| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Population Biology |
2Department of Forestry, Michigan State University, East Lansing, Michigan 48824 USA
Received for publication March 30, 2004. Accepted for publication September 17, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: admixture • chloroplast DNA • microsatellite • Pinus resinosa • population genetics • refugia
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Simon et al., 1986
; Allendorf et al., 1982
; Mosseler et al., 1991
), in stark contrast to the rather high levels of polymorphism found in nearly every other pine. Among studied gymnosperms (mostly conifers) an average of 71% of allozyme loci are polymorphic at the species level (Hamrick and Godt, 1990
). Hamrick et al. (1992)
examined 93 genetic studies of pine species and found an average heterozygosity of 0.157 for allozymes, whereas it is effectively zero in red pine. Mosseler and colleagues (1992)
found virtual monomorphism for a number of randomly amplified polymorphic DNA (RAPD) markers. However, a few variants were observed in Newfoundland, and these were verified by restriction fragment length polymorphism (RFLP)-RAPD analyses (DeVerno and Mosseler, 1997
).
Despite the fact that red pine is by most standards depauperate in genetic variation, polymorphism for chloroplast microsatellite or simple sequence repeat (cpSSR) loci has been found (Echt et al., 1998
). This finding does not necessarily conflict with other studies, because cpSSRs may mutate at greater rates than do most loci (e.g., Di Rienzo et al., 1994
; Provan et al., 1999
). The amount of genetic diversity of cpSSR loci in red pine is small compared to that in other pines (e.g., Powell et al., 1995
; Vendramin et al., 1996
). Clearly, red pine has reduced variation for molecular markers, despite there being millions of individuals spread over its present range across the northcentral and northeastern United States and southcentral and southeastern Canada (Fig. 1).
|
; Wright et al., 1972
). Traits include survival and growth rates (Wright et al., 1972
; Yao, 1974
; Guries and Ager, 1980
), phenology, wood quality (Fowler and Lester, 1970
), and branching characteristics (Mosseler et al., 1992
). Generally, many traits in pines and other forest trees show concordant geographical patterns of genetic variation. Significantly, red pine shares these patterns, even though the overall amounts of genetic differentiation are smaller.
Prominent among geographic patterns in red pine are traits associated with latitude. For example, Vaartaja (1962)
found significant genetic differences in photoperiodic seedling growth response, such that under controlled short day schedules plants from northern populations cease growing and set buds earlier. Such a pattern of dormancy is generally considered a geographic adaptation in forest trees for avoiding early fall frost damage. Similarly, there are significant differences in the timing of flowering; for example, northerly red pine forests undergo anthesis up to 3–4 wk later. Frost damage is a significant problem for red pine, and the average annual frost-free period varies widely, from only 60 d up to 180 d, across the present day native range (Fowler and Lester, 1970
). Dhir (1973)
also found source differences in seedling shoot traits, and some of these traits are correlated with latitude. Northern trees appear to produce smaller seed in provenance tests and also suffer less frost damage (Fowler and Lester, 1970
). In addition, individuals of a small isolated population in Illinois (Fig. 1) produce white pollen (Brenneman, 1956
), a unique condition in red pine, and a rare, usually recessive genetic condition in pines generally (Johnson and Critchfield, 1974
).
Despite geographic genetic variation for important morphological traits, the paucity of variation for genetic markers in red pine indicates that it has suffered severe bottleneck(s) in population size. This also explains why red pine suffers little inbreeding depression upon selfing, in marked contrast to other pines (e.g., Fowler, 1965
). During the last glaciation, many forest trees were restricted to southern refugia, and one hypothesis has it that red pine lost its genetic variability as it persisted in a southern refugia (Fowler and Morris, 1977
). During the Wisconsinian, glaciers covered nearly all of the present range of red pine, and indeed period-spanning macrofossils have been found in northern Georgia (Jackson et al., 1997
, 2000
; S. T. Jackson, personal communication), indicating that red pine did persist in a southern refugium. However, the loss of genetic variation, or at least lack of genetic variation, may alternatively extend much further back in evolutionary history. Red pine is evolutionarily distinct from all other North American pines except for Pinus tropicalis, a Caribbean endemic, and together these two are the only representatives of the subsection Sylvestres, which contains many Eurasian species. An estimate, based on molecular data, of the age of the split between the red pine lineage and that containing all other North American hard pines apart from P. tropicalis, is 120 million years ago (MYA) (Krupkin et al., 1996
). Moreover, several late Cretaceous fossil species, such as P. clementsii, in North American bear remarkable resemblance to P. resinosa. Pinus resinosa may have existed as a distinct lineage for many millions of years.
