- Email: [email protected]
Sequence signatures from DNA amplification fingerprints reveal fine population structure of the dogwood pathogen Discula destructiva
ELSEVIER
FEMS
Microbiology
Letters
145 (1996) 377-383
Sequence signatures from DNA amplification fingerprints reveal fine population structure of the dogwood pathogen Discula destructiva Gustav0 Caetano-AnollQ
a,*, Robert N. Trigiano ‘, Mark T. Windham b
a Department of Ornamental Horticulture and Landscape Design. 252 Ellington Plant Sciences Bldg., The University of Tennessee, Knoxville. ’ Department
of Entomology
and Plant Pathology,
Received 9 September
TN 37901-1071,
USA
The University of Tennessee, Knoxville,
1996; revised 14 September
1996; accepted
TN 37901-1071,
14 September
USA
1996
Abstract Isolates of Discula destructiva Redlin and an undescribed species of Discula, the filamentous fungi that cause anthracnose of flowering (Cornusflorida L.) and Pacific (Cornus nuttalli Aud.) dogwoods, were analyzed for genetic variation by nucleic acid scanning with arbitrary mini-hairpin oligonucleotide primers. While the fungal population was highly homogeneous throughout the disease range in eastern and western North America, the generation of arbitrary signatures from amplification profiles (ASAP) distinguished isolates from the northeast, middle and southern Appalachian mountain range, and western United States and Canada. ASAP involves a dual-step arbitrary primer-based amplification procedure that generates a large number of informative allelic characters by amplification of monomorphic DNA fingerprints. ASAP analyses showed a fine fungal population structure consistent with the recent and separate introduction of the pathogen on the west and east coasts of North America. Keywords:
Arbitrary
primer;
ASAP;
C0rnu.r; DNA fmgerprinting;
1. Introduction Dogwoods are prized ornamental understory trees native to North American deciduous forests. Two species, flowering (Cornus florida L.) and Pacific (C. nuttalli Aud.) dogwoods, have recently been the major subject of a devastating disease outbreak that has destroyed native dogwood stands throughout the eastern and western forests, respectively. The disease
* Corresponding author. Tel.: +l (615) 974 8841; Fax: +l (615) 974 2765; E-mail: 037%1097/96/$12.00
[email protected]
Copyright PIISO378-1097(96)00435-l
0 1996 Federation
of European
Discula
was first reported in the late 1970’s in the Seattle area [l] and southeastern New York and southwestern Connecticut [2], spreading in the western coastal areas and highlands associated with the Appalachian range. The sudden appearance, severity, and rapid dissemination of the disease (Fig. 1) has led to the hypothesis that the causal organism is an introduced pathogen [2,3]. The anthracnose-causing pathogen is a filamentous fungus described as Discula destructiva Redlin [4]; however, in 7-8% of cases, an undescribed Discula species was also found associated with the disMicrobiological
Societies. Published
by Elsevier Science B.V
ease [5]. DNA amplification fingerprinting (DAF) showed that D. destructivu was restricted in its genetic diversity, suggesting that the highly homogeneous nature of the pathogen results from either founder effects or genetic bottlenecks maintained by clonal reproduction [3]. In contrast, D. umhrinellu (Berk. and Br.) Morelet, a related european species primarily considered an endophyte of beech (Fugus .~ylvatico L.), oak (Quercus robur L.) and chestnut (Custaneu sutiva Mill.), showed considerable genetic variation [6]. In this study, asexually reproducing fungal populations of the two Discuku species were fingerprinted using a recently developed nucleic acid scanning strategy based on the generation of arbitrary signatures from amplification profiles (ASAP) [7]. This dual-step DNA fingerprinting procedure is based on re-amplification of DAF products with conventional [8] or mini-hairpin [9] oligonucleotide primers. Re-amplification extended scanning of DNA sequence by probing annealing sites within the original DAF amplicons. ASAP can detect increased levels of polymorphic DNA and was here used to analyze the population genetic structure of D. destructive and evaluate the origin and spread of the disease on west and east coasts.
