Table 4) and the proportion of samples with unique genotypes at each site (2 = 15.70; df = 4; P < 0.01; Table 4). Site A exhibited both the highest number of genotypes and the highest proportion of samples with unique genotypes. Within each of the five sampled areas, the probability that two randomly selected samples will share all their bands was lowest for site A (Table 4). Along with the analysis of mean genetic similarity within a stand, these results indicate a greater genetic diversity among the T. latifolia growing at site A, the site most impacted by pollutants (Table 1), than at any of the other sites sampled.
Since it seemed likely that most samples with identical fragment profiles were due to clonal propagation, genetic data from only one ramet of each unique genet were used to estimate the number of loci analyzed per sample, the proportion of polymorphic loci, and heterozygosity within each site (Table 4) as well as Fst. The mean number of loci analyzed per sample ranged from 42.5 at site A to 48.0 at site D, but the proportion of polymorphic loci was relatively low for VNTR loci, ranging only from 0.08 at site D to 0.22 at sites A and E. Heterozygosity among all the samples analyzed from sites A through E was estimated to be 0.14. The estimated heterozygosity among all samples collected within each of the five sites differed significantly from that at each of the other sites (P < 0.05 for all pairwise comparisons) with site A having the highest heterozygosity and site D the lowest (Table 4). The estimate of Fst based on Nei's (1973) method was 0.41 and on Lynch's (1991)
method was 0.32. These estimates of Fst indicate that approximately one-third of the total genetic variability detected was attributable to restricted gene flow between sites. Based on the above Fst values, the estimated number of migrants exchanged per generation between sites ranged from 0.37 to 0.53.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Sharitz et al., 1980
). This result is in accord with allozyme- and VNTR-based estimates of genetic diversity in other plant species, which consistently find variation at VNTR loci to exceed that at allozyme loci (Rogstad, 1996
). The consistently greater variation detected at VNTR loci relative to allozyme loci is presumably due to higher mutation rates combined with lower selective constraints on polymorphism at the presumably neutral VNTR loci. Thus, the greater variability at VNTR loci should make them a more useful marker for population genetic studies where a greater discriminatory power than afforded by allozymes is needed.
Typha latifolia is known to hybridize with T. angustifolia (Sharitz et al., 1980
), with the hybrids typically having an intermediate morphology. Although our casual observations indicated that only plants with T. latifolia morphology were present in the areas we sampled, isolated, relatively small patches of what appeared to be T. angustifolia were observed on a very limited basis elsewhere on the WAFB. This opens the possibility that hybridization between these two species may be occurring to an unknown degree in the areas we sampled and that some of the genetic variation we detected may have been the result of introgression between the species. A previous study of allozyme loci in ten populations of T. angustifolia widely distributed throughout the eastern United States found no genetic variation (Sharitz et al., 1980
). In addition, T. angustifolia was found to be distinguishable from T. latifolia at only a few loci. Therefore, hybridization between these two species, if it has occurred in the areas we sampled, would not seem to have contributed greatly to the higher genetic diversity we detected at VNTR loci relative to allozyme loci. The greater genetic variability at VNTR loci may also be useful in investigating whether hybridization between T. latifolia and T. angustifolia is occurring and how this may be affecting genetic variation on the spatial scales examined in this study.
Our analyses revealed genetic variation at VNTR loci in T. latifolia among samples collected at 8-km intervals along a 320-km transect, between samples from five study sites separated by 50–3000 m, and among samples 3 m apart within 270-m2 plots. Despite the relatively small spatial scale of the five sampling sites at the WAFB, our analyses of VNTR polymorphisms revealed significant genetic differentiation between all of these stands. The genetic similarity data clearly demonstrate these stands to be genetically differentiated, with the ramets within each site being significantly more similar to each other than to the ramets at any of the other study sites. The estimates of Fst (0.32–0.41, with only unique genotypes for each site analyzed) also indicate considerable genetic subdivision among these stands. A survey of 25 studies examining allozyme polymorphisms in long-lived herbaceous perennials found that Fst averaged 0.29 (Hamrick, Godt, and Sherman-Broyles, 1992
). Our estimates of Fst for T. latifolia suggest that gene flow is lower in this species than on average for other long-lived herbaceous perennials, especially given the relatively small area of this study.