Among seven studied populations in the northern half of the range of red pine, one in New Brunswick had much greater diversity for cpSSR loci (Echt et al., 1998
). Coincidentally, among 10 studied populations spread over the southern half of the range, haplotype diversity values for a somewhat different set of cpSSR loci were much larger for three populations in Maine (0.653), New Hampshire (0.277), and Vermont (0.254) than the mean value of 0.028 (Walter and Epperson, 2001
). This suggested but did not prove that there may be an area centered in northern New England where populations have unusually high levels of diversity. This entire area was buried by glaciers during Wisconsin glaciation, and hence cannot be the origin of post-glacial red pine. Such a center of diversity would constitute a contradiction to standard theory about the patterns that evolve when a species spreads from a single origin. If a system has not reached equilibrium, then genetic diversity is greatest at the origin; if it has, then diversity levels should be uniform throughout the system. The situation in red pine appeared to be rather more complex.
Although there may be other factors, the possible role of other refugia, in addition to the one known in the southern Appalachians, would be critical. Even a small second refugial population in the north could have great consequences for the present pattern of genetic variation if it greatly expanded in size as new habitat opened. During the glacial maximum, a series of large islands and extensions of the mainland were exposed due to the reduced level ocean. Some parts of Newfoundland, Gulf of St. Lawrence, and Nova Scotia, as well as then-exposed land that is presently below sea level, was free of all glaciers (e.g., Grant, 1977
; Ives, 1978
; Brookes, 1982
; Josenhans and Lehman, 1999
). Studies of fossil flora from that region and time have rarely if ever been done, but the possibility that a red pine refugium existed there should not be discounted. Moreover, biologists have examined data on the present-day biotic distribution, and have widely accepted that there were well-developed communities in nonglaciated areas of the exposed continental shelf. For example, there are high concentrations of endemic races and subspecies for plants, insects and other animals in Nova Scotia and surrounding areas (e.g., Hamilton and Langor, 1987
; see also review in Pielou, 1992
). Moreover, it appears that red pine reached the previously glaciated New England remarkably early (e.g., Davis, 1983
; Delcourt and Delcourt, 1981
; Watts, 1983
), although researchers rarely distinguish fossil pollen of P. banksiana from P. resinosa.
Evidence of a northern red pine refugium could be gained by examining the geographic distribution of genetic variation in populations in New England and surrounding areas. The present day geographic distributions of haplotypes is related to the histories of distinct lineages, because the chloroplast genome does not recombine (e.g., Neale and Sederoff, 1989
). If the postglacial history of red pine is based on a single southern refugium model, then it seems unlikely that the species gained variation as it spread north or lost variation in the south (due to fragmentation), and then lost the same variants as it spread westward. A more parsimonious explanation is that a second, northern refugium existed. As lineages expanded their ranges from both refugia, they would have formed a contact zone. This contact zone would have greater diversity if the two refugia were genetically differentiated. Thus, a viable hypothesis is that the New England area is an admixture zone. In this study, we examine cpSSR loci for a large number of populations in and around New England in order to determine if indeed there is a center of diversity, and if it has characteristics of an admixture zone. Moreover, we address questions about how gene flow and selection might influence the shape of the zone, as well as the nature of geographic transitions from areas of high diversity to areas with very low diversity.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
100 mg) was cut into small pieces with a razor blade and placed in a 2 mL screw cap centrifuge tube (Molecular Express Ames, Iowa, USA) between a ceramic ball and a ceramic cylinder (both BIO 101 Savant Instruments, Holbrook, New York, USA). A 750-µl cetyltrimethylammonium bromide (CTAB) solution was added, and the samples were ground for 45 s at the 4.5 setting in a FastPrep instrument BIO 101 Savant (BIO 101). After incubation at 65°C for 20 min (lysis), the homogenate was transferred into a centrifuge tube unit and DNA isolation was performed applying the automatic DNA isolation system "AutoGen 850" (AutoGen, Holliston, Massachusetts, USA), at the Genomics Technology Support Facility of Michigan State University. DNA concentrations were measured by fluorometry, and the yield was 10–80 µg/extraction.
Chloroplast markers scored
Sampled trees were genotyped for the same 10 mononucleotide microsatellites reported in our previous study (Walter and Epperson, 2001
). The primers were developed by Vendramin and colleagues (1996)
based on known sequences in Pinus thunbergii. The designations of primer pairs, in terms of original (PT) numbers and the shorter nomenclature used for discussion here (in parenthesis), are: PT9383 (cp2); PT15169 (cp3); PT30204 (cp5); PT36480 (cp6); PT41093 (cp7); PT48210 (cp9); PT71936 (cp12); PT87268 (cp14); PT107148 (cp17); and PT110048 (cp20). The forward primer of each pair was labeled with one of 6- FAM, HEX, or TET (ABI) phosphoramidites. Overlapping allele size ranges were differentiated by using different dye labels in multiplexing.