2. Materials and methods Fungal isolates were collected from diseased symptomatic leaf and twig tissue, and mycelial DNA (721 ng mg-’ of mycelium) extracted from stationary fungal cultures grown on PDVB medium [3] using a method modified from that of Yoon et al. [IO]. In some cases, equal amounts of mycelia from different isolates were bulked and DNA subsequently extracted. DAF reactions (20 ~1) containing 3 pM octamer or mini-hairpin decamer primer and 1 ng pl-’ of template DNA were amplified for 35 cycles of 10 s at 96°C and 10 s at 30°C in a oven thermocycler [8,9]. Mini-hairpin primers harboring very short 3’ terminal arbitrary sequences were from the series H-NNN [7], where H is the highly stable hairpinturn structure GCGAAGC and N is any one DNA base. ASAP reactions [7] were assembled as DAF cocktails containing 9 PM primer and about 0.1 ng ~1~’ template (I : 10:’ diluted DAF reactions) and amplified for 35 cycles of 30 s at 96”C, 30 s at 3O”C, and 30 s at 72°C. Amplification products were separated in polyester-backed 10% T:2% C polyacrylamide-7 M urea slab minigels and stained with silver [l I]. Bands (50-700 bp in length) representing unordered genetic characters (loci) were scored as present
B 1’: MA-12
1969
MA-14 MA-17
1989 1989
MA-16 2 : NY-329 3: MD-1 4: VA-162
1989 1988 1990 1989
5 : 6: 7: 8:
TN-20 TN-1 SC-106 GA-1
1990 1989 1990 1989
9: 10: 11: 12: A: 6 : C : Cl: E :
AL-151 NY-326 VA-17b NC-2 TN-6 TN-30 TN-31 TN-32 TN-33
1992 1962 1990 1990 1989 1993 1993 1993 1993
1988
1989
1990
1991
1992
Fig. I. (A) Areas in eastern United States where dogwood anthracnose has been detected as of 1994. The sampling locations of several 0. destructive (circles) and Discula sp. (squares) isolates are also given, together with isolation dates. Names reflect the state where individual isolates were sampled. The inset shows relevant sampling locations within the Great Smoky Mountains National Park (GSMNP): A, Roaring Fork; B, Ashhopper Creek; C, Sugarland Visitor Center; D, Rich Mountain; and E, Mount Sterling. (B) Impact of dogwood anthracnose on native dogwood populations in Tennessee, following the analysis of IO-tree plots (with trunk diameters of I in at 4 ft above the ground) in 39 forested locations in the GSMNP. Severity of disease is given as percentage of plots with disease detected (at least 1 tree with anthracnose lesions) or severely epidemic ( > 25% of diseased foliage within the plot). The overlaid graph shows the percentage of tree mortality (open circles) that could be attributed to dogwood anthracnose.
319
G. Caetano-Anollh et al. I FEMS Microbiology Letters 145 (1996) 377-383
l.O0.7 0.5 0.40.3-
0.20.1 -
Fig. 2. ASAP analysis of D. destmctivo isolates. (A) Three mini-hairpin primers were use to amplify DAF fragments independently generated using the octamer GTAACGCC from two different batches of D. destructivu NY-329 mycelia (lanes 1 and 2). (B) Analysis with mini-hairpin H-CAA of bulks of D. destructiva representing South Appalachian @A), North Appalachian-Atlantic (b&AA), Pacific northwest (NW) subpopulations, and reference Disc& sp. isolates (see Fig. 4). (C) Analysis of isolates obtained from a C. kousa orchard in Martha’s Vineyard, Massachussets. Primer H-GCA was used to amplify DAF fragments generated with primer GAAACGCC. Note the ‘clonal’ nature of isolates examined. Molecular markers are in indicated in kb.
(1) or absent (0), and genetic relationships evaluated by phylogenetic analysis using parsimony (PAUP) (version 3.1; Illinois Nat. Hist. Surv., Champaign, IL) with an exhaustive search for minimal trees, and phenetic and numerical analysis using the unweighted pair group cluster analysis using the arithmetic means (UPGMA) method and principal coor-
dinate analysis (PCO) with the NTSYS-pc program (version 1.7; Exeter Software, Setauket, NY). Jaccard similarity coefficients were used instead of Dice coefficients (cf. [3]), providing different estimates of diversity, but similar genetic relationships. Gene diversity estimates for each ASAP locus were calculated to assess the overall distribution of varia-
Table 1 ASAP analysis of a selected group of eastern Discula isolates ASAP primer
H-CGA H-TCC H-CAT H-TCA H-TGA H-CCT H-GTC H-CTA
Average no. of bands/primer
21.2 f 4.4 21.8 f 2.6 31.0fO.O 14.7 f. 3.1 14.1 f 1.1 23.7 + 1.1 16.1 +_3.9 17.2k3.1
Total bands
40 36 31 33 27 28 32 40
Polymorphic
bands:
Total
Within D. destructiva
Within Discula sp.
14 18 0 20 6 6 20 24
20 5 0 10 8 1 6 10
31 21 0 32 23 I 29 37
Unique fingerprints
100 100 0 83 83 83 100 100
(%)
380
G. Cuetam-Anolks
A
et al. I FEMS Microbiology Letters 145 (1996) 377.-383
Discula species
Discula species I
North AppalachianAtlantic
I
I
NW
P
South Appalachian
(
WI -
B
I.
I.
I.
0.4
0.5
0.6
a
0.7
.I.,.,
I
0.8
0.9
1.0
0.4
.,
.,
0.5
.
0.6
,
0.7
,
,
0.8
*,
,
0.9
,
1.0
DAF Fig. 3. Principal coordinate analysis (PCO) (A) and cluster analysis using the UPGMA algorithm (B) of a subgroup of D. destructivn isolates depicting different geographical locations in the eastern United States. Three isolates of an undescribed Discula species (boxed numbers) were used as reference and were defined as outgroup. DAF profiles were produced with 10 conventional primers used previously [3]. ASAP analysis used 8 mini-hairpin decamers (Table 1) to amplify DAF fragments produced with the heptamer CGAGCTG. UPGMA dendrograms have scales that indicate relative genetic similarity of isolates based on correlation coefficients of cluster groups
bility between and within subpopulations using the statistics of Nei [12] as applied to loci defined by arbitrary amplification [ 131.
