Range-wide genetic assessment of F. pennsylvanica reveals a high frequency of parentage from Fraxinus cultivars and extensive admixture between F. pennsylvanica and F. americana

doi:10.21203/rs.3.rs-1359299/v1

PDF

Research Article

Range-wide genetic assessment of F. pennsylvanica reveals a high frequency of parentage from Fraxinus cultivars and extensive admixture between F. pennsylvanica and F. americana

https://doi.org/10.21203/rs.3.rs-1359299/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted 21 Mar, 2022

You are reading this latest preprint version

Conservation of the gene pool of F. pennsylvanica (green ash), the species most susceptible to emerald ash borer (EAB), requires a range-wide genetic assessment of the population dynamics to guide seed collection efforts. An assessment of gene flow into natural stands from ubiquitously distributed green ash and F. americana (white ash) cultivars is also essential. We used 16 EST-SSR markers to genotype 1291 trees from 48 naturally regenerated populations of green ash, 19 green ash cultivars and 10 white ash from a Fraxinus species collection. We did not find evidence for latitudinal gradients in allele richness or higher differentiation in northern range edge populations. Analysis of population substructure revealed two major groups of green ash, one dominant in northwestern locations, one dominant in southern locations and extensive admixture in midwestern and eastern locations. We identified a third group as white ash, including 140 individuals with interspecific admixture in eastern and southern populations. We detected cultivar parentage with high confidence in 172 individuals in 34 of the 48 populations. We conclude that high dispersal capacity, broad adaptability, and gene flow from many cultivars into predominately northern regions, plays a role in the comparatively low regional population substructure in the Northwest. The high frequency of cultivar propagule dispersal and extent of interspecific admixture detected in our study suggests that under rapidly changing conditions, local populations may or may not be locally adapted and standing genetic variation may not be of long standing.

Parentage

cultivar

provenance

admixture

Fraxinus pennsylvanica

Fraxinus americana

Green ash (F. pennsylvanica Marsh), one of the most wide-ranging of the North American Fraxinus, occurs in a multiplicity of forested ecosystems across Eastern and central North America (Poland & McCullough 2006). High phenotypic plasticity, cold tolerance, salt tolerance, ease of clonal propagation, rapid growth, an attractive canopy, and prior to the accidental introduction of the emerald ash borer (Agrilus planipennis Fairmaire), few serious insect pests, made green ash desirable species for urban landscaping, shelterbelts, and riparian buffers (Schultz et al. 2004; MacFarlane & Meyer 2005). Grafted clones of green and white ash cultivars were widely planted as street and park trees throughout the United States and Canada to replace the American elms lost to Dutch elm disease (Poland & McCullough 2006; Raupp et al. 2006). Trees from improved green ash germplasm for shelterbelts and buffer zones were widely planted in rural areas in the United States and Canada, starting in the 1930s (Dawson & Read 1964; Cunningham 1988). An understanding of population dynamics across local and regional scales, including an assessment of gene flow from cultivars and the extent of admixture between sympatric Fraxinus species is essential for designing cost-effective strategies for seed collection, informing strategies for EAB interventions and designing restoration projects.

Propagule dispersal mechanisms in Fraxinus suggest that gene flow among populations will be high. Fraxinus fruits are samaras, indehiscent winged achenes that enable both anemochorous (wind) and hydrochorous (water) seed dispersal. Green ash samaras float for at least two days and maintain viability after immersion in water for over two weeks (Schmiedel & Tackenberg 2013). Hydrochorous seed dispersal is the most likely mechanism for the exceptionally rapid spread of green ash in Central European floodplain forests where this species is not native; more than 970 km/year in some regions (Schmiedel 2010). Rapid hydrochorous seed dispersal, coupled with anemochorous seed and pollen dispersal, could potentially minimize local differentiation while maintaining high standing genetic variation across broad regional scales. Populations of F. excelsior (European ash) in Britain and France show minimal differentiation (F_st = 0.025), suggesting extensive propagule exchange across broad geographical regions (Sutherland et al. 2010). A similar study in Ireland also detected very low differentiation, little indication of inbreeding and high genetic diversity throughout the island (Beatty et al. 2015). However, a larger population study across most of the range of F. excelsior, while supporting regional panmixia among British, western European, and central European populations, found strong genetic differentiation between the three Swedish populations and the southeast European populations (Heuertz et al. 2004). A more detailed study of far northern range edge populations revealed high population differentiation and loss of genetic diversity relative to the more southern populations, the expected signal of postglacial colonization (Tollefsrud et al. 2016). Based on these data, it might be reasonable to assume that the range-wide population dynamics of F. pennsylvanica would be similar. However, the native range of F. pennsylvanica lacks the altitudinal and coastal heterogeneity present within the range of F. excelsior. Patterns of glacial advance and retreat in North America were different than those in Europe and Neolithic human impacts on the landscape differed substantially from those in Europe (Jackson et al. 2000; Veloz et al. 2012; Izumi & Bartlein 2016). The absence of impassable geographical barriers in the central and eastern United States and Canada and the high dispersal capacity of Fraxinus suggests that panmixia, with little genetic differentiation even at the range edges, may be present across most of the range. Alternatively, the sharp climate contrasts between the Great Plains, the Gulf coasts and the Atlantic coasts may have resulted in regional differentiation as a result of adaptive variation.

Green ash clonal cultivars could potentially swamp local provenances with nonlocal pollen and seed sources, especially at the edges of native range, where populations of local origin are small and widely scattered. Although studies of assisted gene flow among conspecific natural populations have attracted interest as an adaptive forest management strategy (Girardin et al. 2021), studies of gene flow from forest tree cultivars into wild conspecifics have primarily focused on the impact of plantation forestry on native gene pools. (Sampson & Byrne 2008; Steinitz et al. 2012; Ramírez-Valiente & Robledo-Arnuncio 2015; Nishio et al. 2021). Given the very extensive use of male green and white ash cultivars in urban forestry and the dispersal capacity of both pollen and seed, gene flow into natural populations is certainly possible and may be extensive.