The breeding system of T. latifolia seems likely to have contributed to some degree to the significant genetic differentiation we detected over very short interstand distances between the WAFB sites. Typha latifolia is thought to be primarily self-pollinating (Mashburn, Sharitz, and Smith, 1978
). However, the possibility that seeds assumed to be the result of selfing were actually produced without fertilization via agamospermy has not been rigorously examined. That T. latifolia commonly produces seed by selfing is supported by correlations between allozyme-based measures of Fst and plant breeding systems. A comprehensive review of seed plants by Hamrick and Godt (1989)
found the average Fst for 78 selfing species to be 0.51 and that for 134 outcrossing wind-pollinated species to be 0.09. The estimates of Fst for the closely situated sites of T. latifolia in this study are the highest VNTR-based estimates yet reported for plants (Table 5) and are more consistent with allozyme-based estimates found for selfing species. In addition, the results of this study indicate that vegetative propagation is common within stands of T. latifolia. Both selfing (or agamospermy) and clonal growth should promote genetic homogeneity within populations but tend to increase genetic differences between populations (Les, 1991
; Waycott, Walker, and James, 1995
), which is what we have found among the stands of T. latifolia we examined on the WAFB.
|
; Busemeyer et al., 1997
; Keane, Smith, and Rogstad, 1998
; Kumar and Rogstad, 1998
; Lim, Pelikan, and Rogstad, in press
). Only VNTR analyses of several agamospermous plant species have yielded higher estimates of genetic similarity within and between populations than we observed in T. latifolia (Van Heusden et al., 1991;
Kraft, Nybom, and Werlemark, 1996
). Although estimates of genetic similarity among samples within a site on the WAFB may have been biased upwards by the inclusion of ramets of the same clone in the analysis, estimates of genetic similarity between samples from different sites on the WAFB or among samples along the LDT did not include any possible ramets from the same clone and yet were equally as high as those among samples within a site. Hybridization between T. latifolia and T. angustifolia should result in a higher level of genetic diversity among hybrids relative to that found in individuals of either species alone. Therefore, even if some proportion of the leaves we sampled were hybrids between these two species, this would not alter our finding that the stands of cattails we analyzed were characterized by very low levels of genetic variation.
The extremely low level of genetic variation we observed at VNTR loci in T. latifolia reflects what has been previously found at allozyme loci in this species (Mashburn, Sharitz, and Smith, 1978
; Sharitz et al., 1980
). In this study, as well as in the allozyme studies, samples were collected from widespread geographic locations and diverse environmental conditions, suggesting that the lack of genetic diversity detected at these markers is characteristic of the species over a broad area and not a result of local sampling phenomena. The relatively low levels of genetic variation at VNTR and allozyme loci in T. latifolia are surprising in view of the general trend for species having broader geographic, climatic, habitat, and ecological ranges to harbor higher amounts of genetic diversity (Nevo, 1988
). In Hamrick and Godt's (1989)
review of allozyme polymorphism in seed plants, species with widespread ranges (N = 87) had significantly higher levels of genetic diversity than more narrowly distributed species (N = 134). In T. latifolia, estimates of genetic diversity at allozyme and VNTR loci are among the lowest values reported for plant species. Widespread species were also characterized by Fst values averaging 0.21 compared to 0.24 for narrowly distributed species. Our estimates of Fst in closely situated sites of T. latifolia are much larger than the average value for even the narrowly distributed species.
Very low levels of genetic variability detected in several other plant species have been attributed to high inbreeding (selfing and kin mating) and/or asexual reproduction. As previously noted, several agamospermous species have been found to be nearly invariant at VNTR loci across extremely large areas (Van Heusden et al., 1991
; Kraft, Nybom, and Werlemark, 1996
). However, other plant species with a breeding system similar to Typha exhibit much greater genetic diversity than has been observed in T. latifolia (e.g., Jeffries and Gottlieb, 1983
; Alberte et al., 1994
). Thus, the breeding system of T. latifolia alone may not entirely explain the low genetic diversity found in this species. Other factors that may possibly be contributing to the low genetic variation in T. latifolia include founder events, selection, and the organization and dynamics of the genome.