PCR amplification and multiplexing
All polymerase chain reactions (PCRs) were carried out using 20 ng of template DNA in a total volume of 10 µl. Optimized conditions for all reactions were achieved with a PCR buffer (20 mmol/L Tris-HCl [pH 8.75], 10 mmol/L KCl, 10 mmol/L (NH4)2 SO4, 2 mmol/L Mg SO4, 0.1% Triton X-100, 100 µg/mL bovine serum albumin [BSA], 6% sucrose and 0.1 mmol/L cresol red), 3.5 mmol/L Mg Cl2, 200 _µmol/L each dNTP, 250 _µmol/L each primer, and 0.25 units of AmpliTaq Gold (Perkin Elmer, Wellesley, Massachusetts, USA). The PCR amplifications were performed in an MJ Research (Waltham, Massachusetts, USA) thermocycler Model PTC-100, applying a touchdown amplification protocol: three step cycling with two cycles at 94°C for 1 min, 65°C for 1 min, 70°C for 35 s; 18 cycles at 93°C for 45 s, 64°C for 45 s (with lowering the annealing temperature from 64°C to 55.5°C by the 18th cycle), 70°C for 45 s; and 20 cycles at 92°C for 30 s, 55°C for 30 s, 70°C for 60 s followed by a final extension step at 70°C for 5 min.
Fluorescently labeled PCR products were multiplexed on a 0.40 mm, 6% denaturing polyacrylamide gel (7 mol/L urea) along with a 350-base pair (bp) size standard and an external standard (to control for any possible intergel size variation, which in fact did not occur). Samples were electrophoresed at 2500 V/30 W for 5.5 to 6 h. Labeled microsatellite fragments were run on a 373 ABI (Applied Biosystems, Foster City, California, USA) sequencer, using GeneScan, giving 1-bp resolution. As will be shown, a great majority of trees had the same haplotype, i.e., the same constellation of alleles at all loci. All sampled trees that had variant alleles were retested (re-amplified and genotyped) twice, and in all cases values were consistent. In addition, 256 or 8% of the total (313 trees x 10 loci, or 3130) single-locus genotypic scores for common alleles were designated for random retesting, and again there were zero errors of genotyping.
Statistical analysis
Standard measures of genetic diversity and differentiation were calculated for haplotype frequencies using Lewis and Zaykin's (2001)
GDA program (version 1.0 d16c). For each population sample, the number of haplotypes observed and effective "heterozygosity" or diversity of haplotypes (Weir, 1996
) was calculated. Weir's "
" estimate of the average amount of differentiation of haplotype frequencies among populations was also estimated. Nei's (1978)
measures of genetic distances among pairs of populations were used for the Mantel spatial statistics. Special diversity statistics designed to exploit differences in the numbers of repeats among alleles (e.g., "RST," Goldstein et al., 1995
) were not appropriate for our data set. Polymorphic loci had only two or three alleles, and in the latter cases at least one allele was in very low frequency. Thus there cannot be substantial additional information to exploit, and the additional assumptions, which might be violated, need not be made.
The genetic distance matrix could not be used to construct Unweighted Pair Group Method with Arithmatic Mean (UPGMA) dendrograms, because of various issues, including occurrences of many "ties," where multiple pairs had the same distance values. It is worth noting that loci are not independent, because of the lack of recombination. Loci are completely dependent with respect to dispersal and drift, but not for mutation. Moreover, in our data, essentially all of the information on locus-specific alleles is contained in the haplotypes themselves. In effect, the treatment of haplotypes themselves as "alleles" of a locus is for our purposes consistent with infinite "alleles" (haplotypes) assumptions. Under the infinite alleles mutation model with genetic drift, Nei's measure is linear with time (Nei, 1987
), if there is no migration (although it should be noted that the assumption of no migration is invalid in this as well as most other studies of genetic variation within species).
