Geographic Patterns of RAPD Variation in ... - Semantic Scholar

Published online May 6, 2005

Geographic Patterns of RAPD Variation in Cultivated Flax Yong-Bi Fu*

Reproduced from Crop Science. Published by Crop Science Society of America. All copyrights reserved.

ABSTRACT

et al., 2002c), and phenotypic (Diederichsen and Hammer, 1995) studies suggests that the pale flax (L. angustifolium Huds.) is the most likely progenitor of cultivated flax and the likely place of origin was within the Near East region (Zohary and Hopf, 2000). However, the Near East does not fully match the dominant regions of great flax diversity: Indian Subcontinent, Abyssinian, Near East, and Mediterranean (Vavilov 1926; 1951). The flax types from the Indian Subcontinent and Abyssinian regions were found to be morphologically more diverse than those from the Near East and Mediterranean regions (Vavilov, 1926; 1951). It is possible that Vavilov’s regions of flax diversity reflected only the multiple domestications which occurred at these regions over time, not necessarily the true place of flax origin, as the origin of a crop plant is diffuse in both time and space (Harlan 1956; 1971). If this reasoning is correct, analyses of geographical patterns of flax variation should yield more information on the number of flax domestication centers and how these centers are genetically connected (Harlan, 1975; Sokal et al., 1991) and less on the region of flax origin (Harlan, 1986). However, such geographic analyses have received little attention (Vavilov, 1926; Elladi, 1940; Dillman, 1953; Chandra and Makhija, 1979), particularly through the application of molecular markers. Exploitation of flax genetic resources for flax improvement requires knowledge of the range and structure of the genetic variability present in flax gene pools, but comprehensive characterizations of the existing flax germplasm held in the world collections are lacking, particularly using molecular techniques. Early characterization efforts based on agrobotanical characters revealed four major regions of flax diversity as mentioned above (Vavilov, 1926; 1951) and several unique groups of cultivated flax such as fiber flax, intermediate flax, largeseeded flax, and dehiscent flax (Dillman, 1953; Kulpa and Danert, 1962). Recent RAPD analyses showed (i) that RAPD variation in flax was generally low; (ii) that more variation existed in landraces than cultivars; (iii) that more variation was detected in linseed, than fiber, flax; (iv) that North American linseed cultivars had more variation than those from the other countries; (v) that more variation was found in the Canadian linseed cultivars released before 1980 than those released after 1980; and (vi) that the genetic erosion in the centurylong North American linseed breeding programs was not significant (Fu et al., 2002b; 2003b). While these variation patterns are significant for the improvement of North American flax, they are limited in helping us understand how large the genetic variation in a flax collection is, how the flax gene pool is structured, how a core set

Geographic studies of plant molecular diversity can provide insights into plant domestication and enhance plant germplasm management and utilization, but such studies are lacking in cultivated flax (Linum usitatissimum L.). The objective of this study was to assess the geographic patterns of flax variability in a world collection of cultivated flax by random amplified polymorphic DNA (RAPD) markers. Sixteen RAPD primers were applied to screen 2727 flax accessions representing 63 countries and one group of unknown origin, and 149 RAPD bands were scored for each accession. Analyses of the RAPD data revealed a wide range of occurrence frequencies of polymorphic bands from 0.0004 to 0.9978 with an average of 0.537. The majority (84.2%) of the RAPD variation resided within accessions of each country and only 15.8% of the variation was present among accessions of different countries. Grouping the accessions into 12 major regions explained 8.2% of the RAPD variation. Accessions from the East Asia and European regions were most diverse, but accessions from the regions of Indian Subcontinent and Africa were most distinct. Accessions from the West Asia region were genetically more related to those from the Africa region and less to those from the Indian Subcontinent region. These findings are significant for understanding flax domestication and also are useful in classifying intraspecific diversity of cultivated flax, establishing a core subset of the flax collection, and exploring new sources of genes for flax improvement.

P

lant gene resources of Canada (PGRC; the Canadian national seed genebank) at Saskatoon maintains a flax collection consisting of 2813 active accessions of cultivated flax from 69 countries and 54 accessions of 26 wild species in the genus Linum (Diederichsen et al., 2002). To support the management and utilization of flax germplasm, the flax collection was characterized from 1999 to 2001 by the RAPD technique (Williams et al., 1990). The characterization generated DNA fingerprints for each accession at 149 RAPD loci screened from 16 informative RAPD primers. These DNA fingerprints represent the first comprehensive molecular dataset ever obtained for a genebank collection of such size and, thus, provide an opportunity to address not only the issues associated with the management and utilization of flax germplasm (Fu et al., 2002a) but also the domestication of cultivated flax (Fu et al., 2002c). Flax has been cultivated for oil and fiber for several thousand years, but its domestication remains poorly understood (Zohary and Hopf, 2000). Evidence from archaeological (Helbaek, 1959; Zohary and Hopf, 2000), genetic (Tammes, 1923; Gill and Yermanos, 1967; Fu Plant Gene Resources of Canada, Saskatoon Research Centre, Agriculture and Agri-Food Canada, 107 Science Place, Saskatoon, SK, Canada S7N 0X2. Received 4 June 2004. Plant Genetic Resources. *Corresponding author ([email protected]). Published in Crop Sci. 45:1084–1091 (2005). doi:10.2135/cropsci2004.0345 © Crop Science Society of America 677 S. Segoe Rd., Madison, WI 53711 USA

Abbreviations: AMOVA, analysis of molecular variance; PGRC, Plant Gene Resources of Canada; RAPD, random amplified polymorphic DNA; UBC, University of British Columbia.