Regardless of the causal mechanism(s), if the extremely low levels of genetic diversity detected at VNTR and allozyme loci reflect the extent of genetic variation in the entire genome, then T. latifolia has adapted to grow successfully across a broad spectrum of ecological conditions despite possessing relatively little genetic diversity among individuals. It would also suggest that much of the morphological and physiological differences observed between T. latifolia populations may be due to environmental influences on the phenotype rather than genotypic differences between individuals. The ability of T. latifolia to adapt to a myriad of environmental challenges may be the result of remarkably flexible phenotypic responses that are mediated by a relatively invariant genome. This finding would certainly not be consistent with the widely held view in conservation biology that low levels of genetic diversity imply little capacity to respond to environmental change (Lynch, 1996
).
Although VNTR- and allozyme-based estimates of genetic diversity in T. latifolia are both low, an absence of polymorphism at allozyme loci is not necessarily indicative of a similar lack of genetic variation at VNTR loci. Eelgrass, Zostera marina L., is a marine angiosperm that, like T. latifolia, is capable of clonal propagation by rhizomatous growth and often occurs in large monospecific stands. Investigations of allozyme polymorphisms in Z. marina revealed virtually no genetic diversity within a population and extremely low levels of genetic differentiation between geographically distant populations of eelgrass (McMillian, 1982
). However, VNTR-based estimates of genetic similarity within and between populations of Z. marina range from 0.44 to 0.68 (Alberte et al., 1994
), which are typical of the range of values reported for plants, including those with relatively high levels of allozyme polymorphism. The VNTR-based estimate of Fst between three geographically disjunct eelgrass populations 30–150 km apart was 0.18, reflecting less population subdivision than that found between T. latifolia at the more closely spaced sites sampled in this study.
In contrast to previous allozyme studies, VNTR probes provided much greater resolution for detecting clonal growth in T. latifolia. This study indicates that the well-demonstrated ability of T. latifolia to propagate vegetatively has a substantial influence on the genetic composition of natural stands. Assuming all samples with identical fragment profiles to be part of the same genet, we found that 61% of the 220 samples we analyzed on the WAFB were ramets of one multiramet clone or another. Although this estimate may include some samples from different genets that shared all their bands by chance, we feel the frequency of these occurrences was relatively low, since the estimated probability of two randomly selected ramets from different genets having identical banding patterns within each of the five stands we sampled ranged from 2 to 6%. Moreover, we expect that more closely spaced sampling would likely reveal an even greater proportion of the shoots in our study plots to be coclonal ramets. Thus, the majority of shoots in these stands seem likely to have been produced asexually. Consequently less genetic diversity would be expected within these stands than if all the shoots developed from sexually produced seeds.
Patterns of genotypic diversity showed the T. latifolia stands we sampled to be multiclonal and not dominated by a single genotype. In addition, clones grew intermingled and ranged up to 39 m in size. Aspen clones also intermingle, but show high diversity (heterozygosity = 0.76), unlike cattails (Rogstad, Nybom, and Schaal, 1991
; Rogstad, 1996
). Studies of clonal propagation in other plant species have typically found populations to contain multiple clones of intermediate frequency with clones being restricted to one or a few populations (Ellstrand and Roose, 1987
). The consequences (e.g., mating, competitive combining ability) of clonal plants growing intermingled vs. clones excluding one another remain to be explored.
Chemical and toxicity analyses of water and sediment samples indicated that at least some of the locations on the WAFB where we investigated VNTR polymorphisms in T. latifolia differed in the extent to which they were impacted by pollutants. Several studies have documented correlations between exposure to contaminants and significant differences in allozyme frequencies in natural populations of animals (Gillespie and Guttman, 1989
, 1993). Our findings here suggest that differences in the amount of contaminants at the WAFB cattail study plots may be a potential factor contributing to their genetic differentiation.