All spatial statistics were calculated using the PASSAGE program of Rosenberg (2002)
. Moran's I statistics were calculated for each mean-adjusted variable Zi for population sample i, by I = (
ij wijZiZj)/(Wi Zi2), where n is the number of populations, wij is a measure of weight reflecting spatial proximity for a pair of populations i and j, and W is the sum of all weights (Cliff and Ord, 1981
). In the first instance, the wij are binary indicators of inclusion (1) or exclusion (0) of a pair of populations in a particular distance class, or range of geographic distances, using great circle distances separating populations. A set (correlogram) of the unweighted statistics for mutually exclusive and exhaustive distance classes was calculated. To determine if there is finer scale autocorrelation of genetic diversity within distance class one, we calculated weighted I statistics, such that pairs of populations within distance class one are weighted by inverse functions of distance of separation. Spatial directionalities were determined by calculating circular spatial autocorrelation statistics (Oden and Sokal, 1986
), where pairs of populations are classified by distance and compass directions between populations. In all of the above cases, the expected value of I statistics under the null hypothesis of a spatially random distribution is equal to –1/(n – 1), or –0.036. Standard errors and tests of significance were based on the randomization null hypothesis. Spatial statistics could be utilized for only some of the haplotypes, not for those (haplotypes V, VII, VIII, and IX) that occurred in only one population. The geographic distribution of haplotype diversity values was analyzed using the same spatial statistical methods. In addition, Mantel statistics were calculated to measure the overall association of genetic distance with geographic distance (Cliff and Ord, 1981
). In this case, the small Illinois sample (the entire population consisted of four trees, all with haplotype I, and it is located hundreds of kilometers from any other natural stands of red pine) was excluded as potentially having over-influence. The two West Virginia samples were combined in order to have a better representation of the most southeastern region, and for similar reasons the two Pennsylvania population samples were combined.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
|
), with four polymorphic loci (characters), each with ordered states. Haplotypes II, III, IV, VI, and VII all differ from haplotype I by a single base pair change in fragment size at one of the four polymorphic SSRs. In all most parsimonious cladograms, haplotype I is the most interior node, and haplotypes II, III, IV, VI, and VII are all one step to the exterior (Fig. 2). Haplotype V differs from I solely in that its fragment size (162) for cp14 is 2 bp smaller than that (164) for I, and differs from VI solely in that VI has the intermediate size (163); thus V appears to have arisen from VI. Mutational paths to haplotypes VIII and IX, both differing from I by a single base-pair difference at each of two SSRs, cannot be resolved by parsimony alone. For haplotype VIII, VII and IV are equally parsimonious intermediates. Similar roles are represented by II and III in the derivation of IX. However, based on the geographic distribution of haplotypes, VIII is more likely derivative of VI than of IV. The one tree that carried haplotype VIII was found in the same New Hampshire sample in which seven other trees with haplotype VI were found. All of the other 10 trees with haplotype VI were found in one of the isolated populations in West Virginia. In contrast, haplotype IV was found in three trees in total, and two of these were in Ontario, although one was found in a different New Hampshire sample. Similarly, haplotype IX appears somewhat more likely derived from II than from III. Haplotype II is more frequent than III (27 compared to 8 trees in total) and it is more widespread. In the eastern Ontario population (787) in which the only tree with haplotype IX was found, all of the other trees had haplotype I. However, haplotype II was found in the spatially proximal population 746. In contrast, haplotype III was found only in the northeast extreme of the species range (New Hampshire, Maine, and Nova Scotia). The diversity of haplotypes, He, within populations averaged 0.152, but exhibited a wide range. Many population samples had zero diversity, but values ranged up to 0.653, more than four times the mean. The number of haplotypes averaged 1.86, with a range from one to four. The only private haplotypes, i.e., restricted to a single population, were the four that each were found in a single tree, as mentioned above. All populations with high diversity are concentrated into a narrow, band-shaped area centered in northeast New England. The overall measure of haplotype differentiation among populations, , for all haplotypes combined was 0.33 (Table 4), and this value is largely determined by frequency differences in haplotype I. Among separate measures of for each haplotype, values for the four unique haplotypes were near zero, but this is expected because is not well-defined in such cases. Among the other haplotypes, ranged from 0.18 to 0.73. The large value for haplotype VI (0.73) is striking and consistent with the fact that VI is found in only two samples, fixed in the sample from one of the isolated populations in West Virginia, and in high frequency in one of the samples from New Hampshire. Values for haplotypes II and III were both similar to that for I, despite the fact that haplotype II is widespread like haplotype I, but haplotype III is found only in parts of New England.
|
|
|
The Mantel test, measuring the overall association of genetic distance of haplotype frequencies among populations with geographic distances separating populations, was not significant. There is little genetic isolation by distance of the form generated by genetic drift and limited migration. However, the correlogram of Mantel values indicates some spatial structure (Fig. 3). Most notably, there is a large negative and nominally significant (P < 0.01) value for distance class eight (1138– 1412 km).
Directional autocorrelation analyses did not detect trends of haplotype frequencies with latitude or longitude or any other direction. Out of 40 statistics (eight distance/direction classes and five nonprivate haplotypes), only one was nominally significant and none of the five distance/direction correlograms were significant. In contrast, the distance/direction correlogram for diversity values was statistically significant (P < 0.05). The highest correlation was 0.38 (P = 0.001), and it occurred for populations separated by distances from 150 to 600 km (insufficient numbers of populations were separated by less than 150 km given the direction classes, hence directional autocorrelation cannot be calculated for distances less than 150 km) and bearing directions 0° to 60°, or a typical bearing of 30°. This indicates a high level of correlation at short distances along an essentially southwest to northeast axis. This is consistent with the map showing a tilted, band-shaped area of concentration of populations with high diversity along such an axis, in the northeast parts of New England extending northeast into Nova Scotia and the southeastern parts of New Brunswick (Fig. 4).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Walter and Epperson, 2001
). In contrast, much larger numbers of alleles are typically found for cpSSRs in comparable studies of other pine species (Powell et al., 1995
; Provan et al., 1999
). A total of nine haplotypes were found.