1084

1085


FU: RAPD VARIATION IN CULTIVATED FLAX

of flax germplasm should be sampled, and how the flax germplasm could be effectively managed. To address these issues requires a molecular screening of the world collection of cultivated flax and a pattern analysis of germplasm variability (Fu et al., 2002a). The objective of this analysis was to assess the geographic patterns of RAPD variation in the world collection of cultivated flax maintained at PGRC. Molecular markers are less affected by environmental factors than those morphological and botanical characters previously used to describe flax variation and should be more informative in determining geographic patterns of flax diversity (Harlan, 1986). MATERIALS AND METHODS Plant Materials The PGRC collection of flax germplasm is largely a duplication of the flax collection maintained in the United States National Plant Germplasm System. Newly acquired germplasm include 94 accessions developed in Canada and 571 accessions of diverse origin from Russia, Germany, and the Czech Republic. During rejuvenation from 1999 to 2001, 238 additional accessions with defined phenotypic differences were newly selected from mixed samples. Currently, the PGRC flax collection consists of 2813 active accessions of cultivated flax and stands as the seventh largest collection of flax genetic resources in the world (Diederichsen et al., 2002). A total of 2727 accessions of cultivated flax (Table 1) were subjected to RAPD characterization from 1999 to 2001. These accessions represented 63 countries with some of uncertain origin as a control. To assess the region(s) with the greatest diversity, accessions of known country origin were grouped according to the region classification of Zhukovsky (1968) [and later described by Zeven and Zhukovsky (1975)] for 12 crop diversity centers, not the eight “centers of origin” of Vavilov (1926, 1951). Zhukovsky’s classification was an expansion of Vavilov’s centers from eight to 12 regions by taking into account the evidence newly accumulated and thus is more informative (Zeven and Zhukovsky, 1975). On the basis of Zhukovsky’s classification, no accessions represented the second region (Indochina–Indonesia), while there were 909 accessions originating from the ninth region (Europe–Siberia). To reduce the impact of accession sizes in this analysis, the ninth region was separated to two subregions: East Europe (9a) and West Europe (9b).

DNA Extraction and RAPD Analysis About 10 to 15 seeds of each accession were sampled from the PGRC flax collection and grown in a greenhouse at the Saskatoon Research Centre, Agriculture and Agri-Food Canada. Young leaves were collected from 10 3- to 4-wk-old seedlings, bulked for each accession, freeze-dried for 3 to 5 d, and stored at –80⬚C. DNA was extracted from 50 mg of dry tissue in a 2-mL Eppendorf tube by shaking with three glass beads, adding 1 mL of 95⬚C extraction buffer [0.1 M Tris-HCl (pH 8.0), 10 mM EDTA, 1 M KCl], and incubating at 80⬚C for 10 min with occasional agitation. The homogenate was centrifuged to remove cell debris. The supernatant was treated with 5 ␮L of 10 g L⫺1 RNase A and DNA was precipitated with isopropanol. The DNA was suspended in 50 ␮L of water and then quantified by fluorimetry using Hoechst 33258 stain (Sigma Chemical Co., St. Louis, MO). Typical yields were 10 to 15 ␮g of DNA and all of the DNA samples were diluted to 2.5 ng ␮L⫺1.

Each polymerase chain reaction (PCR) contained 10 ng of DNA as template, 4 U of Taq DNA polymerase (BRL, Mississauga, ON), 50 mM KCl, 2.5 mM MgCl2, 200 ␮M of each dNTP, and 0.2 ␮M decamer primer [University of British Columbia (UBC), Vancouver, BC]. All PCR reactions were performed in the PTC-200 DNA Engine thermocycler (MJ Research, Watertown, MA) with 1.5 min at 95⬚C, followed by 35 cycles of 20 s at 95⬚C, 1 min at 36⬚C (ramp 1⬚C/s), 1.5 min at 72⬚C, and a final 7 min cycle at 72⬚C. PCR products were separated in 2% (w/v) agarose gels in 1⫻ TAE by electrophoresis at 100 V for 3 h. Gels were stained in ethidium bromide and photographed on a digital gel-documentation system (Stratagene, La Jolla, CA). RAPD analyses were performed with 16 informative UBC RAPD primers (Table 2). These primers were selected from previous primer screenings and various tests of reproducibility as described in Fu et al. (2002b, 2003a). The analyses generated about 1800 gel images, from which a total of 149 polymorphic bands were scored independently by two individuals as present (1), absent (0), or uncertain (9) for each accession.