Considering just the four sites for which toxicity and/or chemistry analyses were available, these data indicate that site A has been impacted more severely by pollutants than the other sites. Overall, our analyses of VNTR markers indicated that genetic diversity among the samples at site A was higher than that within any of the other sites we examined on the WAFB. Genetic similarity among the samples at site A was significantly lower than that at each of the other sites, while the estimate of heterozygosity at site A was significantly higher. Higher genetic diversity at site A was also indicated by the greater number of genotypes, higher proportion of individuals with unique genotypes, higher proportion of polymorphic loci, and lower probability of two samples within a site sharing all their bands by chance at site A relative to the other sites. The toxicity and chemistry data combined also suggest that site E may have been impacted by contaminants to a greater extent than either site B or site D. Among these sites, the samples from site E exhibited significantly higher genetic diversity than those from sites B and D, which did not differ significantly, at least in terms of within-stand genetic similarity. Thus, the greater impact by pollutants at site A and possibly site E may be contributing to the significantly higher levels of genetic diversity we detected at sites A and E relative to sites B and D.
Several of the organic contaminants detected at the WAFB are suspected mutagens (Sittig, 1991
), but we know of no empirical studies of the mutagenecity these in the genus Typha. If these pollutants are inducing mutations in cattails, we would expect the frequency of private alleles to be higher at the more impacted sites since mutations are random events and migration between the sites sampled on the WAFB is extremely low as evidenced by the estimates of Fst and Nm. We detected a total of nine private alleles across the five sites, and all of them were found at either site A (three) or site E (six). This finding is consistent with the hypothesis that mutation rates are higher at the more impacted sites, but the total number of private alleles and the frequency of each are very low, suggesting that mutation alone has not as of yet contributed extensively to the higher genetic diversity at the most impacted sites. The greater number of private alleles at site E may also reflect a lower migration rate between the pond and wetland relative to that between sites within the wetland. The four sites at the wetland are separated by no more than 0.75 km, while site E is located 3 km from the wetland.
Alternatively or in addition, the higher genetic diversity we observed at the most polluted sites may also be due to contaminant-induced changes in the selective pressures on plants growing at these locations. There are several possible scenarios in which changing selective pressures on T. latifolia due to increased levels of contaminants can lead to higher within-stand genetic diversity. First, genets with the highest heterozygosities may be competitively superior (Mitton, 1993
) to more homozygous genets in polluted areas. Thus, rising levels of contaminants in an area may select for genets with increased heterozygosities relative to uncontaminated areas. Second, seedling recruitment into established stands of T. latifolia is thought to be very low (McNaughton, 1968
). However, if increasing pollution causes greater mortality among established or establishing ramets, this may permit higher seedling colonization at the more polluted sites. If the colonizing seeds come from genetically differentiated populations, they may introduce greater genetic diversity into the more impacted areas. Third, genets that were most common within an area prior to the greater impact by pollutants may be gradually replaced, through clonal spread, by previously rare genets that are better adapted to the higher levels of contamination. During the period of gradual replacement, overall genetic diversity within a stand should increase. Finally, given that it has been suggested that members of hybrid swarms may play an important role in colonizing disturbed habitats (Briggs and Walters, 1984
), the possibility exists that some portion of the increased genetic variation detected at polluted sites is due to a higher rate of colonization by T. latifolia x T. angustifolia hybrids.