The observed genetic differentiation among populations and the spatial pattern of genetic variation both indicate that there is little genetic isolation by distance over the species range. A low level of isolation by distance is broadly consistent with observations in other pines and with the high levels of dispersal of pollen and seed for red pine and typical of pines (Epperson, 2003
). Moran's I statistics and Mantel statistics both showed that there is little spatial autocorrelation of the observed genetic variation in terms of haplotype frequencies. Haplotype I had by far the greatest total frequency (0.869), and it is very widespread, indeed found in nearly every population. The second most frequent haplotype, II (0.060), is also widespread. Four haplotypes, V, VII, VIII, and IX, were each found in only a single tree, each in quite different regions. While it is possible that each one of these occurs in regions outside where it was found, albeit at low frequencies and not sampled, another explanation is that each is a fairly recent mutation. All four are at exterior nodes of the cladogram (Fig. 2). The remaining haplotypes, III, IV, and VI, all have low global frequencies (0.007–0.038) and show various degrees of geographic restriction.
Despite the low amount of genetic isolation by distance, genetic variation exhibits a strikingly nonrandom spatial pattern that indicates limitations to gene flow are controlled by latitude rather than distance. Haplotype diversity values are highly autocorrelated. The Moran statistic for distance class one (259 km) was very large (0.31) and significant, the value for distance two was considerably smaller (0.12), and values for larger distance classes were small and often negative. This is due to a strong concentration of populations with high diversities in parts of New England and immediate surroundings. However, even within this area high diversity values are further concentrated, as indicated by analyses of weighted Moran I statistics within distance class one. Populations with the highest diversity are concentrated into a narrow band (Fig. 4), and the orientation of this is on a southwest to northeast axis, as was determined by directional autocorrelation analysis. The results indicate that there is little gene flow north or south across this band. It is worth pointing out that this area is often not represented in red pine experiments on genetics of morphological traits and provenance trials.
The results also reveal remarkable narrowness of the area and sharpness of its boundaries. The boundaries have not been smoothed as would be expected from even fairly low levels of gene flow. All of the sampled populations in eastern Ontario and southern Quebec have low diversity, as do the even more spatially proximal areas of northern New Brunswick and northern New York. In contrast, all of the populations in the band-shaped area have high levels of diversity, including the one (743) more southeast in New York. In summary, all and only populations with high diversity occur in a narrow band extending from southeastern New York toward the northeast through New Hampshire and Maine, to Nova Scotia and southern New Brunswick. This is wholly inconsistent with equilibrium models of selectively neutral markers under isolation by distance (e.g., Epperson, 2003
), which would predict that diversity level is uniform over the species range. Red pine is either not as near equilibrium as would be expected based on its high levels of dispersal of seed and pollen, or the cpSSRs have been subject to selection (if so, most likely as hitchhiking effects), or both. Selection could cause such a pattern either at equilibrium or with nonequilibrium and cannot be excluded. Nonetheless, it is also likely that the distribution is not near equilibrium despite the fact that high levels of gene flow are expected based on the dispersal characteristics of seed and pollen.
Because the geographic pattern of haplotypes is entirely inconsistent with equilibrium neutral models, explanations should focus on processes that are not at equilibrium. One of the simplest is that the area with relatively high diversity was the origin of red pine as the glacier retreated, and genetic diversity was lost in populations expanding from it. However, this area was buried by the glaciers and thus cannot be the origin. Moreover, it is known that red pine did persist in the south, because macrofossils of red pine, dating from 16 000 and 18 000 yr ago, have been found in northern Georgia. Generally, eastern pines are considered to have survived in the southeastern United States and expanded northward after the glacial maximum. Among the newly sampled populations, it is most telling that haplotype VI was found only in a New Hampshire population, whereas previously it was found only in one of the populations in West Virginia: indeed in that sample all 10 trees were haplotype VI. This strongly suggests that the populations in the high diversity zone have been influenced by the southern refugia.
It appears unlikely that a single southern refugium model could produce the geographic pattern. First, it would require either that red pine gained variation as it moved northward but not in the south or it lost variation in the south after it expanded northward. The latter is reasonable, since decreases in regional population size and fragmentation and isolation clearly have occurred in the southeast. The southeastern populations sampled do have low diversity and relatively high differentiation. Most striking is the fixation of different haplotypes in the samples from the two West Virginia populations that are separated by only 50 km. Second, red pine would had to have lost genetic variation again as it spread from New England westward to its range boundaries. It would have had to lose precisely the same variants it had either gained in the northeast or lost in the southeast during the first phase, which seems unlikely. When red pine spread toward the west it did so as a rapidly expanding regional population (Davis, 1983
), which remains very large today. Throughout the modern northern range of red pine there are millions of red pine trees. Substantial drift and particularly loss of variants are unlikely for a large and rapidly growing regional population.