Consensus on the taxonomy of the Fraxinus in North America remains elusive, but F. pennsylvanica and F. americana are universally recognized based on diploidy and morphology (Whittemore et al. 2018). Some authors assign other species names to polyploid trees having morphological characteristics similar to F. americana (Nesom 2010; Campbell 2017; Whittemore et al. 2018). However, difficulty arises in field collections, as morphology may not reveal ploidy and even the distinction between green and white ash is not clear in some sites. Both species may occur in the same site and are sympatric over most of their respective native ranges (Little & Elbert 1979). Verbal reports of individual interspecific hybrids in natural settings are frequent; three authors report viable seed from experimental crosses (Wright 1953; Dugal 1971; Taylor 1973). Cimmzam Cimmaron®, one of the cultivars included in this study, although patented as a green ash cultivar, has a leaf scar more typical of white ash than green ash, suggesting possible admixture with white ash (Zampini 1992).

In this investigation we genotyped 48 naturally regenerated populations of green ash (1291 trees) and 19 ash cultivars with 16 EST-SSR markers to assess population differentiation, population substructure, and gene flow from cultivars. We included 10 white ash individuals from a species collection we had genotyped previously to enable detection of misidentification and identify possible hybrids (Noakes et al. 2014). We found broadly regional population substructure consisting of a Northwestern/Northern group and a Southern group largely coincident with northwestern/northern and southern locations and frequent admixture between these groups in the central and eastern United States. We detected parentage from cultivars in 34 of the 48 populations and extensive cultivar parentage (23–50%) in eight populations. Analysis of population substructure also revealed 140 cases of admixture between green ash and white ash and three sites containing only white ash. We discuss the implications of our findings for conservation of the existing gene pool of the North American Fraxinus, now threatened by the inexorable advance of EAB.

2.1 Study area and species characteristics

The native range of green ash spans more than 23 degrees of latitude and 45 degrees of longitude, extending from Cape Breton Island in the Atlantic Ocean westward to Alberta, Canada and southward to the Gulf Coast of the Eastern United States (Little & Elbert 1979; Prasad & Iverson 2003). In the Great Plains, green ash is locally abundant in the Temperate Prairie, West-central Semi-arid Prairie and South-central Semi-arid Prairie ecoregions. In the Eastern Temperate and Northern Forests ecoregions, green ash is a significant component of eight forested ecosystems types, including white-red-jack pine in the Great Lakes states and provinces, loblolly-shortleaf pine in the Gulf coastal plains and the Piedmont, oak-pine in the Appalachians, oak-hickory in the more mesophytic areas of the central and eastern United States, oak-gum-cypress in southern bottomlands, elm-ash-cottonwood in seasonally flooded bottomland and aspen-birch on glacial till in cold, moist climates (Garrison et al. 1977). Green ash woodlands can also persist in seasonally dry creek beds, upland forests, and seasonally dry urban environments across the entire native range.

Unlike many of the Fraxinus species, green ash is strictly dioecious (each tree is either male or female), a characteristic that could predispose small local populations to invasion from propagules of clonal cultivars if sex ratios became severely unbalanced. Green ash is strictly diploid (n = 23) while white ash (Fraxinus americana L), a closely related species sympatric with green ash in the central and eastern United States, includes both diploid (n = 23) and polyploid individuals, although many authors assign different species names to the polyploids (Hardin 1974; Black & Beckmann 1983; Nesom 2010; Whittemore et al. 2018).

2.2 Sample collection design

We collected leaf or twig samples from adult trees (> 10 cm DBH) determined to be green ash based on taxonomic traits and habitat in 48 naturally regenerated sites, using nested sampling when possible, to identify site-specific processes (Table 1, Fig. 1). All nested populations were between 10 and 53 kilometers from one another. We did not collect samples in southern Michigan, northern Indiana, or northern Ohio, as surviving adult trees were unlikely to be representative of the genetic diversity existing before EAB invasion. Most of the loss, now estimated at more than 100 million trees, has occurred in this region and in the Great Lakes region of southern Ontario (Liu 2018). Populations were identified as those trees occurring within approximately a square kilometer of naturally regenerated forest or woodland. Maps shown are in the USA_Contiguous_Albers_Equal_Area_Conic_USGS_version projection.