Validating pollution-induced selection as a contributing factor to the greater genetic diversity we detected at the most polluted sites requires demonstrating that genets with particular alleles have different fitnesses when growing in areas containing different levels of the specific contaminants that were found in our study areas. Each of the hazardous organic compounds detected in the water and sediment samples at the WAFB has been demonstrated to have negative effects on fitness components in a variety of organisms in laboratory tests (Verschueren, 1983
), but the concentrations of these organic compounds at the sites we sampled were below the permissible freshwater concentrations established by the U.S. Environmental Protection Agency (Sittig, 1991
). In addition, a study of T. latifolia grown in greenhouse mesocosms in which the aqueous concentrations of trichloroethylene (40 ppm) and benzene (25 ppm) were greater than that detected at any of the sites on the WAFB we sampled failed to find evidence of increased mortality relative to controls over a 3-wk period (Childress, 1997
). Nonetheless, the effects of persistently low, but environmentally relevant, concentrations of complex mixtures of these organic compounds, as well as iron, on the fitness of T. latifolia genets in nature are unknown. It is also not necessary that the contaminants act directly on T. latifolia in order for the level of pollutants at a site to affect stand genetic diversity. For example, contaminants may alter the diversity of microorganisms present in the sediments, which in turn changes the selective challenges confronting T. latifolia. We should also note that the history of contamination at the sites we sampled prior to 1995 (the first year of contaminant analysis) is unknown, but contaminant concentrations may have been higher when the base was actively used. Also, sampling for contaminants was not continuous and thus the degree to which pollutant concentrations have fluctuated through time is unknown.
Therefore, the associations we observed between genetic variation and the amount of contamination in an area does not necessarily establish environmental differences in pollution as a contributing factor to the significantly higher cattail genetic diversity at the most polluted sites. Studies assessing the effect of pollution on the genetic structure of geographically separated populations of plants must also consider the possible effects of variables such as founding events, genetic drift, selection on other traits, and other historical factors on genetic variation that are independent of any effect of environmental contaminants. By sampling T. latifolia stands over a relatively small spatial scale, including sites immediately upstream and downstream of the point sources of pollution as proposed by Gillespie and Guttman (1989)
we may have been able to control for, as best as is possible in natural populations, the effects of these confounding variables. Nonetheless, while the higher genetic diversity we observed among plants growing at the more contaminated sites may be due to the greater concentration of pollutants in these areas, we are unable to eliminate the possible effects of factors independent of the level of pollution on the genetic structure of the T. latifolia stands we investigated.
There are only a few published studies, mostly for animal species (Gillespie and Guttman, 1993
), that have documented a relationship between exposure to pollutants and differences in genetic variation in natural populations. In some of these studies, reduced levels of genetic diversity were detected in populations inhabiting contaminated sites, whereas others found elevated genetic variation at polluted sites (Gillespie and Guttman, 1989
). In contrast to our results for T. latifolia, the only other study known to us where VNTR loci are used to investigate the effect of pollution on plant population genetics detected lower genetic diversity among wild red raspberry (Rubus idaeus L.; Rosaceae) stands at contaminated sites on the WAFB (Keane, Smith, and Rogstad, 1998
). These studies seem to suggest that pollution-induced changes on genetic diversity are variable, making it difficult to generalize about the effects of pollutants on genetic variation in nature, although more research is needed to determine whether predominant response patterns exist.
While the results of this study do not conclusively demonstrate that environmental differences in pollution have contributed to the differences in genetic diversity we observed between the stands of T. latifolia we sampled, they do suggest that genetic diversity at VNTR loci has the potential to act as a sensitive research tool to examine whether pollution-induced changes in the genetic structure of natural populations of plants occur. There is an ever-increasing variety and quantity of chemicals present in the environment as a result of human activity. Pollution-induced changes in genetic diversity may have long-ranging consequences for a population's survival, but empirical studies assessing the impact of pollutants on the genetic diversity and structure of natural populations of plants are generally lacking. All previous investigations of this sort of which we are aware have utilized allozyme markers, but the typically greater genetic variation at VNTR loci suggests these loci may provide even greater sensitivity for detecting genetic changes associated environmental pollution. Ideally, both studies that examine variation at VNTR loci in natural populations of plants for several generations prior to and following impact by pollutants and those that manipulate pollutant levels across several populations are needed to ascertain the utility of VNTR markers as a tool for monitoring contaminant-induced effects on genetic structure. Such studies will also contribute to our understanding of the long-term consequences of environmental pollutants on the genetic stability and resiliency of natural plant populations.
|
FOOTNOTES |
|---|
6 Author for correspondence (FAX: 513-556-5299; e-mail: steven.rogstad{at}uc.edu)
.