Data Analysis The scored bands were analyzed for the level of polymorphism with respect to each primer, the accession country of origin, and the proposed accession region by counting the number of polymorphic bands and generating the summary statistics on the band frequencies. To visualize the variation pattern, the numbers of polymorphic bands were plotted against their frequencies of occurrence in all assayed accessions. To assess the impact of accession size on the polymorphism observed for each country and region, a regression was made by SAS PROC REG (SAS Institute, Inc., 2004) on the number of accessions over the number of polymorphic bands, the mean band frequency, and within-country (or within-region) variation measured from the sum of squares from the below analysis of molecular variance (AMOVA; Excoffier et al., 1992). To assess RAPD variations across the countries and regions, an AMOVA was performed by Arlequin version 2.001 (Schneider et al., 2002). This analysis enables partitioning of the total RAPD variation into within- and among-country (or region) variation components, and provides a measure of inter-country (or inter-region) genetic distances as the proportion of the total RAPD variation residing between flax accessions of any two countries (or regions) (called Phi statistic; Excoffier et al., 1992). Models involving one level of structuring (country) and two levels of structuring (region and country) were applied. Significance of resulting variance components and inter-country (or inter-region) genetic distances was tested with 10 100 random permutations. To assess the genetic relationships of the flax accessions of different countries and regions, the inter-country (and inter-region) distance matrices of the Phi statistic were analyzed by NTSYS-pc 2.01 (Rohlf, 1997) and clustered with the algorithm of unweighted pair–group methods using arithmetic averages.

RESULTS AND DISCUSSION RAPD Variation A total of 149 polymorphic RAPD bands screened from 16 informative RAPD primers were analyzed in this study. The number of polymorphic bands scored for any specific primer ranged from four (for the primer UBC 790) to 16 (for the primer UBC 337) with an average of nine (Table 2). These bands were presumably detected randomly from the 2n (⫽30) flax chromosomes

1086

CROP SCIENCE, VOL. 45, MAY–JUNE 2005

Table 1. Random amplified polymorphic DNA (RAPD) variations for 2727 flax accessions representing 63 countries, 12 proposed regions, and one group of unknown origin. Region and code† East Asia (1) Australia–New Zealand (3)


Indian Subcontinent (4) Central Asia (5)

West Asia (6)

Mediterranean (7)

Africa (8)

East Europe (9a)

West Europe (9b)

Central America (10) South America (11)

North America (12) Unknown(13)

Country

Number of accessions

Number of polymorphic bands

Mean band frequency‡

Within-country variation§

China Japan South Korea Australia New Zealand India Pakistan Kazakhstan Tajikistan Uzbekistan Afghanistan Iran Iraq Israel Syria Turkey Armenia Azerbaijan Georgia Algeria Cyprus Egypt Greece Italy Malta Morocco Tunisia Portugal Spain Eritrea Ethiopia Kenya South Africa Zimbabwe Belarus Bulgaria Czech Republic Hungary Lithuania Poland Romania Russia Ukraine Yugoslavia Austria Belgium Finland France Germany Ireland Netherlands Sweden United Kingdom Costa Rica Guatemala Nicaragua Argentina Brazil Chile Uruguay Venezuela Canada USA Unknown

56 12 1 32 2 286 103 5 1 1 47 62 13 6 9 125 1 2 9 3 3 22 29 9 1 40 7 1 8 1 58 2 1 1 18 11 93 136 13 14 13 309 26 5 1 2 1 157 27 11 63 5 4 1 1 1 168 7 1 8 1 113 539 19

119 92 na 119 28 136 105 41 na na 113 118 74 62 60 121 na 15 80 45 46 112 121 94 na 112 60 na 77 na 112 31 na na 76 80 129 132 70 96 92 133 100 61 na 30 na 132 111 77 122 52 73 na na na 131 82 na 80 na 129 138 107

0.5722 0.559 na 0.5402 0.5 0.554 0.5665 0.6073 na na 0.5654 0.5372 0.608 0.4957 0.5613 0.5468 na 0.5 0.5132 0.5222 0.5507 0.549 0.5203 0.5643 na 0.5447 0.5341 na 0.4923 na 0.5014 0.5 na na 0.5234 0.523 0.5345 0.5108 0.5693 0.5257 0.5484 0.5302 0.5234 0.5311 na 0.5 na 0.5211 0.5466 0.5432 0.536 0.5548 0.4623 na na na 0.5123 0.5236 na 0.5549 na 0.5163 0.5327 0.5326