3. Results and discussion 3.1. ASAP
optimization
ASAP reactions were first optimized for specificity,
yield and reproducibility. Primer, template and magnesium concentrations were analogous to those identified in the analysis of plant DNA [9]. Mini-hairpin primers required at least 6 uM primer to produce reproducible fingerprints, whereas standard octamers required a minimum of 9 uM concentration. ASAP required 0.001-10 ng pl-’ template. ASAP patterns were reproduced between independent amplifications, DNA isolations, or even different batches of harvested mycelia (e.g., Fig. 2A).
381
G. Caetano-AnoN& et al. I FEMS Microbiology Letters 145 (1996) 377-383
A
r MA-14 -I
;;I;;
NAA-1
MA-18 i TN-33 7 TN-8 SA-1b TN-30 Discula sp. I’ 0.3
1 0.4
”
‘1’
0.5
0.6
1 0.7
m 1. 0.6
1 0.9
-TN-32
i
‘I
1.0
0.70
Fig. 4. Cluster analysis with the UPGMA algorithm of bulked fungal samples representing eastern (SA and NAA) and Pacific northwest (NW) subpopulations of D. destructiva (A) and of isolates from an ornamental planting of Chinese dogwood in Martha’s Vineyard, Massachussets, and from flowering dogwoods in Tennessee’s GSMNP (B). Dendrograms have scales that indicate relative genetic similarity of isolates based on correlation coefficients of cluster groups. ASAP analysis of DAF profiles originahy generated with the octamer GTAACGCC were produced using 9 mini-hairpin decamers (H-CCT, H-TGA, H-CTI’, H-CCA, H-CTC, H-CAG, H-CAA, H-CTA, HGGA) in (A). DAF profiles amplified with primer GAAACGCC were analyzed using 6 mini-hairpin decamers (H-GCA, H-GTT, HGGG, H-CAT, H-CGG, and H-CTA) in (B). Bulked fungal samples contained equal amounts of DNA from individual isolates (in parentheses): SA-la (SC-101, SC-106, VA-161, VA-162) SA-lb (TN-l, TN-9, TNlO, TN-14, TN-15 TN-16, TN-20, TN-22), SA-B (AL-149, AL-151, AL-154, GA-l), NAA-1 (MA-11, MA-14, NY-332, NY-329), NAA-2 (NJ-135, PA-129, MD-l, MD-121), NW-l (M2a, M5a, M7a, M21d, M13c, Ml7a, M12e), NW-2 (Cl, C2, C3, Comu, Me3d, Me6, MelOb), and Discula sp. (NC-2, VA-17b, NY-326). See Fig. 1A and [3] for names and locations of fungal isolates. The observed reduction of fungal diversity following the migration of the disease from coastal northeastern regions to the Appalachian mountains matches the historical dissemination of the disease through native flowering dogwood populations.
3.2. ASAP
analysis of isolates representative
of an
eastern subpopulation
Monomorphic
DAF profiles of a subgroup of D. (Fig. 1A) representing an extant fungal subpopulation [3] were produced using primers CGAGCTG or GTAACGCC. Fingerprints contained 23 and 33 bands (50-700 bp), 67 and 54% of which were polymorphic with bands from Discula sp., respectively. ASAP analysis of heptamer-amplified templates produced 267 bands (20.7 (+ 2.4) bands/primer), 68% of which were polymorphic within all species (167 polymorphic characters; 23.2 (+ 13.0) polymorphisms/primer) and 5 and 13 characteristic of one or two D. destructiva isolates, respectively (Table 1). While DAF of individual D. destructiva isolates produced an average 3% polymorphisms per primer [3], our results show that ASAP can detect polymorphic DNA at levels 14-fold higher. The large number of informative unordered allelic characters obtained by amplification from the monomorphic DAF fragments enabled a study of population structure. These characters are heritable [7] but cannot be assigned to any specific DAF amplicon. Genetic relationships were estimated from DAF and destructiva isolates
ASAP data using an ordination technique (PCO) (Fig. 3A), a distance matrix method (UPGMA) (Fig. 3B), and parsimony (PAUP) (not shown). DAF analysis was unable to distinguish D. destructiva isolates from each other, confhming the highly homogeneous nature of the pathogen at the genetic level [3]. In contrast, ASAP analysis clustered the different isolates in two groups, labelled South Appalachian (SA) and North Appalachian-Atlantic (NAA). Within the SA group, isolates from the boundary of the disease in Alabama and Georgia (SA-2) were even separated from the rest. Genetic diversity estimates for the NAA and SA groups were 0.28 and 0.25, respectively (GUT = 0.48). Similar results were obtained by ASAP analysis of octameramplified fragments (not shown). Our results demonstrate the superior performance of ASAP in the dissection of a highly uniform population at the genetic level, highlighting potential applications in the study of processes that determine genetic variation in fungi. 3.3. ASAP
analysis of bulked isolates from
dejined
subpopulations
ASAP analysis was coupled with bulking
of myce-
382
G Caetano-Am&s
et ul. I FEMS Microbiology