Table 1

Locations and descriptive statistics for the 48 populations genotyped.
Population	Identifier	N	Latitude	Longitude	dLGM (km)^a	A_r^b	F_ST^c	P_a^d
Saskatchewan 1	SK1	24	50.79	-103.89	343.7	6.6	0.104	1
Manitoba 2	MB2	26	50.56	-96.63	410.7	8	0.065	3
Alberta 1	AB1	27	50.02	-110.75	216.5	6.4	0.111	1
Manitoba 1	MB1	27	49.69	-95.28	432.4	6.6	0.11	1
Quebec 1	QB1	35	48.68	-71.90	227.1	6.3	0.111	2
Minnesota 2	MN2	28	47.63	-94.61	250.7	7.6	0.088	2
Minnesota 1	MN1	32	47.07	-93.93	238.7	7.1	0.085	0
North Dakota 1	ND1	27	46.83	-100.83	-106.1	6.4	0.12	0
North Dakota 2	ND2	27	46.80	-100.81	-109.6	6.3	0.099	1
Nova Scotia 3	NS3	32	46.72	-60.93	128.6	6.4	0.217	0
Montana 1	MT1	35	46.37	-105.06	-49.4	6.4	0.119	2
Montana 2	MT2	24	46.34	-105.07	-45.5	5.7	0.158	1
New Brunswick 2	NB2	25	45.95	-66.88	233.6	6.3	0.216	0
New Brunswick 1	NB1	29	45.87	-66.51	230.5	6.3	0.138	2
South Dakota 2	SD2	29	45.74	-97.23	75.7	6.7	0.074	0
South Dakota 1	SD1	25	45.65	-97.13	90.5	7	0.099	0
Quebec 3	QB3	30	45.53	-76.02	590.8	6.9	0.082	1
Wisconsin 2	WI2	27	45.31	-88.56	171	8.1	0.103	4
Wisconsin 1	WI1	27	45.13	-88.44	174	7.9	0.057	0
Vermont 1	VT1	27	44.97	-73.27	276.2	5.9	0.111	1
Ontario 1	ON1	25	44.94	-73.13	191.5	9.2	0.032	6
Nova Scotia 2	NS2	24	44.90	-63.84	124.8	7.3	0.081	3
Wyoming 1	WY1	12	44.77	-104.67	-205.9	5.1	0.112	0
Vermont 2	VT2	24	44.74	-73.26	293.8	6.6	0.077	0
Wyoming 2	WY2	28	44.53	-104.08	-235.6	6.8	0.092	2
Iowa 1	IA1	28	42.04	-93.60	-9.8	6.3	0.128	4
Iowa 2	IA2	31	42.04	-93.64	-8.8	6.4	0.088	0
Nebraska 2	NE2	27	41.30	-102.12	-522	6.1	0.131	0
Nebraska 1	NE1	30	41.24	-101.67	-551.4	5.8	0.126	1
Illinois 2	IL2	29	40.96	-89.43	-19.3	6.9	0.084	0
Illinois 1	IL1	30	40.44	-89.96	-79.1	7.1	0.099	1
Ohio 1	OH1	30	40.12	-82.97	-13.6	9.1	0.069	3
Ohio 2	OH2	24	40.01	-82.84	-7.6	7.7	0.093	1
Missouri 2	MO2	30	39.01	-92.77	-341.4	7.8	0.072	4
Missouri 1	MO1	29	38.87	-92.30	-323.4	7.8	0.152	3
Virginia 2	VA2	27	38.83	-77.50	-363.6	8.7	0.071	4
Kansas 1	KS1	29	38.72	-94.96	-359.1	7.1	0.087	4
Virginia 1	VA1	29	37.62	-78.21	-392.4	8.4	0.069	1
Tennessee 1	TN1	29	35.69	-83.50	-450.6	6	0.238	2
Oklahoma 1	OK1	10	35.40	-95.59	-700.2	4.3	0.164	0
South Carolina 2	SC2	17	33.14	-79.82	-718.7	5.6	0.155	2
South Carolina 1	SC1	28	32.95	-79.75	-738.3	7.2	0.115	2
Mississippi 1	MS1	26	32.82	-90.81	-851.9	6	0.154	0
Mississippi 2	MS2	27	32.71	-90.79	-840.7	7.6	0.112	2
Texas 1	TX1	21	30.56	-95.75	-1171.6	6.6	0.1	0
Louisiana 1	LA1	29	30.41	-91.67	-1092.7	8	0.094	3
Louisiana 2	LA2	24	30.39	-89.73	-1075.5	6	0.146	0
Texas 2	TX2	31	29.17	-95.50	-1305.4	6.7	0.142	0
^aDistance from the LGM: negative number site outside of LGM, positive number site within LGM, ^bAllele richness, ^cPopulation-specific F_ST, ^dPrivate alleles

Cultivars were from the collection at the Minnesota Landscape Arboretum. A cultivar checklist made in 1983 listed 23 valid F. pennsylvanica cultivars (Santamour Jr & McArdle 1983). We initially evaluated 20 named cultivars (Online Resource 1), 12 of which were on the 1983 list and eight of which were named after this date. ‘Lednaw Aerial’ was an exact match to ‘Summit’. We discovered later that ‘Lednaw Aerial’ is a bud sport (a somatic mutation) of ‘Summit’, which explained their identical genetic profiles. The majority of these cultivars originated from northern states or provinces (Online Resource 1). The 10 white ash included as comparators were identified as white ash based on morphological characteristics, primarily the shape of bud scar and dormant bud, and ITS sequences.

2.3 DNA Extraction, PCR, and Genotyping.

DNA extraction techniques and PCR protocols were as previously described (Noakes et al. 2014). We used 16 EST-SSR markers developed previously for Fraxinus to identify genetic variation within and among populations (Noakes et al. 2014; Lane et al. 2016). Fifteen were chosen from the green ash transcriptome and one from F. excelsior (GenBank: FR639289.1). Amplicon length was measured with an ABI 3730xl capillary electrophoresis device (Applied Biosystems, Foster City, CA) and the resulting genotypes were scored with GeneMapper version 4.1 (Applied Biosystems).

2.4 Statistical analyses

We employed GenAlEx version 6.512b2 (Peakall & Smouse 2006, 2012) to detect zygotic and gametic disequilibrium, perform principle coordinate analysis (PCA) based on genetic distances, detect private alleles, assess isolation-by-distance, and F-statistics (Smouse et al. 1986). Where appropriate, we corrected for multiple comparisons using the Holm-Bonferroni method (Holm 1979). We used the Bayesian approach implemented in Geste 2.0 to estimate population-specific differentiation, permitting evaluation of spatial factors as predictor variables for the genetic distinctness of each population relative to all the other populations (Foll & Gaggiotti 2006). Analysis parameters used were 100 pilot runs, a sample size of 10,000, thinning interval of 20, and a burn-in of 50,000.

Evidence for reduction of effective population size was examined with Bottleneck version 1.2.02 using the Wilcoxon sign rank test with 10⁴ replications, 0 for the proportion of single mutation model, and 36% for the variance of geometric distributions in the two-phase model with the recommended settings for microsatellite data (Cornuet & Luikart 1996; Luikart et al. 1998; Piry et al. 1999). A test was also done for a mode shift in allele frequency distribution in each population with the same application. The M-ratio, the ratio of the number of alleles to the range of allele sizes, can detect bottlenecks that occurred further back in the population history than the tests described above, provided that the assumptions are met (Garza & Williamson 2001).

Population structure and admixture were estimated with Structure version 2.3.4 (Pritchard et al. 2000) with 50,000 MCMC burn-in iterations followed by 100,000 MCMC iterations. All of the data, (populations, cultivars and the white ash comparators) were included. We initially tested K (the number of proposed genetic groups) from two to 20, with 10 replicates for each K. The K at which the data were most likely was inferred using the Evanno method as implemented in Structure Harvester (Evanno et al. 2005; Earl & vonHoldt 2012). Individuals were counted as admixed if the estimated proportion of membership in a single group was < 0.85. A value > 0.85 falls within average 95% confidence intervals estimates for data ranging from 0.85 to 0.99 in this analysis.