27.05 26.20 na 27.18 21.00 22.03 17.43 15.20 na na 26.60 25.39 17.71 20.33 21.83 24.26 na 12.00 25.19 23.33 24.00 26.71 26.22 30.22 na 25.18 18.67 na 24.86 na 24.17 27.00 na na 21.67 23.35 24.48 27.37 23.74 25.79 29.42 25.58 25.67 25.70 na 24.00 na 27.22 25.60 22.96 23.88 20.70 32.17 na na na 25.34 26.43 na 24.89 na 25.63 24.96 26.59

† Based on the region classification for crop diversity first proposed by Zhukovsky (1968) and later described by Zeven and Zhukovsky (1975). The names of some proposed regions were also slightly modified for simplicity. No flax accessions were available from the second region (Indochina–Indonesia). ‡ na, not applicable. § Within-country variation calculated from the sum of squares from analysis of molecular variance (AMOVA).

and would have covered the whole genome with an average genetic distance up to 20 cM between any two bands. However, the true genome coverage by these RAPD loci requires further investigation. The occurrence frequencies of the 149 polymorphic bands ranged from 0.0004 to 0.9978 with an average of

0.537. Most of the polymorphic bands appear to be either frequently or infrequently present in all accessions (Fig. 1). Eighteen polymorphic bands (12%) were present in most of the accessions (with frequency of 0.95 or greater) and 42 polymorphic bands (28%) present only in a small proportion of accessions (with fre-


1087

Table 2. Variations of random amplified polymorphic DNA (RAPD) markers observed for 16 RAPD primers synthesized by University of British Columbia (UBC).


Primer UBC73 UBC248 UBC290 UBC301 UBC336 UBC337 UBC365 UBC396 UBC403 UBC465 UBC548 UBC586 UBC601 UBC731 UBC775 UBC790

Band frequency

Number of polymorphic bands

Mean

Mininum

Maximum

10 9 11 9 8 16 11 9 8 8 12 7 9 13 5 4

0.388 0.703 0.451 0.652 0.434 0.427 0.562 0.525 0.341 0.709 0.669 0.644 0.517 0.512 0.562 0.636

0.003 0.043 0.000 0.123 0.005 0.001 0.053 0.044 0.006 0.001 0.144 0.404 0.005 0.003 0.168 0.244

0.948 0.984 1.000 0.994 0.939 0.978 0.996 0.998 0.854 0.998 0.984 0.988 0.926 1.000 0.963 0.902

quency of 0.15 or lower). Such a pattern of variation is expected for a plant germplasm collection with many accessions representing breeding lines or cultivars, as artificial selection within inbreeding species generally forces both dominant and recessive alleles toward fixation (Villand et al., 1998). For polymorphic bands detected by specific primers, the average band frequency varied greatly from 0.341 for the primer UBC403 to 0.709 for the primer UBC465 (Table 2), indicating the heterogeneity of the detected flax genomes. Thus, a large amount of genetic diversity exists in the world collection of cultivated flax.

Variation among Accessions of Different Countries The number of accessions for each country varied greatly (Table 1). There were 16 countries with only one accession and 47 countries with two to 539 accessions. The country with the most accessions was the USA (with 539 accessions), followed by Russia (309), India (286), Argentina (168), France (157), Hungary (136), Turkey (125), Canada (113), and Pakistan (103). This distribution of accession sizes across countries of origin was significantly associated (r ⫽ 0.6; P ⬍ 0.0001) with the variation in the number of polymorphic bands observed for each country. Thus, a country with more accessions displayed more polymorphic bands. However, associations of accession size were not observed with the mean band frequency and the within-country variation calculated from the sum of squares from AMOVA. The mean band frequency varied among the 47 countries and ranged from 0.46 (UK) to 0.61 (Iraq and Kazakhstan). The within-country variation also varied among the 47 countries and ranged from 12.0 (Azerbaijan) to 32.2 (UK) (Table 1). The countries with the most withincountry variations were UK (32.2), Italy (30.2), Romania (29.4), Hungary (27.4), Australia (27.2), and France (27.2). On the basis of AMOVA with one level of structuring, the among-country variation was significant, and it accounted for 15.8% of the total RAPD variation in

Fig. 1. Number of polymorphic markers of random amplified polymorphic DNA (RAPD) in relation to their frequencies of occurrence in all flax accessions.

the flax accessions assayed (Table 3). Thus, the majority (84.2%) of the RAPD variation still resided within the accessions of a country. Assessments on the genetic relationships of flax accessions of 47 countries with more than one accession revealed two major distinguishable groups (Fig. 2). The first group included accessions from India and Pakistan and the other consisted of accessions from Iran, Syria, Turkey, and Iraq. These two groups largely corresponded with their geographic regions, but accessions from the other countries were clustered in various subgroups with less correspondence with geographic origins. For example, accessions from Algeria and Morocco are genetically similar to those from Australia, Greece, and Hungary. Accessions from Canada and USA were genetically similar to those from New Zealand and Kenya. These results appear to indicate the genetic differentiation among the flax accessions of various countries were relatively weak, except for those of the first two distinguishable groups.

Variation among Accessions of Proposed Regions The flax accessions assayed represented only 11 out of 12 regions originally proposed by Zhukovsky (1968) for crop diversity centers (Table 1). Also, the Central America region was underrepresented with only three accessions (Table 4). The region with the most number of accessions was North America (i.e., having 652 accessions), followed by East Europe (638 accession), Indian Subcontinent (389), and West Europe (271). However, this distribution of accessions for the proposed regions was not significantly associated with the variation in the number of polymorphic bands, the mean band frequency, and the within-country variation measured from the sum of squares from AMOVA. The number of polymorphic bands for each region ranged from 35 (Central America) to 141 (North America). There was a narrow

1088


Table 3. Results for analysis of molecular variance (AMOVA) for 2727 flax accessions representing 63 countries, 12 proposed regions, and one group of unknown origin based on 149 random amplified polymorphic DNA (RAPD) markers and two models of structuring.


Model/source One level model Among countries Within countries Total Two level model Among regions Among countries with regions Within countries Total

df

Sum of squares

Variance component

Percentage of variation

63 2663 2726

6 601.70 32 926.86 39 528.56

2.33 12.36 14.69

15.84 84.16

⬍0.0001

12 51 2663 2726

4 474.56 2 127.14 32 926.86 39 528.56

1.22 1.21 12.36 14.79

8.24 8.15 83.61

⬍0.0001 ⬍0.0001

P-value†

† The probability that the among-country (or among-region) variance component was larger than zero, as computed from 10 100 random permutations.

mean band frequency at the region level, ranging from 0.506 (Africa) to 0.579 (East Asia). East Asia displayed the most within-region variation (28.6), followed by East Europe (27.6). Excluding the underrepresented region of Central America, accessions from the Indian Subcontinent region had the lowest within-region variation (21.7). This indicates accessions within the Indian Subcontinent region were genetically more similar than those within the other regions. On the basis of AMOVA with two levels of structuring, the among-region variation was significant, and it accounted for 8.2% of the total RAPD variation (Table 3). Thus, RAPD variation exists among accessions of different regions but is relatively small. Assessments on the RAPD variations of among-region accessions revealed accessions from the Indian Subcontinent region were the most genetically distinct from those of the other 10 proposed regions, followed by the accessions from Africa, West Asia, and Central Asia (Table 5). Accessions from the Indian Subcontinent region displayed the largest genetic distances with accessions of the other regions (ranging from 0.19–0.32), followed by those of Africa (0.14–0.27), West Asia (0.06–0.22), and Central Asia (0.05–0.19). The genetic distances obtained among the other seven regions largely were smaller than 0.10. Assessing the genetic relationships of the flax accessions originating from different regions revealed four clear, distinct clusters representing the Indian Subcontinent, Africa, West Asia–Central Asia, and the other eight proposed regions including Mediterranean. Accessions from the Indian Subcontinent region were genetically less related to those from the other regions, followed by those from Africa, West Asia– Central Asia, and the remaining regions. Further assessment on these regional patterns with respect to the original four centers of flax diversity proposed by Vavilov (1926; 1951) confirmed N.I. Vavilov’s observations on flax diversity as reflected by morphological and botanic trait evaluation.

Variation Patterns and Centers of Flax Diversity Fig. 2. Genetic relationships of 2711 flax accessions representing 47 countries and one group of unknown origin in a dendrogram based on their differences reflected in random amplified polymorphic DNA (RAPD) markers. The region code is given in the parentheses following the country label.

The RAPD analysis presented here represented the first attempt to assess the regional patterns of molecular variation in a crop species from such a large scale of plant germplasm. The variation patterns revealed here (Fig. 2 and 3) probably reflect all the possible

1089


Table 4. Comparison of random amplified polymorphic DNA (RAPD) variations for flax accessions of different regions.


Region and code East Asia (1) Australia–New Zealand (3) Indian Subcontinent (4) Central Asia (5) West Asia (6) Mediterranean (7) Africa (8) East Europe (9a) West Europe (9b) Central America (10) South America (11) North America (12) Unknown (13) Mean§

Number of accessions

Number of polym. bands†

Mean band frequency

Within-region variation‡

69 34 389 116 165 123 63 638 271 3 185 652 19 209.8 (227)

121 122 136 127 127 138 114 140 134 35 132 141 107 121.1 (128.3)

0.579 0.527 0.550 0.558 0.565 0.539 0.506 0.525 0.532 0.519 0.518 0.530 0.533 0.537 (0.539)

28.60 27.17 21.67 27.04 26.41 26.60 25.48 27.56 27.15 18.33 25.40 25.37 26.59 25.64 (26.25)

† Number of polymorphic bands. ‡ Within-region variation calculated from the sum of squares from analysis of molecular variance (AMOVA). § Value in parentheses was obtained excluding flax accessions from Central America.

“centers of diversity” (or distributions of flax variability) presently existing in the world gene pool of cultivated flax. These should include those, if any, newly developed by extensive exchanges of flax germplasm that have taken place worldwide over the last 60 yr (Peeters, 1988). However, a comparison of these regional patterns of flax variation with those traditional “centers of diversity” formulated in 1920s by N.I. Vavilov (1926, 1951) revealed little discrepancy. This may reflect the fact that the exchanged germplasm was rarely utilized in flax breeding programs and still maintained most of the initial variability captured from originating countries. It is difficult to interpret these regional patterns with respect to the place of flax origin (Zeven and Zhukovsky, 1975; Durrant, 1976), but these findings appear to offer some insight to flax domestication, if one reasons with the diffuse theory of Harlan (1986). Possibly, these regional patterns of flax variability reflect the multiple centers of flax cultivation (Harlan, 1971), since there are significant differences in diversity across regions. Also, some centers of domestication are genetically more related than the others (Fig. 3) and might reflect the geographical direction of flax cultivation. For example, African flax shows a greater relatedness to those from the West Asia and Central Asia regions but a lower

relatedness to those from the Indian Subcontinent region. This implies that African flax may have been spread in domestication from the Near East region, and not the Indian Subcontinent region. If this reasoning is correct, connecting these centers of domestication in chronological order will enhance our understanding of flax domestication. Testing this reasoning could proceed with DNA sequence analyses of flax from various regions (Jaenicke-Despres et al., 2003), which allows for molecular dating of domestication events. However, archaeobotanical support from these regions is definitely needed, particularly for the Indian Subcontinent and Africa regions where flax fossils were largely lacking (Zohary and Hopf, 2000).

Implications for Germplasm Management and Utilization These regional patterns of RAPD variation are significant for the management and utilization of flax germplasm, particularly for those in the PGRC collection. The PGRC collection, although well representative for some countries and regions, could be further enhanced if specific field collections of flax germplasm are performed in the regions of large variation such as East

Table 5. Genetic distances among flax accessions of different regions measured by the proportion of the total random amplified polymorphic DNA (RAPD) variation residing between regions (⌽st; below the diagonal). Significance of each ⌽st value is given above the diagonal as the probability obtained by 10 100 random permutations that a random ⌽st value was greater than observed value. Region and code East Asia (1) Australia–New Zealand (3) Indian Subcontinent (4) Central Asia (5) West Asia (6) Mediterranean (7) Africa (8) East Europe (9a) West Europe (9b) Central America (10) South America (11) North America (12) Unknown (13) * at P ⬍ 0.05. ** at P ⬍ 0.01. *** at P ⬍ 0.001. ns, not significant.

1 0.07 0.25 0.10 0.16 0.05 0.17 0.02 0.04 0.05 0.10 0.07 0.02

3

4

5

6

7

8

9a

9b

10

11

12

13

***

*** ***

*** *** ***

*** *** *** ***

*** *** *** *** ***

*** *** *** *** *** ***

*** *** *** *** *** *** ***

*** *** *** *** *** *** *** ***

ns * *** ** ** ns * ns ns

*** *** *** *** *** *** *** *** *** **

*** *** *** *** *** *** *** *** *** ns ***

ns *** *** *** *** ** *** ns ns ns *** ***

0.23 0.07 0.11 0.03 0.17 0.06 0.04 0.08 0.03 0.05 0.06

0.19 0.20 0.22 0.27 0.24 0.24 0.32 0.24 0.25 0.27

0.06 0.05 0.14 0.10 0.09 0.13 0.10 0.13 0.09

0.11 0.19 0.16 0.16 0.22 0.14 0.18 0.15

0.14 0.04 0.03 0.07 0.05 0.06 0.03

0.16 0.15 0.19 0.19 0.18 0.17

0.02 0.02 0.08 0.04 0.01

0.01 0.05 0.03 0.01

0.11 0.06 0.08

0.04 0.02

0.04


1090


Fig. 3. Genetic relationships of 2727 flax accessions representing 12 proposed regions and one group of unknown origin in a dendrogram based on their differences reflected in random amplified polymorphic DNA (RAPD) markers. Australia–NewZ stands for the Australia–New Zealand region; Indian Subcont. for the Indian Subcontinent region.

Asia and Europe and of great distinctness such as Indian Subcontinent and Africa. Incorporating these regional patterns of variation in evaluations of intraspecific classification of flax diversity such as those of Dillman (1953) and Kulpa and Danert (1962) would make the grouping of intraspecific diversity more informative for germplasm management (Diederichsen and Fu, 2004). Exploring these variation patterns with respect to environmental factors or connecting these patterns to the resistance patterns of flax diseases may result in grouping of accessions with similar adaptedness or disease resistance and thus would facilitate the search for unique genotypes from the collection. Establishing a core subset of flax accessions from the PGRC collection can enhance the management and utilization of flax germplasm (Brown and Spillane, 1999), but challenges remain in the applications of sampling to capture the existing variability. However, the geographic patterns of flax variability revealed here provide some baseline information for assigning the weighting each country or region could contribute to a core subset and for developing effective sampling strategies. Accessions from the Indian Subcontinent, Africa, and West Asia regions may be weighted more than those from other regions. Large weighting could also be given to the accessions from the regions and countries of great diversity (Tables 1 and 4). To capture more variation in a core subset, a random sampling should be stratified at the country level, as more variation resided among countries than among proposed regions. Generally, a

random stratified sampling with size allocation proportion to the level of variation existing in a country should be considered (Brown, 1989), but such sampling may not necessarily be optimal. This issue can be addressed, given the entire collection was characterized. Efforts are being made to assess the effectiveness of different sampling methods in capturing genetic variation in a core subset, and eventually a well-representative core subset will be developed for the PGRC collection. Our analysis revealed that a large amount of genetic diversity existed in the flax collection. About 84% of the variation detected resided within the accessions of a country and 16% among accessions of different countries. Thus, selection of breeding materials could be made from the flax germplasm within a country for flax improvement. Also, about 8% of the RAPD variation resided among accessions of different regions (Table 3), and selection of breeding materials from germplasm of different regions is also possible. Exploring new source of genes within the existing gene pool of cultivated flax might be made initially toward accessions originating from the Indian Subcontinent and Africa regions, where the most distinct germplasm exists. To broaden the genetic base of breeding populations, inclusion of some flax germplasm originating from the East Asia and East Europe regions should be considered, at least for initial screening. Developing specific core subsets of flax accessions on the basis of these variation patterns to facilitate the initial screening of genes of particular interest is also feasible. These core subsets could save some effort in the exploration of new breeding germplasm for flax improvement.

CONCLUSIONS The RAPD analysis reported here represented the first attempt to assess the regional patterns of molecular genetic variation in a crop species from such a large scale of plant germplasm. The PGRC flax collection displayed a large amount of genetic diversity in flax germplasm. The majority (84.2%) of the RAPD variation resided within accessions of each country and only 15.8% of the variation was present among accessions of different countries. Grouping the accessions into 12 major regions explained 8.2% of the RAPD variation. Accessions from the East Asia and European regions were most diverse, but accessions from the regions of Indian Subcontinent and Africa were most distinct. Accessions from the West Asia region were genetically more related to those from the Africa region and less to those from the Indian Subcontinent region. These findings provide additional insights to flax domestication and also are useful in classifying intraspecific diversity of cultivated flax, establishing a core subset of the flax collection, and exploring new sources of genes for flax improvement. ACKNOWLEDGMENTS The author thanks Gregory Peterson for his technical assistance on the RAPD characterization of the flax germplasm; Dallas Kessler, Dave Williams, and Peter Kusters for their


assistance in sampling seeds and planting them in the greenhouse; and Rong-Cai Yang and Axel Diederichsen for their helpful comments on the manuscript.


REFERENCES Brown, A.H.D. 1989. Core collection: A practical approach to genetic resources management. Genome 31:818–824. Brown, A.H.D., and C. Spillane. 1999. Implementing core collections– principles, procedures, progress, problems and promise. p. 1–9. In R.C. Johnson and T. Hodgkin (ed.) Collections for today and tomorrow. International Plant Genetic Resources Institute, Rome, Italy. Chandra, S., and P.O. Makhija. 1979. Genetic variability and interrelationships of some quantitative characters in geographic groups of exotic linseed (Linum usitassimum L.) germplasm. Acta Agron. Acad. Sci. Hung. 28:65–70. Diederichsen, A., and Y.B. Fu. 2005. Phenotypic and molecular (RAPD) differentiation of four infraspecific groups of cultivated flax (Linum usitatissimum L. subsp. usitatissimum). Genet. Resour. Crop Evol. (in press). Diederichsen, A., and K. Hammer. 1995. Variation of cultivated flax (Linum usitatissimum L. subsp. usitatissimum) and its wild progenitor pale flax (subsp. angustifolium (Huds.) Thell.). Genet. Resour. Crop Evol. 42:263–272. Diederichsen, A., J.P. Raney, Y.B. Fu, and K.W. Richards. 2002. Diversity in the flax collection at Plant Gene Resources of Canada. Flax Inst. U.S. Proc. 59:138–143. Dillman, A.C. 1953. Classification of flax varieties. USDA, Tech. Bull. 1064. U.S. Gov. Print. Office, Washington, DC. Durrant, A. 1976. Flax and linseed. p. 190–193. In N.W. Simmonds (ed.) Evolution of crop plants. Longman, London. Elladi, V.N. 1940. Linum usitatissimum (L.) Vav. Consp. Nov.–Len (Russ.). p. 109–207. In E.V. Vul’f and N.I. Vavilov (ed.) Kul’turnaja flora SSSR, prjadil’nye, Vol. 5, Part 1, Sel’chozgiz. Moskva, Leningrad. (In Russian). Excoffier, L., P.E. Smouse, and J.M. Quattro. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: Application to human mitochondrial DNA restriction data. Genetics 131:479–491. Fu, Y.B., A. Diederichsen, and K.W. Richards. 2002a. Molecular characterization of 2800 flax accessions at Plant Gene Resources of Canada with RAPD markers. Flax Inst. U.S. Proc. 59:144–149. Fu, Y.B., A. Diederichsen, K.W. Richards, and G. Peterson. 2002b. Genetic diversity of flax (Linum usitatissimum L.) cultivars and landraces as revealed by RAPDs. Genet. Resour. Crop Evol. 49: 167–174. Fu, Y.B., G.W. Peterson, A. Diederichsen, and K.W. Richards. 2002c. RAPD analysis of genetic relationships of seven flax species in the genus Linum L. Genet. Resour. Crop Evol. 49:253–259. Fu, Y.B., S. Guerin, G.W. Peterson, J.E. Carlson, and K.W. Richards. 2003a. Assessment of bulking strategies for RAPD analyses of flax germplasm. Genet. Resour. Crop Evol. 50:743–746. Fu, Y.B., G.G. Rowland, S.D. Duguid, and K.W. Richards. 2003b. RAPD analysis of 54 North American flax cultivars. Crop Sci. 43: 1510–1515.

1091

Gill, K.S., and D.M. Yermanos. 1967. Cytogenetic studies on the genus Linum. Crop Sci. 7:623–631. Harlan, J.R. 1956. Distribution and utilization of natural variability in cultivated plants. Brookhaven Symp. Biol. 9:191–206. Harlan, J.R. 1971. Agricultural origins: Centers and noncenters. Science (Washington, DC) 174:468–474. Harlan, J.R. 1975. Geographic patterns of variation in some cultivated plants. J. Hered. 66:184–191. Harlan, J.R. 1986. Plant domestication: Diffuse origins and diffusions. p. 21–34. In C. Barigozzi (ed.) The origin and domestication of cultivated plants, Elsevier, New York. Helbaek, H. 1959. Domestication of food plants in the Old World. Science (Washington, DC) 130:365–372. Jaenicke-Despres, V., E.S. Buckler, B.D. Smith, M.T.P. Gilbert, A. Cooper, J. Doebley, and S. Paabo. 2003. Early allelic selection in maize as revealed by ancient DNA. Science (Washington, DC) 302:1206–1208. Kulpa, W., and S. Danert. 1962. Zur Systematik von Linum usitatissimum L. Kulturpflanze. Beiheft 3:341–388. Peeters, J.P. 1988. The emergence of new centres of diversity: Evidence from barley. Theor. Appl. Genet. 76:17–24. Rohlf, F.J. 1997. NTSYS-pc 2.1. Numerical taxonomy and multivariate analysis system. Exeter Software, Setauket, NY. SAS Institute, Inc. 2004. The SAS system for windows V8.02. SAS Institute Incorporated, Cary, NC. Schneider, S., D. Roessli, and L. Excoffier. 2002. Arlequin ver 2.001: A software for population genetics data analysis. Genetics and Biometry Laboratory, Univ. Geneva, Switzland (Available at http://lgb. unige.ch/arlequin/software; verified 6 January 2005). Sokal, R.R., N.L. Oden, and C. Wilson. 1991. Genetic evidence for the spread of agriculture in Europe by demic diffusion. Nature (London) 351:143–145. Tammes, T. 1923. Das genotypische Verhaeltnis zwischen wildem Linum angustifolium und dem Kulturlein, Linum usitatissimum. Genetica (The Hague) 5:61–76. Vavilov, N.I. 1926. Studies on the origin of cultivated plants. Inst. Bot. Appl. et d’ Amelior. des Plantes. State Press, Leningrad. (In Russian and English). Vavilov, N.I. 1951. The origin, variation, immunity and breeding of cultivated plants. Chron. Bot. 13:1–366. Villand, J., P.W. Shroch, T. Lai, P. Hanson, C.G. Kuo, and J. Nienhuis. 1998. Genetic variation among tomato accessions from primary and secondary centers of diversity. Crop Sci. 38:1339–1347. Williams, J.K.G., A.R. Kubelik, K.J. Livak, J.A. Rafalski, and S.V. Tingey. 1990. DNA polymorphisms amplified by arbitrary primers are useful as genetic markers. Nucleic Acids Res. 18:6531–6535. Zeven, A.C., and P.M. Zhukovsky. 1975. Dictionary of cultivated plants and their centres of diversity. Centre for Agricultural Publishing and Documentation, Wageningen, the Netherlands. Zhukovsky, P.M. 1968. New centres of the origin and new gene centres of cultivated plants including specifically endemic micro-centres of species closely allied to cultivated species. Bot. Zh. (Leningrad) 53: 430–460 (In Russian with English abstract). Zohary, D., and M. Hopf. 2000. Domestication of plants in the old world: The origin and spread of cultivated plants in West Asia, Europe and the Nile Valley. 3rd ed. Oxford Univ. Press.