ha or genomic DNA from samples representative of arbitrarily defined geographical subpopulations. The strategy simplifies the study of any large group of individual isolates, especially when portraying genetic diversity of a highly uniform population [3]. A total of 297 ASAP bands were produced with 9 mini-hairpin decamers from 8 bulks. Fig. 2B shows representative ASAP profiles. In these experiments, bulks were representative of the pooled individual isolates, as all bands produced by each isolate were present in the fingerprints of the bulked DNA (not shown). Although bulking underestimates within population diversity and provides more conservative estimates than analysis of individual isolates, a northwestern (NW) fungal subpopulation was differentiated from northeastern (NAA) and southeastern (SA) counterparts (Fig. 4A). Most variability in diwas detected between subpopulations versity (GUT = 0.69, ranging 0.50-0.88 with primer). Despite the limited number of isolates examined, results suggest that D. destructiva may have had separate centers of origin in west and east coasts, as subpopulations differed in genetic constitution and diversity. Under this scenario, introduction of the pathogen in both coasts was probably independent from each other, and was influenced by regional biotic and abiotic factors and by a differential adaptation of the pathogen to Pacific and flowering dogwoods. Moreover, the pathogen appears not to have been disseminated by nursery exchanges between coasts, of the disat least during the initial establishment ease. 3.4. ASAP analysis of eastern isolates from restricted locales introduction, natural selection Upon pathogen and subsequent genetic drift within geographical pockets (perhaps disjunct) could result in the establishment of isolates with variant genetic constitutions. The study of isolates from an orchard in Massachussets (representing subpopulation NAA- I ; cf. Fig. 4A) and several locations in GSMNP (representing SA- 1b) may support this view and perhaps explain the variant and disjunct SA-B subpopulation (Fig. 4A) selected to represent a southern disease boundary in 1994. ASAP identified two distinct Tennessee isolates clustering separately from the rest
Letters 145 11996) 377-3X3
(Fig. 4B). The high homogeneity of fungal isolates obtained from the Massachussets orchard (Fig. 2C) was surprising, but expected when invoking the ‘clonal’ nature of the spread of dogwood anthracnose exclusively via asexual spores. The absence of a sexual stage and of gene flow from native to introduced populations (probably due to species-specific differences between Disc&), compounded with short contact time (< 15-20 years), should help explain the very low levels of genetic diversity now uncovered by ASAP analysis in eastern and western fungal subpopulations.
Acknowledgments This study was funded by the Tennessee Agricultural Experiment Station and grants from USDA (91-3424 l-5929) and the Horticultural Research Institute. L. Phillips is thanked for technical assistance and E. Schilling and M. Cruzan for suggestions on the manuscript.
References Disc& species associated with anthracnose of dogwood in the pacific northwest. Plant Dis. 67, 1290. Hibben, C.R. and Daughtrey, M.L. (1988) Dogwood anthracnose in northeastern United States. Plant Dis. 72, 199-203. Trigiano, R.N., Caetano-Anollis, G., Bassam, B.J. and Windham, M.T. (1995) DNA amplification fingerprinting provides evidence that Discuh destructiva, the cause of dogwood anthracnose in North America, is an introduced pathogen. Mycologia 87, 49&500. Redlin, SC. (1991) Disc& destructiva sp. nov., cause of dogwood anthracnose. Mycologia 83, 633-642. Windham, M.T., Erbaugh, E.K., Montgomery-Dee, M.E. and Trigiano, R.N. (1994) Frequency of Disculu destructiva Redlin and an undescribed Discula species from dogwood tissue. Phytopathology 84, 778. Haemmerli, U.A., Brindle, U.E., Petrini, 0. and McDermott, J.M. (1992) Differentiation of isolates of Discula umbrinei~rr (telomorph: Apiognumonia errcrbunda) from beech, chestnut, and oak using randomly amplified polymorphic DNA markers. Mol. Plant-Microbe Interact. 5, 479483. Caetano-Anollbs, G. and Gresshoff, P.M. (1996) Generation of sequence signatures from DNA amplification fingerprints with mini-hairpin and microsatellite primers. Biotechniques 20, 10~1056. Caetano-AnollCs, G., Bassam, B.J. and Cresshoff, P.M. (1991)
[I] Salogga. D.S. and Ammirati, J.F. (1983)
[2] [3]
[4] [S]
[6]
[7]
[8]
G. Caetano-AnoNPs et al. I FEMS Microbiology Letters 145 (1996) 377-383 DNA amplification
fingerprinting
using very short arbitrary
oligonucleotide primers. Bioltechnology 9, 553-557. [9] Caetano-Anolles, G. and Gresshoff, P.M. (1994) DNA amplification fingerprinting using arbitrary mini-hairpin oligonucleotide primers. Bioltechnology 12, 619623. [lo] Yoon, C.K., Glawe, D.A. and Shaw, P.D. (1991) A method for rapid small-scale preparation of fungal DNA. Mycologia 83, 835-838. [l l] Bassam, B.J., Caetano-Anolles, G. and Gresshoff, P.M. (1991)
Fast and sensitive silver staining
383 of DNA in polyacrylamide
gels. Anal. Biochem. 196, 80-83. [12] Nei, M. (1973) Analysis of gene diversity in subdivided populations. Proc. Natl. Acad. Sci. USA 70, 3321-3323. [13] Dawson, I.K., Chalmers, K.J., Waugh, R. and Powell, W. (1993) Detection and analysis of genetic variation in Hordeum spontaneum populations from Israel using RAPD markers. Mol. Ecology 2, 151-159.
FEMS
Microbiology
Letters
145 (1996) 377-383
Sequence signatures from DNA amplification fingerprints reveal fine population structure of the dogwood pathogen Discula destructiva Gustav0 Caetano-AnollQ
a,*, Robert N. Trigiano ‘, Mark T. Windham b
a Department of Ornamental Horticulture and Landscape Design. 252 Ellington Plant Sciences Bldg., The University of Tennessee, Knoxville. ’ Department
of Entomology
and Plant Pathology,
Received 9 September
TN 37901-1071,
USA
The University of Tennessee, Knoxville,
1996; revised 14 September
1996; accepted
TN 37901-1071,
14 September
USA
1996
Abstract Isolates of Discula destructiva Redlin and an undescribed species of Discula, the filamentous fungi that cause anthracnose of flowering (Cornusflorida L.) and Pacific (Cornus nuttalli Aud.) dogwoods, were analyzed for genetic variation by nucleic acid scanning with arbitrary mini-hairpin oligonucleotide primers. While the fungal population was highly homogeneous throughout the disease range in eastern and western North America, the generation of arbitrary signatures from amplification profiles (ASAP) distinguished isolates from the northeast, middle and southern Appalachian mountain range, and western United States and Canada. ASAP involves a dual-step arbitrary primer-based amplification procedure that generates a large number of informative allelic characters by amplification of monomorphic DNA fingerprints. ASAP analyses showed a fine fungal population structure consistent with the recent and separate introduction of the pathogen on the west and east coasts of North America. Keywords:
Arbitrary
primer;
ASAP;
C0rnu.r; DNA fmgerprinting;
1. Introduction Dogwoods are prized ornamental understory trees native to North American deciduous forests. Two species, flowering (Cornus florida L.) and Pacific (C. nuttalli Aud.) dogwoods, have recently been the major subject of a devastating disease outbreak that has destroyed native dogwood stands throughout the eastern and western forests, respectively. The disease
* Corresponding author. Tel.: +l (615) 974 8841; Fax: +l (615) 974 2765; E-mail: 037%1097/96/$12.00
[email protected]
Copyright PIISO378-1097(96)00435-l
0 1996 Federation
of European
Discula
was first reported in the late 1970’s in the Seattle area [l] and southeastern New York and southwestern Connecticut [2], spreading in the western coastal areas and highlands associated with the Appalachian range. The sudden appearance, severity, and rapid dissemination of the disease (Fig. 1) has led to the hypothesis that the causal organism is an introduced pathogen [2,3]. The anthracnose-causing pathogen is a filamentous fungus described as Discula destructiva Redlin [4]; however, in 7-8% of cases, an undescribed Discula species was also found associated with the disMicrobiological
Societies. Published
by Elsevier Science B.V
ease [5]. DNA amplification fingerprinting (DAF) showed that D. destructivu was restricted in its genetic diversity, suggesting that the highly homogeneous nature of the pathogen results from either founder effects or genetic bottlenecks maintained by clonal reproduction [3]. In contrast, D. umhrinellu (Berk. and Br.) Morelet, a related european species primarily considered an endophyte of beech (Fugus .~ylvatico L.), oak (Quercus robur L.) and chestnut (Custaneu sutiva Mill.), showed considerable genetic variation [6]. In this study, asexually reproducing fungal populations of the two Discuku species were fingerprinted using a recently developed nucleic acid scanning strategy based on the generation of arbitrary signatures from amplification profiles (ASAP) [7]. This dual-step DNA fingerprinting procedure is based on re-amplification of DAF products with conventional [8] or mini-hairpin [9] oligonucleotide primers. Re-amplification extended scanning of DNA sequence by probing annealing sites within the original DAF amplicons. ASAP can detect increased levels of polymorphic DNA and was here used to analyze the population genetic structure of D. destructive and evaluate the origin and spread of the disease on west and east coasts.
2. Materials and methods Fungal isolates were collected from diseased symptomatic leaf and twig tissue, and mycelial DNA (721 ng mg-’ of mycelium) extracted from stationary fungal cultures grown on PDVB medium [3] using a method modified from that of Yoon et al. [IO]. In some cases, equal amounts of mycelia from different isolates were bulked and DNA subsequently extracted. DAF reactions (20 ~1) containing 3 pM octamer or mini-hairpin decamer primer and 1 ng pl-’ of template DNA were amplified for 35 cycles of 10 s at 96°C and 10 s at 30°C in a oven thermocycler [8,9]. Mini-hairpin primers harboring very short 3’ terminal arbitrary sequences were from the series H-NNN [7], where H is the highly stable hairpinturn structure GCGAAGC and N is any one DNA base. ASAP reactions [7] were assembled as DAF cocktails containing 9 PM primer and about 0.1 ng ~1~’ template (I : 10:’ diluted DAF reactions) and amplified for 35 cycles of 30 s at 96”C, 30 s at 3O”C, and 30 s at 72°C. Amplification products were separated in polyester-backed 10% T:2% C polyacrylamide-7 M urea slab minigels and stained with silver [l I]. Bands (50-700 bp in length) representing unordered genetic characters (loci) were scored as present
B 1’: MA-12
1969
MA-14 MA-17
1989 1989
MA-16 2 : NY-329 3: MD-1 4: VA-162
1989 1988 1990 1989
5 : 6: 7: 8:
TN-20 TN-1 SC-106 GA-1
1990 1989 1990 1989
9: 10: 11: 12: A: 6 : C : Cl: E :
AL-151 NY-326 VA-17b NC-2 TN-6 TN-30 TN-31 TN-32 TN-33
1992 1962 1990 1990 1989 1993 1993 1993 1993
1988
1989
1990
1991
1992
Fig. I. (A) Areas in eastern United States where dogwood anthracnose has been detected as of 1994. The sampling locations of several 0. destructive (circles) and Discula sp. (squares) isolates are also given, together with isolation dates. Names reflect the state where individual isolates were sampled. The inset shows relevant sampling locations within the Great Smoky Mountains National Park (GSMNP): A, Roaring Fork; B, Ashhopper Creek; C, Sugarland Visitor Center; D, Rich Mountain; and E, Mount Sterling. (B) Impact of dogwood anthracnose on native dogwood populations in Tennessee, following the analysis of IO-tree plots (with trunk diameters of I in at 4 ft above the ground) in 39 forested locations in the GSMNP. Severity of disease is given as percentage of plots with disease detected (at least 1 tree with anthracnose lesions) or severely epidemic ( > 25% of diseased foliage within the plot). The overlaid graph shows the percentage of tree mortality (open circles) that could be attributed to dogwood anthracnose.
319
G. Caetano-Anollh et al. I FEMS Microbiology Letters 145 (1996) 377-383
l.O0.7 0.5 0.40.3-
0.20.1 -
Fig. 2. ASAP analysis of D. destmctivo isolates. (A) Three mini-hairpin primers were use to amplify DAF fragments independently generated using the octamer GTAACGCC from two different batches of D. destructivu NY-329 mycelia (lanes 1 and 2). (B) Analysis with mini-hairpin H-CAA of bulks of D. destructiva representing South Appalachian @A), North Appalachian-Atlantic (b&AA), Pacific northwest (NW) subpopulations, and reference Disc& sp. isolates (see Fig. 4). (C) Analysis of isolates obtained from a C. kousa orchard in Martha’s Vineyard, Massachussets. Primer H-GCA was used to amplify DAF fragments generated with primer GAAACGCC. Note the ‘clonal’ nature of isolates examined. Molecular markers are in indicated in kb.
(1) or absent (0), and genetic relationships evaluated by phylogenetic analysis using parsimony (PAUP) (version 3.1; Illinois Nat. Hist. Surv., Champaign, IL) with an exhaustive search for minimal trees, and phenetic and numerical analysis using the unweighted pair group cluster analysis using the arithmetic means (UPGMA) method and principal coor-
dinate analysis (PCO) with the NTSYS-pc program (version 1.7; Exeter Software, Setauket, NY). Jaccard similarity coefficients were used instead of Dice coefficients (cf. [3]), providing different estimates of diversity, but similar genetic relationships. Gene diversity estimates for each ASAP locus were calculated to assess the overall distribution of varia-
Table 1 ASAP analysis of a selected group of eastern Discula isolates ASAP primer
H-CGA H-TCC H-CAT H-TCA H-TGA H-CCT H-GTC H-CTA
Average no. of bands/primer
21.2 f 4.4 21.8 f 2.6 31.0fO.O 14.7 f. 3.1 14.1 f 1.1 23.7 + 1.1 16.1 +_3.9 17.2k3.1
Total bands
40 36 31 33 27 28 32 40
Polymorphic
bands:
Total
Within D. destructiva
Within Discula sp.
14 18 0 20 6 6 20 24
20 5 0 10 8 1 6 10
31 21 0 32 23 I 29 37
Unique fingerprints
100 100 0 83 83 83 100 100
(%)
380
G. Cuetam-Anolks
A
et al. I FEMS Microbiology Letters 145 (1996) 377.-383
Discula species
Discula species I
North AppalachianAtlantic
I
I
NW
P
South Appalachian
(
WI -
B
I.
I.
I.
0.4
0.5
0.6
a
0.7
.I.,.,
I
0.8
0.9
1.0
0.4
.,
.,
0.5
.
0.6
,
0.7
,
,
0.8
*,
,
0.9
,
1.0
DAF Fig. 3. Principal coordinate analysis (PCO) (A) and cluster analysis using the UPGMA algorithm (B) of a subgroup of D. destructivn isolates depicting different geographical locations in the eastern United States. Three isolates of an undescribed Discula species (boxed numbers) were used as reference and were defined as outgroup. DAF profiles were produced with 10 conventional primers used previously [3]. ASAP analysis used 8 mini-hairpin decamers (Table 1) to amplify DAF fragments produced with the heptamer CGAGCTG. UPGMA dendrograms have scales that indicate relative genetic similarity of isolates based on correlation coefficients of cluster groups
bility between and within subpopulations using the statistics of Nei [12] as applied to loci defined by arbitrary amplification [ 131.
3. Results and discussion 3.1. ASAP
optimization
ASAP reactions were first optimized for specificity,
yield and reproducibility. Primer, template and magnesium concentrations were analogous to those identified in the analysis of plant DNA [9]. Mini-hairpin primers required at least 6 uM primer to produce reproducible fingerprints, whereas standard octamers required a minimum of 9 uM concentration. ASAP required 0.001-10 ng pl-’ template. ASAP patterns were reproduced between independent amplifications, DNA isolations, or even different batches of harvested mycelia (e.g., Fig. 2A).
381
G. Caetano-AnoN& et al. I FEMS Microbiology Letters 145 (1996) 377-383
A
r MA-14 -I
;;I;;
NAA-1
MA-18 i TN-33 7 TN-8 SA-1b TN-30 Discula sp. I’ 0.3
1 0.4
”
‘1’
0.5
0.6
1 0.7
m 1. 0.6
1 0.9
-TN-32
i
‘I
1.0
0.70
Fig. 4. Cluster analysis with the UPGMA algorithm of bulked fungal samples representing eastern (SA and NAA) and Pacific northwest (NW) subpopulations of D. destructiva (A) and of isolates from an ornamental planting of Chinese dogwood in Martha’s Vineyard, Massachussets, and from flowering dogwoods in Tennessee’s GSMNP (B). Dendrograms have scales that indicate relative genetic similarity of isolates based on correlation coefficients of cluster groups. ASAP analysis of DAF profiles originahy generated with the octamer GTAACGCC were produced using 9 mini-hairpin decamers (H-CCT, H-TGA, H-CTI’, H-CCA, H-CTC, H-CAG, H-CAA, H-CTA, HGGA) in (A). DAF profiles amplified with primer GAAACGCC were analyzed using 6 mini-hairpin decamers (H-GCA, H-GTT, HGGG, H-CAT, H-CGG, and H-CTA) in (B). Bulked fungal samples contained equal amounts of DNA from individual isolates (in parentheses): SA-la (SC-101, SC-106, VA-161, VA-162) SA-lb (TN-l, TN-9, TNlO, TN-14, TN-15 TN-16, TN-20, TN-22), SA-B (AL-149, AL-151, AL-154, GA-l), NAA-1 (MA-11, MA-14, NY-332, NY-329), NAA-2 (NJ-135, PA-129, MD-l, MD-121), NW-l (M2a, M5a, M7a, M21d, M13c, Ml7a, M12e), NW-2 (Cl, C2, C3, Comu, Me3d, Me6, MelOb), and Discula sp. (NC-2, VA-17b, NY-326). See Fig. 1A and [3] for names and locations of fungal isolates. The observed reduction of fungal diversity following the migration of the disease from coastal northeastern regions to the Appalachian mountains matches the historical dissemination of the disease through native flowering dogwood populations.
3.2. ASAP
analysis of isolates representative
of an
eastern subpopulation
Monomorphic
DAF profiles of a subgroup of D. (Fig. 1A) representing an extant fungal subpopulation [3] were produced using primers CGAGCTG or GTAACGCC. Fingerprints contained 23 and 33 bands (50-700 bp), 67 and 54% of which were polymorphic with bands from Discula sp., respectively. ASAP analysis of heptamer-amplified templates produced 267 bands (20.7 (+ 2.4) bands/primer), 68% of which were polymorphic within all species (167 polymorphic characters; 23.2 (+ 13.0) polymorphisms/primer) and 5 and 13 characteristic of one or two D. destructiva isolates, respectively (Table 1). While DAF of individual D. destructiva isolates produced an average 3% polymorphisms per primer [3], our results show that ASAP can detect polymorphic DNA at levels 14-fold higher. The large number of informative unordered allelic characters obtained by amplification from the monomorphic DAF fragments enabled a study of population structure. These characters are heritable [7] but cannot be assigned to any specific DAF amplicon. Genetic relationships were estimated from DAF and destructiva isolates
ASAP data using an ordination technique (PCO) (Fig. 3A), a distance matrix method (UPGMA) (Fig. 3B), and parsimony (PAUP) (not shown). DAF analysis was unable to distinguish D. destructiva isolates from each other, confhming the highly homogeneous nature of the pathogen at the genetic level [3]. In contrast, ASAP analysis clustered the different isolates in two groups, labelled South Appalachian (SA) and North Appalachian-Atlantic (NAA). Within the SA group, isolates from the boundary of the disease in Alabama and Georgia (SA-2) were even separated from the rest. Genetic diversity estimates for the NAA and SA groups were 0.28 and 0.25, respectively (GUT = 0.48). Similar results were obtained by ASAP analysis of octameramplified fragments (not shown). Our results demonstrate the superior performance of ASAP in the dissection of a highly uniform population at the genetic level, highlighting potential applications in the study of processes that determine genetic variation in fungi. 3.3. ASAP
analysis of bulked isolates from
dejined
subpopulations
ASAP analysis was coupled with bulking
of myce-
382
G Caetano-Am&s
et ul. I FEMS Microbiology
ha or genomic DNA from samples representative of arbitrarily defined geographical subpopulations. The strategy simplifies the study of any large group of individual isolates, especially when portraying genetic diversity of a highly uniform population [3]. A total of 297 ASAP bands were produced with 9 mini-hairpin decamers from 8 bulks. Fig. 2B shows representative ASAP profiles. In these experiments, bulks were representative of the pooled individual isolates, as all bands produced by each isolate were present in the fingerprints of the bulked DNA (not shown). Although bulking underestimates within population diversity and provides more conservative estimates than analysis of individual isolates, a northwestern (NW) fungal subpopulation was differentiated from northeastern (NAA) and southeastern (SA) counterparts (Fig. 4A). Most variability in diwas detected between subpopulations versity (GUT = 0.69, ranging 0.50-0.88 with primer). Despite the limited number of isolates examined, results suggest that D. destructiva may have had separate centers of origin in west and east coasts, as subpopulations differed in genetic constitution and diversity. Under this scenario, introduction of the pathogen in both coasts was probably independent from each other, and was influenced by regional biotic and abiotic factors and by a differential adaptation of the pathogen to Pacific and flowering dogwoods. Moreover, the pathogen appears not to have been disseminated by nursery exchanges between coasts, of the disat least during the initial establishment ease. 3.4. ASAP analysis of eastern isolates from restricted locales introduction, natural selection Upon pathogen and subsequent genetic drift within geographical pockets (perhaps disjunct) could result in the establishment of isolates with variant genetic constitutions. The study of isolates from an orchard in Massachussets (representing subpopulation NAA- I ; cf. Fig. 4A) and several locations in GSMNP (representing SA- 1b) may support this view and perhaps explain the variant and disjunct SA-B subpopulation (Fig. 4A) selected to represent a southern disease boundary in 1994. ASAP identified two distinct Tennessee isolates clustering separately from the rest
Letters 145 11996) 377-3X3
(Fig. 4B). The high homogeneity of fungal isolates obtained from the Massachussets orchard (Fig. 2C) was surprising, but expected when invoking the ‘clonal’ nature of the spread of dogwood anthracnose exclusively via asexual spores. The absence of a sexual stage and of gene flow from native to introduced populations (probably due to species-specific differences between Disc&), compounded with short contact time (< 15-20 years), should help explain the very low levels of genetic diversity now uncovered by ASAP analysis in eastern and western fungal subpopulations.
Acknowledgments This study was funded by the Tennessee Agricultural Experiment Station and grants from USDA (91-3424 l-5929) and the Horticultural Research Institute. L. Phillips is thanked for technical assistance and E. Schilling and M. Cruzan for suggestions on the manuscript.
References Disc& species associated with anthracnose of dogwood in the pacific northwest. Plant Dis. 67, 1290. Hibben, C.R. and Daughtrey, M.L. (1988) Dogwood anthracnose in northeastern United States. Plant Dis. 72, 199-203. Trigiano, R.N., Caetano-Anollis, G., Bassam, B.J. and Windham, M.T. (1995) DNA amplification fingerprinting provides evidence that Discuh destructiva, the cause of dogwood anthracnose in North America, is an introduced pathogen. Mycologia 87, 49&500. Redlin, SC. (1991) Disc& destructiva sp. nov., cause of dogwood anthracnose. Mycologia 83, 633-642. Windham, M.T., Erbaugh, E.K., Montgomery-Dee, M.E. and Trigiano, R.N. (1994) Frequency of Disculu destructiva Redlin and an undescribed Discula species from dogwood tissue. Phytopathology 84, 778. Haemmerli, U.A., Brindle, U.E., Petrini, 0. and McDermott, J.M. (1992) Differentiation of isolates of Discula umbrinei~rr (telomorph: Apiognumonia errcrbunda) from beech, chestnut, and oak using randomly amplified polymorphic DNA markers. Mol. Plant-Microbe Interact. 5, 479483. Caetano-Anollbs, G. and Gresshoff, P.M. (1996) Generation of sequence signatures from DNA amplification fingerprints with mini-hairpin and microsatellite primers. Biotechniques 20, 10~1056. Caetano-AnollCs, G., Bassam, B.J. and Cresshoff, P.M. (1991)
[I] Salogga. D.S. and Ammirati, J.F. (1983)
[2] [3]
[4] [S]
[6]
[7]
[8]
G. Caetano-AnoNPs et al. I FEMS Microbiology Letters 145 (1996) 377-383 DNA amplification
fingerprinting
using very short arbitrary
oligonucleotide primers. Bioltechnology 9, 553-557. [9] Caetano-Anolles, G. and Gresshoff, P.M. (1994) DNA amplification fingerprinting using arbitrary mini-hairpin oligonucleotide primers. Bioltechnology 12, 619623. [lo] Yoon, C.K., Glawe, D.A. and Shaw, P.D. (1991) A method for rapid small-scale preparation of fungal DNA. Mycologia 83, 835-838. [l l] Bassam, B.J., Caetano-Anolles, G. and Gresshoff, P.M. (1991)
Fast and sensitive silver staining
383 of DNA in polyacrylamide
gels. Anal. Biochem. 196, 80-83. [12] Nei, M. (1973) Analysis of gene diversity in subdivided populations. Proc. Natl. Acad. Sci. USA 70, 3321-3323. [13] Dawson, I.K., Chalmers, K.J., Waugh, R. and Powell, W. (1993) Detection and analysis of genetic variation in Hordeum spontaneum populations from Israel using RAPD markers. Mol. Ecology 2, 151-159.