Parentage analysis was done with Cervus version 3.0.7 (Marshall et al. 1998; Kalinowski et al. 2007). The mean polymorphic information content(PIC) of the data used to assign parentage was 0.7418 and the combined exclusion probability with one parent known exceeded 0.9999. These values are similar to a parentage analysis of cultivated Pacific oysters in which the authors found that the combined exclusion power of 12 microsatellites for identification of one parent was 1 (100% correct identification of the true parent) while more than 50 SNPs were required for the same result.(Liu et al. 2017). The settings for parentage analysis simulations for the cultivars were 10,000 offspring, 19 candidate parents for the cultivar parentage analysis, 0.05 proportion of candidate pollen parents sampled, 0. 05 proportion of loci mistyped, and a minimum of 12 loci typed, yielding a critical LOD score of 2.14 for a strictly confident parent/offspring pairs (95% confidence).

3.1 Descriptive statistics

Zygotic disequilibrium was not detected in 25 of the of the 42 populations. Of the 17 populations remaining, disequilibrium was detected at one marker only in 12 populations and at two markers only in five populations (MN2, IL1, IL2, MO1, and MS2). Less than 1% of the gametic disequilibrium tests had significant p-values after correction for multiple comparisons. Allele richness ranged from 4.3 in the Oklahoma 1 population to 9.2 in the Ontario 1 population (Table 1).

3.2 Measures of differentiation and bottlenecks

Molecular variance within populations accounted for 70% of the total, followed by the variance within individuals (16%) and last, the variance among populations (14%). Pairwise F_ST values ranged from 0.013 (SK1-AB1) to 0.204 (ON1-MO2). We found a moderate but significant signature of isolation-by-distance (R² = 0.15, p = 0.001). Allelic richness showed no indication of a latitudinal gradient (R² = 0.012, p = 0.933), a longitudinal gradient (R² = 0.052, p = 0.119), a relationship with distance from the last glacial maximum (R² = 0.024, p = 0.297) or a relationship with number of detected cultivar parents (R² = 0.032, p = 0.117). Population-specific F_ST estimates, a measure of the differentiation of each population from all of the others, ranged from 0.03 in Ontario 1 to 0.24 in Tennessee 1. The three populations with the highest population-specific F_ST estimates (TN1, NS3, and NB2) had neither the lowest level of allele richness nor the highest number of private alleles. Posterior model probabilities for an association of population-specific F_ST estimates with latitude (P = 0.052) and longitude (P = 0.023) did not suggest latitudinal or longitudinal gradients. The distance from the last glacial maximum (dLGM) had the highest posterior probability (P = 0.111) but this value was lower than the minimum (0.15) recommended (Foll & Gaggiotti 2006).

All 48 populations fit the two-phase model (TPM) for mutation drift equilibrium and 47 of the 48 populations had the expected L-shaped distribution of allele frequencies, suggesting no recent population bottlenecks. The Oklahoma 1 population consisted of only 10 trees, a sample size well below the 20–30 recommended for the bottleneck test (Luikart et al. 1998). We calculated M-ratios on the ten dinucleotide repeat markers alone. None of the 48 populations had an M-ratio above the critical value (0.68).

3.2 Genetic structure and population admixture

The data were most likely at K = 3 genetic groups when assessed with STRUCTURE (Fig. 2, Fig. 3a). Two groups mapped on two broad geographical regions, the northwestern and the southern part of the native range. The third group has a geographically dispersed distribution, primarily in the eastern part of the native range. When visualized at K = 4, the Northwestern group splits into two groups, Northwestern and Northern (Fig. 3b), while the remaining two groups remain largely unchanged. A high frequency of individuals with admixture from all three groups occurred in the Ohio and Virginia sites and to a lesser extent in the Ontario site. Individuals in three sites (NS3, NB3 and TN1) cluster in group three, with minimal admixture. The MO1 site contains mostly unadmixed group three individuals, while the WI2 site contains group three individuals with some admixture from the Northwestern group. Examination of the admixture in the 13 local population pairs revealed 10 pairs with similar admixtures and three discordant pairs: New Brunswick, Wisconsin, and Missouri.

The first and second principle coordinates of the PCA analysis (Online Resource 2) explain 29.09% and 17.19% of the observed variation respectively. Four clusters are evident, the minimally admixed third group that we identify as white ash (see below), a moderately to highly admixed group, a minimally admixed Southern group and minimally admixed Northwest group.

3.3 White ash and white ash admixture

Seven of the 10 presumed white ash comparators were unadmixed members of the third group (Fig. 4). While this does not prove that the third group is white ash, it is strongly suggestive. Trees FA_7 and FA_9 grouped with Northwestern/Northern green ash while tree FA_3 is admixed. This result indicates that identification by morphology and ITS sequence does not necessarily detect admixture or even species in some cases, despite expert close examination.

If we accept group three as white ash and classify individuals having 15–85% estimates for white ash as admixed, then 140 of the total population (10.8%) are white-ash-green ash admixtures. The incidence of admixture is unevenly distributed. Twenty-two populations have no admixed individuals, while the majority of sampled individuals at the NS2 (88%), SC1 (68%), and VA2 sites (63%) are admixed. A high incidence of admixture also occurs at the ON1 (48%), SC2(47%), OH2 (42%), WI2 (37%), and VA1 sites (31%). Following our criteria, all of the individuals sampled at the NS3, NB2, and TN1 sites are white ash.