In contrast, the observed geographic pattern is entirely consistent with there having been two refugia, one in the northeast in addition to that in the southern Appalachians. Thus, while it may be possible to construct other more complicated nonequilibrium processes that are consistent with the data or to invoke various selection models, we focus on this possibility as more parsimonious. Geologists have increasingly emphasized that there were unglaciated areas in the northeast (Grant, 1977
; Ives, 1978
; Brookes, 1982
). Many ecologists have long been convinced that terrestrial ecosystems persisted there as refugia throughout the Wisconsinian (e.g., Fernald, 1911
; Pielou, 1992
). During the glaciation, the level of ocean dropped, exposing a series of islands and extensions of the mainland from Newfoundland extending well south of Nova Scotia. Much of this exposed area (as well as small parts of the present coast line) was not covered by the Laurentide ice sheet, and much of it also was not covered by smaller glaciers. Unfortunately, there has been very little study of fossil flora in this region, much of which is now covered by the Atlantic, and we know of no finds of red pine fossils in this region. However, it is known that period mastodon fossil bones are frequently found by fisherman trawling in areas that were above the Wisconsinian sea level and that mastodons primarily consumed conifers, including pines (Pielou, 1992
). The two refugial populations could have expanded and met in the area where 10-fold greater diversity is found today. It is possible that much of the discrepant paucity of genetic variation in red pine relative to other pines could have been in place well before the Wisconsinian. If so, then there would have been only minor genetic differentiation between the two refugia, caused both by decreasing population sizes as they contracted into the refugia and by mutations that occurred during isolation. Nonetheless, with respect to the genetic diversity of the entire species today, that differentiation would be the major controlling factor. Very recently, other researchers have developed several polymorphic nuclear SSRs in Pinus resinosa, and they found that populations in Newfoundland and certain parts of New Brunswick were highly differentiated from other populations, strongly supporting our interpretation that there was a second, northeast refugium (J. Boys, M. Cherry and S. Dayanandan, unpublished manuscript). Moreover, they found that the diversity of nSSRs was markedly higher in the same general region as the band-shaped area of high cpSSR diversity, supporting the idea that this area is a contact zone.
It appears that there were two refugia, and if both had significant influences on red pine today, then haplotype I was common in both. The fact that haplotype I has very high frequency throughout the range somewhat limits distinctions that can be made between two refugial lineages. Haplotype II is both the second most frequent and is also widespread, being repeatedly found in all regions, the northeast, south, and northwest. If there were two important refugia, haplotype II was probably present in both. In contrast, the distribution of haplotype VI suggests that it was only in the southern refugia. Haplotype VIII is most likely a recent mutant of haplotype VI, and it was found in a single tree, in the same population in New Hampshire. Other haplotypes, V, VII, and IX, were also found in single trees and likely are recent mutants. It is worth noting that haplotype V is very likely (barring a 2-bp mutation) a mutant of VI and was found only in a single tree in central Wisconsin, near the southern limit of the present western range. This suggests that the southern refugia did spread far west. Haplotype VII is almost certainly a mutant from haplotype I and hence could be derived from either refugium. Haplotype IX is more likely a mutant of II and also could be from either refugia. Most informative is that haplotype III is found only in the area of high diversity. It is far more common near the eastern end of the band, in Nova Scotia and eastern Maine, and outside of this more restricted area it was found only in a single tree in New Hampshire. Thus haplotype III would likely be from the northeastern refugium. Haplotype IV is quite interesting in that it was found in the far northwest edge of the species range, and at similar latitudes in New Hampshire and Nova Scotia, suggesting it was derived from the northern refugium. Overall, the results suggest that if a second refugium did exist in the north and then spread, it did so along a northern latitudinal band. The southern refugium would have spread somewhat north and then west along a southern latitudinal band. This would further suggest that the two may differ in their fitness with respect to latitude, as might be expected based on adaptiveness of photoperiodic responses.
In summary, a remarkably narrow and sharply bounded band-shaped area of high diversity was found. One might have expected to find intermediate diversity in nearby regions, for example in southeastern Quebec. Instead there is a sharp drop. Further, northern New Brunswick and similar latitudes in Quebec and eastern Ontario are nearly fixed for haplotype I. If the band of high diversity is an admixture zone, it appears to have remained remarkably tight since forming some 10 000 yr ago, when the pollen record indicates red pine had reached this region (Davis, 1983
). Why would not pollen and seed dispersal have mixed the lineages and smoothed out the admixture area? One answer, consistent with the observed patterns for both haplotype frequencies and genetic variation of photoperiodic and other morphological traits, is that two refugial lineages have spread essentially westward along latitudinal lines, because of latitudinal selection between the lineages. Recent studies that have detected marked genetic differentiation associated with selection mediated environmental factors include Mitton and Duran (2004)
. It should be noted that a potential alternative is that the contact zone is very recent.