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Antonius, K., and H. Nybom. 1994 DNA fingerprinting reveals significant amounts of genetic variation in a wild raspberry Rubus idaeus population. Molecular Ecology 3: 177–180.
Briggs, D. and S. M. Walters. 1984 Plant variation and evolution. Cambridge University Press, Cambridge.
Busemeyer, D., S. Pelikan, R. S. Kennedy, and S. H. Rogstad. 1997 Genetic diversity of Philippine Rubus moluccanus L. (Rosaceae) populations examined with VNTR DNA probes. Journal of Tropical Ecology 13: 867–884.[ISI]
Childress, C. R. 1997 Constructed wetlands for the treatment of wastewater containing metals and volatile organic compounds. Master's thesis, University of Cincinnati, Cincinnati, OH.
Dickerman, J., and R. G. Wetzel. 1985 Clonal growth in Typha latifolia: population dynamics and demography of ramets. Journal of Ecology 73: 535–552.[CrossRef][ISI]
Ellstrand, N. C., and M. Roose. 1987 Patterns of genotypic diversity in clonal plant species. American Journal of Botany 74: 123–131.[CrossRef][ISI]
Fishbein, L., W. G. Flamm, and H. L. Falk. 1970 Chemical mutagens; environmental effects on biological systems. Academic Press, New York, NY.
Gillespie R. B., and S. I. Guttman. 1989 Effects of contaminants on the frequencies of allozymes in populations of the central stoneroller. Enivronmental Toxicology and Chemistry 8: 309–317.
———, and ———. 1993 Allozyme frequency analysis of aquatic populations as an indicator of contaminant-induced impacts. In J. W. Gorsuch, F. J. Dwyer, C. G. Ingersoll, and T. W. La Point [eds.], Environmental toxicology and risk assessment, vol. 2, 134–145. American Society for Testing and Materials, Philadelpia, PA.
Good, P. 1994 Permutation tests: a practical guide to resampling methods for testing hypotheses. Springer-Verlag, New York, NY.
Guttman, S. I. 1994 Population genetic structure and ecotoxicology. Environmental Health Perspectives 102: 97–100.
Hamilton, M. B. 1997 Genetic fingerprint-inferred population subdivision and spatial genetic tests for isolation by distance and adaptation in the coastal plant Limonium carolinianum. Evolution 51: 1457–1468.
Hamrick, J. L., and M. J. W. Godt. 1989 Allozyme diversity in plants. In A. H. D. Brown, M. T. Clegg, A. L. Kahler, and B. S. Weir [eds.], Plant population genetics, breeding, and resources, 43–63. Sinauer, Sunderland, MA.
———, ———, and S. L. Sherman-Broyles. 1992 Factors influencing levels of genetic diversity in woody plant species. New Forests 6: 95–124.[CrossRef]
Jeffreys, A. J., N. J. Royle, V. Wilson, and Z. Wong. 1988 Spontaneous mutation rates to new lenght alleles at tandem-repetitive hypervariable loci in human DNA. Nature 332: 278–281.[CrossRef][Medline]
Jeffries, R. J., and L. D. Gottlieb. 1983 Genetic variation within and between populations of the asexual plant Puccinellia x phryganodes. Canadian Journal of Botany 61: 774–779.
Jin, L., and R. Chakraborty. 1993 A bias-corrected estimate of heterozygosity for single-probe multilocus DNA fingerprints. Molecular Biology and Evolution 10: 1112–1114.[ISI][Medline]
Keane, B., M. K. Smith, and S. H. Rogstad. 1998 Genetic variation in red raspberries (Rubus idaeus L.; Rosaceae) from sites differing in organic pollutants compared with synthetic tandem repeat DNA probes. Environmental Toxicology and Chemistry 17: 2023–2034.
Kimura, M., and T. Ohta. 1971 Theoretical aspects of population genetics. Princeton University Press, Princeton, NJ.
Kraft, T., H. Nybom, and G. Werlemark. 1996 DNA fingerprint variation in some blackberry species (Rubus subg. Rubus, Rosaceae). Plant Systematics and Evolution 199: 93–108.
0.05) as determined by permutation tests