3.4 Cultivars and cultivar parentage

Fourteen cultivars from northern sources dominate this collection: five from Minnesota, five from North Dakota, two from Iowa, one from Alberta and one from Wisconsin. Fifteen of these 19 cultivars are in the Northwestern/Northern genetic group, with minimal admixture of the other two groups (Fig. 4). Cimmzam Cimmaron® is in the white ash group. ‘Emerald’, ‘Hollywood’ and ‘Centerpoint’ are admixtures of the Northwest and Southern groups. The mean polymorphic information content of the data used to assign parentage (PIC) was 0.7418 and the combined exclusion probability exceeded 0.9999. Parentage analysis detected 173 high confidence parent-offspring matches in which one of the parents was identified as one of the 19 cultivars (Table 2). One high-confidence parent offspring match was between the two cultivars ‘Johnson Leprechaun’ and ‘Honeyshade’. The frequency of cultivar parentage was highest in the Northwest, the region where most of the cultivars originated. However, widespread planting of clonally propagated green ash landscaping cultivars in locations far distant from the reported origin is revealed in the distribution of some cultivar offspring (Table 2). ‘Leeds Prairie Dome’, reported as originating in North Dakota, was identified as the parent of 25 progeny distributed across seven states and three provinces, from southern Saskatchewan to northern Quebec and south to western Nebraska. ‘Emerald’, reported as originating from Arlington, Nebraska was identified as the parent of six progeny, from Manitoba to Mississippi. Cimmzam Cimmaron’ was identified as the parent of one tree in each of five sites (NS3, NB2, ON1, NE2 and OH2 (Table 2). Only ‘Hollywood’ has no parent-offspring matches.

Table 2

Number (N) and incidence (I) of cultivar parentage by population and cultivar
	Berg	CC	CP	Em	HF	HS	HW	Jwl	JL	Kin	LPD	Man	MS	New	Pat	RPS	SD	Sum	WDC	N	I
SK1	1					2					1			3		1		3	1	12	0.5
MB2	1			1									1							3	0.12
AB1	1							1			2		2	1	1			3		11	0.42
MB1						1									4					5	0.19
QB1	1			1							4					2		1		9	0.32
MN2																			3	3	0.1
MN1	1					1		1			1			1					2	7	0,23
ND1			1		2	1					4					1				9	0.29
ND2	6				1			2	1		2							1		13	0.41
NS3		1																		1	0.03
MT1	1					1					2			2		4	1			11	0.32
MT2								1			3		2	2				1		9	0.26
NB2		1																		1	0.03
NB1						1										2		1		4	0.11
SD2						3							1			2		1		7	0.18
SD1					1						1		2			1				5	0.13
QB3	1								2		1					1		3		8	0.2
WI1									1					1		1			1	4	0.1
VT1	1			1	1				1		1							1		6	0.14
ON1		1																		1	0.02
WY1																		1		1	0.02
WY2	1			1				1			1							1		5	0.11
IA2								1	1	1		1	4	1					1	10	0.22
NE2		1				1					2					2				6	0.13
NE1						2										1			1	4	0.08
IL2						2		1					2							5	0.1
IL1				1																1	0.02
OH2		1																		1	0.02
MO1												1								1	0.02
KS1			1		2								1							4	0.08
OK1			2							1										3	0.06
MS1				1																1	0.02
LA1			1																	1	0.02
N	15	5	5	6	7	15	0	8	6	2	25	2	15	11	5	18	1	17	9

Cultivar parentage was not detected in 14 of the 48 sites (WI2, NS2, VT2, IA1, OH1, MO2, VA2, VA1, TN1, SC1, SC2, MS2, TX1, TX2). Cultivar parentage comprised 23–50% of the individuals tested at eight sites, seven located in the northwestern region of the range (SK1, AB1, MN1, ND1, ND2, MT1 and MT2) and one (QB1) in the northeast (Table 2). Twelve individuals of the 24 sampled at the Saskatchewan site had parentage from seven different cultivars. As most of the cultivars examined are males (Online Resource 1), offspring likely resulted from pollen flow into the population. We infer that the offspring of the female trees ‘Jewel’ and ‘Mandan’ are the result of pollen flow from naturally regenerated trees followed by seed dispersal back into the stand. We found no trees that were exact or fuzzy matches to the cultivars and no trees with both parents from the set of 19 cultivars.

We initially designed our range-wide study to assess the population differentiation and population substructure of a widely distributed, ecologically important native tree species acutely threatened by the emerald ash borer. We included an assessment of cultivar parentage as the propagule dispersal capacity of Fraxinus is well documented and green and white ash cultivars were extensively planted in urban and agricultural environments over the last 70–80 years. We expected to see low differentiation among green ash populations across broad regional scales, given the species high dispersal capacity, high phenotypic plasticity and high tolerance to abiotic stress. While we observed higher population-specific F_st values than those reported for F. excelsior, evidence for spatial gradients was lacking. We expected to detect evidence of gene flow from cultivars into local stands in the northwest region of the natural range where natural stands are sparsely distributed and where most of the cultivars in this study originated, but we did not expect the frequency of cultivar parentage we detected. We included a small group of white ash comparators in our study to permit us to detect misidentification, but it was not our initial intent to examine the extent of interspecific admixture, as we did not expect to find the locally extensive admixture suggested by the data.

4.1 Incidence of gene flow from landscaping cultivars in naturally regenerated populations

We detected cultivar parentage in 34 of the 48 naturally regenerated stands, most but not all of which occurred in northwest populations. As our investigation of gene flow from landscaping clones does not include all of the named green ash cultivars released in the last 40 years, our results are likely to be an underestimate of the frequency of cultivar parentage in naturally regenerated stands and the geographic extent of cultivar gene flow into these stands. Our investigation is also conservative in that we only identified the first-generation progeny of cultivars. Given the extensive native range of green ash and the ubiquitous deployment of green ash landscaping clones both within and outside of this native range, we felt it was strategically necessary to first identify sites where gene flow from cultivars has definitely occurred. The high frequency of cultivar parentage in the northwest, where green ash populations are small and widely scattered, coupled with an absence of a signature of higher differentiation along the northwestern edge of the range, suggests that some of these populations may consist entirely of propagules dispersed from cultivars and the improved germplasm planted for shelterbelts and riparian buffers in rural communities.

4.2 Impact of cultivar gene flow on adaptive variation

Studies of outcrossing, wind pollinated forest trees have shown that adaptive variation occurs at local and regional scales even when connectivity among populations is high (Pluess et al. 2016; Firmat et al. 2017). However, before speculation on the potentially harmful impact of gene flow from nonlocal cultivars on adaptive variation, consider the meaning of local adaption in this species. Green ash provenance tests conducted over the last ninety years provide phenotypic evidence for local adaptation while at the same time indicating that provenance alone does not necessarily predict growth rate, one measure of adaptation. Height in 60 provenances planted at 10 test sites (common gardens) and measured at age six was not consistent with the expectation that provenances closest to a given test site will grow the fastest. (Steiner et al. 1988). Although southern provenances did suffer injury and mortality from cold temperatures in northern test sites, the tallest and the earliest maturing trees at most of the test sites were from southern Ontario and a ‘central prairie’ region which included provenances from eastern Nebraska, Iowa, and central Illinois. In the Steiner study and in a previous investigation of 13 year old provenances planted in Massachusetts, the northeastern provenances had no growth advantage in test sites closest to northeastern provenances (Santamour 1963). On the other hand, a provenance test of seedlings from 39 Great Plains provenances from North Dakota, South Dakota, Minnesota, Iowa Nebraska and Kansas did reveal evidence of local adaption for drought resistance, with provenances in northwest North Dakota being most drought resistant (Meuli & Shirley 1937). In the light of all of the evidence presented here and with the additional complication of interspecific hybridization (discussed below), we hypothesize that the current standing genetic variation in F. pennsylvanica is the result of adaptation and potentially nonadapative, human-mediated process (i.e. gene flow from landscaping clones), the two of which cannot be directly disentangled without fine scale functional genomics and intensive phenotyping.

4.3 The significance of interspecific admixture

The most recent molecular phylogeny of the genus Fraxinus resulted in incomplete resolution of the North American species F. pennsylvanica, F. americana, and F. velutina (Hinsinger et al. 2013). The authors conclude that this may have resulted from a very rapid radiation or relatively recent gene exchange. This uncertainly is emphasized in the continuing disagreement over what constitutes F. americana and F. pennsylvanica sensu stricto, even though all authors agree that these two species exist. The native ranges of green and white ash are sympatric in the eastern United States (Little & Elbert 1979), but the extent of sympatry depends on the species concepts for F. americana and F. pennsylvanica. We hypothesize that interspecific hybridization and subsequent admixture occurs in ecological contexts where hybrid individuals live long enough to produce fertile progeny, permitting gene flow from one species to another over many generations. This is consistent with the range of admixture observed. Our hypothesis of locally extensive, naturally occurring admixture can be tested with an assessment of genomic architecture in admixed and unadmixed populations. Alternatively, such an assessment could reveal evidence for recent introgression from multiple nonlocal sources as a result of extensive planting of white ash cultivars in the regions where we observed a high incidence of admixture. While it is beyond the scope of this paper to comment on the response of admixed individuals to EAB, this possibility merits further investigation as some studies suggest that white ash may be slightly less susceptible or less attractive to EAB than green ash (Robinett & McCullough 2019).

4.4 Implications for conservation of the North American Fraxinus

Conservation of the gene pool of F. pennsylvanica requires seed collections within populations representing the Northwestern, Northern and Southern groups we detected in this study, but what of the admixed individuals and extensively admixed populations? Admixture among these groups and with F. americana at some sites may be the result of a naturally occurring process predating cultivar development or gene flow from cultivars, or both. The ability of outcrossing, closely related, native, and sympatric species of forest trees to produce fertile, vigorous, and long-lived interspecific hybrids can result in a syngameon, where gene flow among species occurs yet species distinctness persists (Hipp et al. 2019). Conservation plans based on exclusion of admixed individuals could interfere with a long-standing, potentially adaptive process. A second issue is cultivar gene flow. If more detailed genotyping reveals range edge populations consisting of the descendants of many cultivars, does the ecological importance of these stands outweigh their origin? Intensive study of the landscape genomics of the North American Fraxinus species, especially those in which species boundaries are uncertain and admixture is well documented, may shed light on adaptive mechanisms, the actual extent of admixture and the drivers that influence the frequency of intraspecific and interspecific admixture.

Interdisciplinary efforts based on forest monitoring, seed collection, long-term breeding programs, landscape genomics and intensive phenotyping will all be required to preserve the gene pool of the American Fraxinus and restore Fraxinus populations to the landscape. As the EAB infestation continues, climate change accelerates, and the direct impact of the built environment increases, the adaptive landscape of the American Fraxinus is changing in real time, requiring a re-examination of the assumptions that local provenance necessarily means local adaptation and perhaps even more important, that standing genetic variation is of long standing.

Acknowledgements

We appreciate the assistance of federal, state, provincial, county, and municipal governments in the United States and Canada in locating naturally regenerated stands of green ash and granting us permission to collect samples. We thank the Minnesota Landscape Arboretum for allowing us access to their Fraxinus cultivar collection, and the University of Notre Dame Genomics Core Facility for technical advice and assistance with genotyping.

Funding

This work was funded by the National Science Foundation grant I0S1025974 ‘Comparative Genomics of Environmental Stress Responses in North American Hardwoods’. JR-S also acknowledges support from USDA-USFS APHIS grant 18-IA-11242316-105, USDA-APHIS grant 20-JV-11242303-050 and The Tree Fund Foundation Tree Fund grant 18-JD-01. RKS acknowledges support from NIH training grant T32GM075762. JLK acknowledges support from USDA APHIS 18-IA-11242316-105, Michigan Invasive Species Grant Program grant IS18-119, the Commonwealth of Pennsylvania Department of Conservation and Natural Resources Bureau of Forestry 18-CO-11242316-014, and the U.S. Forest Service Special Technology Development Program grant NA-2017-01.

Competing Interests

All of the authors declare that they have no competing financial or nonfinancial interests.

Author Contributions

JR-S, JLK and AGN conceived and designed the work; AGN did the range-wide collections. AGN and TS prepared all of the range-wide samples for genotyping; AGN performed the initial statistical analysis. DLS collected the cultivars, prepared the samples for genotyping and performed the parentage analysis. DLS generated the maps for the figures. RKS performed additional genotyping and assisted with the final figures and tables. AGN and JR-S wrote the first version of paper and DLS wrote the second version including the parentage analysis. JR-S and JKL generated the final version and RKS performed the final editing.

Data Availability

Data will be available from the Dryad Digital Repository after acceptance.

Beatty GE, Brown JA, Cassidy EM, Finlay CMV, McKendrick L, Montgomery WI, Reid N, Tosh DG, Provan J (2015) Lack of genetic structure and evidence for long-distance dispersal in ash (Fraxinus excelsior) populations under threat from an emergent fungal pathogen: implications for restorative planting, vol 11. Tree Genetics & Genomes, p 53
Black CL, Beckmann RL (1983) The Variability of Nuclear DNA and its Implications for Polyploidy in White Ash (Fraxinus americana L.: Oleaceae). Am J Bot 70:1420–1423
Campbell JJN (2017) Green/red and white ashes (Fraxinus sect. Melioides) of east-central North America: taxonomic concepts and polyploidy. Phytoneuron 28:1–36
Cornuet JM, Luikart G (1996) Description and Power Analysis of Two Tests for Detecting Recent Population Bottlenecks From Allele Frequency Data. Genetics 144:2001–2014
Cunningham RA (1988) 27. Genetic improvement of trees and shrubs used in windbreaks, vol 22–23. Agriculture, Ecosystems & Environment, pp 483–498
Dawson DH, Read RA (1964) Guide for selecting superior trees for shelterbelts in the Prairie Plains. Lake States Forest Experiment Station, Forest Service, US Department of …
Dugal AW (1971) A study of Fraxinus americana L. and Fraxinus pennsylvanica Marsh. in the Ottawa district. Carleton University
Earl DA, vonHoldt BM (2012) Structure harvester: a website and program for visualizing structure output and implementing the Evanno method. Conserv Genet Resour 4:359–361
Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clusters of individuals using the software STRUCTURE: A simulation study. Mol Ecol 14:2611–2620
Firmat C, Delzon S, Louvet JM, Parmentier J, Kremer A (2017) Evolutionary dynamics of the leaf phenological cycle in an oak metapopulation along an elevation gradient. J Evol Biol 30:2116–2131
Foll M, Gaggiotti OE (2006) Identifying the environmental factors that determine the genetic structure of populations. Genetics 174:875–891
Garrison GA, Bjugstad AJ, Duncan DA, Lewis ME, Smith DR (1977) Vegetation and environmental features of forest and range ecosystems
Garza JC, Williamson EG (2001) Detection of reduction in population size using data from microsatellite loci. Mol Ecol 10:305–318
Girardin MP, Isabel N, Guo XJ, Lamothe M, Duchesne I, Lenz P (2021) Annual aboveground carbon uptake enhancements from assisted gene flow in boreal black spruce forests are not long-lasting. Nat Commun 12:1169
Hardin JW (1974) Studies of the southeastern United States flora. Iv. Oleaceae. SIDA, Contributions to Botany, 5, 274–285
Heuertz M, Hausman JF, Hardy OJ, Vendramin GG, Frascaria-Lacoste N, Vekemans X (2004) Nuclear microsatellites reveal contrasting patterns of genetic structure between western and southeastern European populations of the common ash (Fraxinus excelsior L.). Evolution 58:976–988
Hinsinger DD, Basak J, Gaudeul M, Cruaud C, Bertolino P, Frascaria-Lacoste N, Bousquet J (2013) The Phylogeny and Biogeographic History of Ashes (Fraxinus, Oleaceae) Highlight the Roles of Migration and Vicariance in the Diversification of Temperate Trees.PLoS ONE, 8, e80431
Hipp AL, Whittemore AT, Garner M, Hahn M, Fitzek E, Guichoux E, Cavender-Bares J, Gugger PF, Manos PS, Pearse IS (2019) Genomic Identity of White Oak Species in an Eastern North American Syngameon1. Ann Mo Bot Gard 104:455–477
Holm S (1979) A simple sequentially rejective multiple test procedure. Scand J Stat 6:65–70
Izumi K, Bartlein PJ (2016) North American paleoclimate reconstructions for the Last Glacial Maximum using an inverse modeling through iterative forward modeling approach applied to pollen data.Geophysical Research Letters, 43, 10,965 – 910,972.
Jackson ST, Webb RS, Anderson KH, Overpeck JT, Webb Iii T, Williams JW, Hansen BCS (2000) Vegetation and environment in Eastern North America during the Last Glacial Maximum. Q Sci Rev 19:489–508
Kalinowski ST, Taper ML, Marshall TC (2007) Revising how the computer program CERVUS accommodates genotyping error increases success in paternity assignment. Mol Ecol 16:1099–1106
Lane T, Best T, Zembower N, Davitt J, Henry N, Xu Y, Koch J, Liang H, McGraw J, Schuster S, Shim D, Coggeshall MV, Carlson JE, Staton ME (2016) The green ash transcriptome and identification of genes responding to abiotic and biotic stresses. BMC Genomics 17:702
Little J, Elbert L (1979) Atlas of United States trees. J Forest 77:188
Liu H (2018) Under Siege: Ash Management in the Wake of the Emerald Ash Borer.Journal of Integrated Pest Management,9
Liu T, Li Q, Kong L, Yu H (2017) Comparison of microsatellites and SNPs for pedigree analysis in the Pacific oyster Crassostrea gigas. Aquacult Int 25:1507–1519
Luikart G, Allendorf FW, Cornuet JM, Sherwin WB (1998) Distortion of allele frequency distributions provides a test for recent population bottlenecks. J Hered 89:238–247
MacFarlane DW, Meyer SP (2005) Characteristics and distribution of potential ash tree hosts for emerald ash borer. For Ecol Manag 213:15–24
Marshall TC, Slate J, Kruuk LEB, Pemberton JM (1998) Statistical confidence for likelihood-based paternity inference in natural populations. Mol Ecol 7:639–655
Meuli LJ, Shirley HL (1937) The Effect of Seed Origin on Drought Resistance of Green Ash in the Prairie-Plains States. J Forest 35:1060–1062
Nesom G (2010) Fraxinus biltmoreana and Fraxinus smallii (Oleaceae), forest trees of the eastern United States. Phytoneuron 51:1–30
Nishio S, Takada N, Terakami S, Takeuchi Y, Kimura MK, Isoda K, Saito T, Iketani H (2021) Genetic structure analysis of cultivated and wild chestnut populations reveals gene flow from cultivars to natural stands. Sci Rep 11:240
Noakes AG, Best T, Staton ME, Koch J, Romero-Severson J (2014) Cross amplification of 15 EST-SSR markers in the genus Fraxinus. Conserv Genet Resour 6:969–970
Peakall R, Smouse PE (2006) GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Mol Ecol Notes 6:288–295
Peakall R, Smouse PE (2012) GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research-an update. Bioinformatics 28:2537–2539
Piry S, Luikart G, Cornuet JM (1999) Computer note. BOTTLENECK: a computer program for detecting recent reductions in the effective size using allele frequency data. J Hered 90:502–503
Pluess AR, Frank A, Heiri C, Lalagüe H, Vendramin GG, Oddou-Muratorio S (2016) Genome–environment association study suggests local adaptation to climate at the regional scale in Fagus sylvatica. New Phytol 210:589–601
Poland TM, McCullough DG (2006) Emerald ash borer: invasion of the urban forest and the threat to North America’s ash resource. J Forest 104:118–124
Prasad A, Iverson L (2003) Little's range and FIA importance value database for 135 eastern US tree species. Northeastern Research Station, USDA Forest Service., Delaware, OH
Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155:945–959
Ramírez-Valiente JA, Robledo-Arnuncio JJ (2015) Exotic gene flow affects fitness trait values but not levels of heritable trait variation in the southernmost population of Scots pine (Pinus sylvestris L. var. nevadensis). Biol Conserv 192:331–342
Raupp MJ, Cumming AB, Raupp EC (2006) Street tree diversity in eastern North America and its potential for tree loss to exotic borers. Arboric Urban Forestry 32:297–304
Robinett MA, McCullough DG (2019) White ash (Fraxinus americana) survival in the core of the emerald ash borer (Agrilus planipennis) invasion. Can J For Res 49:510–520
Sampson JF, Byrne M (2008) Outcrossing between an agroforestry plantation and remnant native populations of Eucalyptus loxophleba. Mol Ecol 17:2769–2781
Santamour FS (1963) Thirteen-year growth of some green ash provenances in the Northeast. Research Note NE-14. U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station, Upper Darby, PA, pp 1–5
Santamour F Jr, McArdle A (1983) Checklist of cultivars of North American ash (Fraxinus) species. J Arboric 9:271–276
Schmiedel D (2010) Fraxinus pennsylvanica in the forests of the floodplain of the central Elbe River, Germany. Weißensee-Verlag, Berlin, Germany
Schmiedel D, Tackenberg O (2013) Hydrochory and water induced germination enhance invasion of Fraxinus pennsylvanica. For Ecol Manag 304:437–443
Schultz RC, Isenhart TM, Simpkins WW, Colletti JP (2004) Riparian forest buffers in agroecosystems – lessons learned from the Bear Creek Watershed, central Iowa. USA Agrofor Syst 61:35–50
Smouse PE, Long JC, Sokal RR (1986) Multiple regression and correlation extensions of the mantel test of matrix correspondence. Syst Zool 35:627–632
Steiner K, Williams M, Dettayes D, Hall R, Eckert R, Bagley W, Lemmien W, Karnosky D, Carter K, Cech F (1988) Juvenile performance in a range-wide provenance test of Fraxinus pennsylvanica Marsh. Silvae Genetica. (Germany, FR)
Steinitz O, Robledo-Arnuncio JJ, Nathan R (2012) Effects of forest plantations on the genetic composition of conspecific native Aleppo pine populations. Mol Ecol 21:300–313
Sutherland BG, Belaj A, Nier S, Cottrell JE, Vaughan P, Hubert S, Russell J K (2010) Molecular biodiversity and population structure in common ash (Fraxinus excelsior L.) in Britain: implications for conservation. Mol Ecol 19:2196–2211
Taylor SMO (1973) Ecological and genetic isolation of Fraxinus americana and Fraxinus pennsylvanica. Univeristy of Michigan
Tollefsrud MM, Myking T, Sønstebø JH, Lygis V, Hietala AM, Heuertz M (2016) Genetic Structure in the Northern Range Margins of Common Ash, Fraxinus excelsior L, vol 11. PLOS ONE, p e0167104
Veloz SD, Williams JW, Blois JL, He F, Otto-Bliesner B, Liu Z (2012) No-analog climates and shifting realized niches during the late quaternary: implications for 21st-century predictions by species distribution models. Glob Change Biol 18:1698–1713
Whittemore AT, Campbell JJN, Xia Z-L, Carlson GH, Atha D, Olsen RT (2018) Ploidy Variation in Fraxinus L. (Oleaceae) of Eastern North America: Genome Size Diversity and Taxonomy in a Suddenly Endangered Genus. Int J Plant Sci 179:377–389
Wright JW (1953) Summary of Tree Breeding Experiments by the Northeastern Forest Experiment Station 1947–1950. United States Department of Agriculture Fortest Service, Upper Darby, PA, p 47
Zampini JW (1992) Green ash tree named Cimmzam. In: Office USP) (ed) Google Patents. James W. Zampini, US, p 9

PDF

Version 1

posted 21 Mar, 2022

You are reading this latest preprint version

Range-wide genetic assessment of F. pennsylvanica reveals a high frequency of parentage from Fraxinus cultivars and extensive admixture between F. pennsylvanica and F. americana

Status:

Version 1

Abstract

Figures

1 Introduction

2 Methods

2.1 Study area and species characteristics

2.2 Sample collection design

2.3 DNA Extraction, PCR, and Genotyping.

2.4 Statistical analyses

3 Results

3.1 Descriptive statistics

3.2 Measures of differentiation and bottlenecks

3.2 Genetic structure and population admixture

3.3 White ash and white ash admixture

3.4 Cultivars and cultivar parentage

4 Discussion

4.1 Incidence of gene flow from landscaping cultivars in naturally regenerated populations

4.2 Impact of cultivar gene flow on adaptive variation

4.3 The significance of interspecific admixture

4.4 Implications for conservation of the North American Fraxinus

Declarations

References

Supplementary Files

Status:

Version 1