|
FOOTNOTES |
|---|
3 Author for reprint requests (Fax: 517 432-1143, Epperson{at}msu.edu
)
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Brenneman W. S. 1956 Red pine indigenous to Illinois?. Journal of Forestry 54: 775
Brookes I. A. 1982 Ice marks in Newfoundland: a history of ideas. Géographie Physique et Quaternaire 36: 139-163
Cliff A. D. J. K. Ord 1981 Spatial processes: methods and applications. Pion, London, UK
Davis M. B. 1983 Quaternary history of deciduous forests of eastern North America and Europe. Annals of the Missouri Botanical Garden 70: 550-563
Delcourt P. A. H. R. Delcourt 1981 Vegetation maps for eastern North America: 40,000 B.P. to the present. In R. Romans [ed.], Geobotany I, 123–166. Plenum, New York, New York, USA
DeVerno L. L. A. Mosseler 1997 Genetic variation in red pine (Pinus resinosa) revealed by RAPD and RAPD-RFLP analysis. Canadian Journal of Forest Research 27: 1316-1320
Dhir N. K. 1973 Geographic variation and relationships among shoot characteristics of red pine seedlings. Minnesota Forestry Research Notes 244: 1-4
Di Rienzo A. A. C. Peterson J. C. Garza A. M. Valdes M. Slatkin N. B. Freimer 1994 Mutational processes of simple-sequence repeat loci in human populations. Proceedings of the National Academy of Sciences, USA 91: 3166-3170
Echt C. S. L. L. DeVerno M. Anzidei G. G. Vendramin 1998 Chloroplast microsatellites reveal population genetic diversity in red pine, Pinus resinosa, Ait. Molecular Ecology 7: 307-316
Epperson B. K. 2003 Geographical genetics. Princeton University Press, Princeton, New Jersey, USA
Fernald M. L. 1911 A botanical expedition to Newfoundland and southern Labrador. Rhodora 13: 109-162
Fowler D. P. 1965 Effects of inbreeding in red pine, Pinus resinosa Ait., II. Pollination studies. Silvae Genetica 14: 12-23
Fowler D. P. D. T. Lester 1970 Genetics of red pine. United States Forest Service, Washington Office Research Paper WO-8
Fowler D. P. R. W. Morris 1977 Genetic diversity in red pine: evidence for low genic heterozygosity. Canadian Journal of Forest Research 7: 343-347
Goldstein D. B. A. R. Linares L. L. Cavalli-Sforza M. W. Feldman 1995 An evaluation of genetic distances for use with microsatellite loci. Genetics 139: 463-471[Abstract]
Grant D. R. 1977 Glacial style and ice limits, the Quaternary stratigraphic record, and changes of land and ocean level in the Atlantic provinces, Canada. Géographie Physique et Quaternaire 31: 247-260
Guries R. A. Ager 1980 Red pine seedling seed orchards: 10 year results. Forest Research Notes, University of Wisconsin 242: 1-4
Hamilton K. G. A. D. W. Langor 1987 Leafhopper fauna of Newfoundland and Cape Breton Islands (Rhynchota: Homoptera: Cicadellidae). Canadian Entomologist 119: 663-695[ISI]
Hamrick J. L. M. J. W. Godt 1990 Allozyme diversity in plant species. In A. H. D. Brown, M. T. Clegg, A. L. Kahler, and B. S. Weir [eds.], Plant population genetics, breeding and genetic resources, 43–63. Sinauer, Sunderland, Massachusetts, USA
Hamrick J. L. M. J. W. Godt S. L. Sherman-Broyles 1992 Factors influencing levels of genetic diversity in woody plant species. New Forests 6: 95-124[CrossRef]
Ives J. D. 1978 The maximum extent of the Laurentide ice sheet along the east coast of North America during the last glaciation. Arctic 31: 24-53[ISI]
Jackson S. T. J. T. Overpeck T. Webb S. E. Keattch K. H. Anderson 1997 Mapped plant-macrofossils and pollen records of late Quaternary vegetation change in eastern North America. Quaternary Science Reviews 16: 1-70
Jackson S. T. R. S. Webb K. H. Anderson J. T. Overpeck T. Webb J. W. Williams B. C. S. Hansen 2000 Vegetation and environment in eastern North America during the last glacial maximum. Quaternary Science Reviews 19: 489-508
Johnson L. C. W. B. Critchfield 1974 A white-pollen variant of bristle cone pine. Journal of Heredity 65: 123
Josenhans H. S. Lehman 1999 Late glacial stratigraphy and history of the Gulf of St. Lawrence, Canada. Canadian Journal of Earth Sciences 36: 1327-1345[CrossRef][ISI]
Krupkin A. B. A. Liston S. H. Strauss 1996 Phylogenetic analysis of the hard pines (Pinus subgenus Pinus, Pinaceae) from chloroplast DNA restriction site analysis. American Journal of Botany 83: 489-498[CrossRef][ISI]
Lewis P. O. D. Zaykin 2001 Genetic data analysis: computer program for the analysis of allelic data, version 1.0 (d16c). Computer program distributed by the authors, website: http://lewis.eeb.uconn.edu/lewishome/software.html
Little E. L. 1971 Atlas of United States trees, vol. 1, Conifers and important hardwoods. USDA Forest Service Miscellaneous Publication No. 1146
Mitton J. B. K. L. Duran 2004 Genetic variation in piñon pine, Pinus edulis, associated with summer precipitation. Molecular Ecology 13: 1259-1264[CrossRef][Medline]
Mosseler A. K. N. Egger G. A. Hughes 1992 Low levels of genetic diversity in red pine confirmed by random amplified polymorphic DNA markers. Canadian Journal of Forest Research 22: 1332-1337
Mosseler A. D. J. Innes B. A. Roberts 1991 Lack of allozymic variation in disjunct Newfoundland populations of red pine (Pinus resinosa). Canadian Journal of Forest Research 21: 525-528[CrossRef]
Neale D. B. R. R. Sederoff 1989 Paternal inheritance of chloroplast DNA and maternal inheritance of mitochondrial DNA in loblolly pine. Theoretical and Applied Genetics 77: 212-216[ISI]
Nei M. 1978 Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89: 583-590
Nei M. 1987 Molecular evolutionary genetics. Columbia University Press, New York, New York, USA
Oden N. L. R. R. Sokal 1986 Directional autocorrelation: an extension of spatial correlograms to two dimensions. Systematic Zoology 35: 608-617[CrossRef][ISI]
Pielou E. C. 1992 After the Ice Age. The return of life to glaciated North America. The University of Chicago Press, Chicago, Illinois, USA
Powell W. M. Morgante R. McDevitt G. Vendramin J. A. Rafalski 1995 Polymorphic simple sequence repeat regions in chloroplast genomes: applications to the population genetics of pines. Proceedings of the National Academy of Sciences, USA 92: 7759-7763
Provan J. N. Soranzo N. J. Wilson D. B. Goldstein W. Powell 1999 A low mutation rate for chloroplast microsatellites. Genetics 153: 943-947
Rosenberg M. S. 2002 PASSAGE. Pattern analysis, spatial statistics, and geographic exegesis, version 1.0. Department of Biology, Arizona State University, Tempe, Arizona, USA
Simon J.-P. Y. Bergeron D. Gagnon 1986 Isozyme uniformity in red pine (Pinus resinosa) in the Abitibi Region, Quebec. Canadian Journal of Forest Research 16: 1133-1135[CrossRef]
Swofford D. L. 2002 PAUP*. Phylogenetic analysis using parsimony (*and other methods), version 4. Sinauer, Sunderland, Massachusetts, USA
Vaartaja O. 1962 Ecotypic variation in photoperiodism of trees with special reference to Pinus resinosa and Thuja occidentalis. Canadian Journal of Botany 40: 849-856
Vendramin G. G. L. R. Lelli P. Rossi M. Morgante 1996 A set of primers for the amplification of 20 chloroplast microsatellites in Pinaceae. Molecular Ecology 5: 595-598[CrossRef][Medline]
Walter R. B. K. Epperson 2001 Geographic pattern of genetic variation in Pinus resinosa: area of greatest diversity is not the origin of postglacial populations. Molecular Ecology 10: 103-111[CrossRef][Medline]
Watts W. A. 1983 Vegetational history of the eastern United States 25,000 to 10,000 years ago. In H. E. Wright, Jr. [ed.], Late-Quaternary environments of the United States, vol. 1, The Late Pleistocene, 294–310. University of Minnesota Press, Minneapolis, Minnesota, USA
Weir B. S. 1996 Genetic data analysis II. Sinauer, Sunderland, Massachusetts, USA
Wright J. W. R. A. Read D. T. Lester C. Merritt C. Mohn 1972 Geographic variation in red pine: 11-year data from the North Central states. Silvae Genetica 21: 205-222
Yao Y. N. 1974 Improved red pine for Michigan. Quarterly Journal of Chinese Forestry 7: 19-29
This article has been cited by other articles:
|
|
 |
|
  F.-L. Shih, S.-Y. Hwang, Y.-P. Cheng, P.-F. Lee, and T.-P. Lin Uniform genetic diversity, low differentiation, and neutral evolution characterize contemporary refuge populations of Taiwan fir (Abies kawakamii, Pinaceae) Am. J. Botany, February 1, 2007; 94(2): 194 - 202. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Boys, M. Cherry, and S. Dayanandan Microsatellite analysis reveals genetically distinct populations of red pine (Pinus resinosa, Pinaceae) Am. J. Botany, May 1, 2005; 92(5): 833 - 841. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME |
