Proc. Natl. Acad. Sci. USA Vol. 90, pp. 9159-9163, October 1993 Population Biology
Randomly amplified polymorphic DNA markers are superior to somatic incompatibility tests for discriminating genotypes in natural populations of the ectomycorrhizal fungus Suillus granulatus
KATHRYN M. JACOBSON*, ORSON K. MILLER, JR.t, AND BRUCE J. TURNER Department of Biology, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061 Communicated by Bruce Wallace, June 21, 1993
ABSTRACT Assessing genetic variation within popula- compatible with B and C, then B and C must be compatible tions and genetic exchange between populations requires an as well. understanding of the distribution and abundance of individual Attempts to determine the genetic control of somatic genotypes within the population. Previous workers have used incompatibility in the Basidiomycetes have not been suc- somatic incompatibility testing to distinguish dones or indi- cessful. Rayner and Todd (2) thus concluded that somatic viduals in natural populations of ectomycorrhizal fungi. How- incompatibility is polygenically controlled. ever, somatic incompatibility tests performed with isolates of Somatic incompatibility testing has recently been applied Suilus granulatus from a natural population revealed a lack of to the description of ectomycorrhizal fungal populations. transitiveness, which brought into question the validity of this Fries (11) concluded that more than one clone of Suillus method. Subsequent studies ofgenetic identity ofthese isolates, luteus (L. ex Fr.) S. F. Gray was associated with a single host using randomly amplified polymorphic DNA (RAPD) markers, tree. Dahlberg and Stenlid (12) used somatic incompatibility conclusively showed that somatically compatible isolates are to investigate the "spreading strategy and clonal size" of not necessarily genetically identical. RAPD marker analysis is Suillus bovinus (L. ex Fr.) Kuntze populations in pine forests more reliable and provides higher resolution of genotype with different stand histories. Finally, Sen (13) showed that distribution in natural populations than does somatic incom- somatic incompatibility reactions could be correlated with patibility testing. This is of particular importance in popula- isozyme patterns for delineating individuals ofS. bovinus and tions of organisms such as ectomycorrhizal fungi in which the Suillus variegatus (Swartz ex Fr.) Kuntze. mating systems are incompletely known. We report here somatic incompatibility tests performed with isolates of Suillus granulatus (L. ex Fr.) Kuntze from a To assess genetic variation within populations and genetic natural population, which caused us to question the assump- exchange between populations, the distribution and abun- tion that somatically compatible isolates of this ectomycor- dance ofindividual genotypes must be known. Determination rhizal fungus are from a single genetic individual (i.e., have of individual genotypes in ectomycorrhizal Basidiomycete the same genotype). We subsequently used randomly ampli- populations is not trivial, however, because the dikaryotic fied polymorphic DNA (RAPD) markers to determine vegetative thalli of these fungi grow below ground in asso- whether somatically compatible S. granulatus isolates were ciation with tree roots. An individual thallus may produce genetically identical. RAPD marker analysis (14, 15) has one or more dikaryotic fruiting bodies (mushrooms), which recently been used in the delineation of distinct genotypes in indicate the local occurrence of a species but provide no closely related individuals in a wide range of organisms, information regarding the abundance of distinct genotypes including fungi (16-21). In contrast to somatic incompatibility (individuals) in the population. testing, isolates are determined to be from the same individ- Researchers working with wood-decomposing and tree- ual only if they have identical patterns for the presence or pathogenic Basidiomycetes (1) have used somatic incompat- absence of markers at the loci examined. ibility testing for delineating different genotypes, which are frequently called clones or individuals. Somatic compatibility is assessed via confrontations of isolates from different METHODS fruiting bodies on nutrient media. If isolates produce a zone Study Site and Fieldwork. A study site (300 m2) was of somatic rejection between them (growth inhibition or established in the Poverty Creek Drainage in Montgomery barrage zones), this is interpreted as an incompatibility County, Virginia. S. granulatus sporocarps were collected reaction, indicating that the two isolates are genetically from the site between May and October of 1987 and 1989 and distinct. However, a compatible reaction (confluent growth mapped relative to 12 marked Pinus strobus trees (100-150 between them) has been interpreted as evidence to indicate years old). Each sporocarp was numbered and bagged indi- that the test isolates are genetically identical (2-10). Based on vidually. somatic incompatibility, these authors discuss the distribu- Isolations. Tissue isolates from each sporocarp were made tion of clones, individuals, or genets in and between fungal on Hagem's agar and subsequently subcultured for storage at populations. Rayner and Todd (2) state: "in nature the 8°C in the Virginia Polytechnic Institute and State University antagonism [somatic incompatibility] is a purely vegetative Culture Collection (VTCC). mechanism which operates to delineate individuals within Somatic Incompatibility Testing. Nine isolates, randomly freely interbreeding populations." Implicit in the assumption chosen from throughout the plot, for the somatic incompat- of somatically compatible isolates being genetically identical ibility tests were removed from the cold and initially grown is that the crosses will exhibit transitiveness: if isolate A is Abbreviation: RAPD, randomly amplified polymorphic DNA. The publication costs of this article were defrayed in part by page charge *Present address: Desert Ecological Research Unit ofNamibia, P.O. payment. This article must therefore be hereby marked "advertisement" Box 1592, Swakopmund, Namibia 9000. in accordance with 18 U.S.C. §1734 solely to indicate this fact. tTo whom reprint requests should be addressed. 9159 Downloaded by guest on October 1, 2021 9160 Population Biology: Jacobson et al. Proc. Natl. Acad Sci. USA 90 (1993) on Hagem's agar. When vigorous growth of unpigmented hyphae was observed, two 4-mm2 plugs were removed from the growing edge of each culture and placed 5 mm apart near the center on one side of a 10-cm-diameter plate. Two plugs from another isolate were similarly placed on the opposite side of the plate, 5 mm from the plugs of the first isolate. Thus, on each plate it was possible to score the interaction of each isolate with itself (a control for the compatibility reac- tion) and with the isolate with which it was being tested. An isolate made from a sporocarp from Claytor Lake, Virginia (30 miles due east of the Poverty Creek site; isolate VT1986), was paired with each of the Poverty Creek isolates as a control for the incompatibility reaction. All 10 isolates were paired with each other, and each cross-combination was repeated at least twice from different subcultures of the isolate retained in the VTCC. Preliminary experiments indi- cated that the strongest incompatibility reactions were ob- served 5-7 weeks after inoculation. RAPD Marker Analysis. Two separate liquid bubble cul- tures of each of the nine Poverty Creek site isolates were initiated from the subcultures used in the somatic incompat- ibility tests. These liquid subcultures were maintained sep- arately throughout the procedure and used for replication of the RAPD marker analysis to ensure that differences in RAPD marker phenotypes observed between isolates were FIG. 1. (A and B) Somatic incompatibility of isolates VT1993 not the result of somatic mutations incurred during the (left) and VT1996 (right), as indicated by a barrage zone (arrows) bulking-up procedure. Cultures were harvested by rapid between the isolates. (A) View of top with above lighting. (B) View freezing in liquid nitrogen and lyophilization. The dried of bottom with above lighting. (C and D) Somatic compatibility of mycelium was ground to a fine powder with a coffee grinder isolates VT1994 (left) and VT1995 (right), as indicated by confluent and stored desiccated at 4°C. DNA was isolated from the growth between the isolates. (C) View of top with above lighting. (D) ground mycelium by the method of Lee and Taylor (22). The View of bottom with above lighting. RAPD marker amplification protocol used was a modification of that by Williams et al. (15). Preliminary experiments VT1995 prevented their delineation as discrete individuals showed that DNA template concentrations of 7-9 ng were (Fig. 2). (For example, somatic compatibility between optimal for the reaction with S. granulatus. Five primers VT1990 and VT1993 suggested that they were from the same were screened. To ensure that results were reproducible, individual, as did somatic compatibility between VT1993 and DNA from both liquid subcultures of each isolate was as- VT1994. However, VT1994 and VT1990 were not compatible sayed with three of the five primers. No differences in and are thus from different individuals.) banding patterns were observed for 0.3- to 1.5-kb bands for RAPD Marker Analysis. Five primers yielded a total of 47 subcultures of the isolates. RAPD markers (Table 1). (Fig. 3 illustrates the RAPD markers identified with three of the primers.) A phenetics program, NTSYS (23), was used to quantify similarities be- RESULTS tween the nine isolates. A similarity matrix, obtained via the Somatic Incompatibility. Somatic incompatibility reactions were observed weekly for at least 6 weeks. Initially, mycelia | = P. strobus . from each plug grew out in a concentric fashion, regardless X = S. granulatus of the juxtaposition of an adjacent plug of the same or different isolate. Aerial hyphae from both isolates were observed growing within the same zone with no apparent avoidance behavior. Growth between the four plugs was confluent prior to the development of a barrage zone at 3-6 weeks between nonself plugs (Fig. 1). This zone was ob- served as a line, when illuminated from the bottom, and as a raised ridge within a depressed zone, when illuminated from 0 / the top. Barrage zones were never seen between plugs from 1984 ()2006 the same isolate but were seen between all Poverty Creek isolates and the Claytor Lake isolate. Thus, the barrage zone was interpreted as an indication of somatic incompatibility between isolates. Isolates that did not develop barrage zones . were considered somatically compatible (Fig. 1). All crosses were repeated a second time and all results were reproduc- ible. In addition, no differences in growth rate or pigment A production were observed between cultures collected in 1987 (N-@1993 N and 1989, suggesting that culture age did not affect the results 1996 lli = 1 meter of the crosses. 1963 Somatic incompatibility identified four distinct clones in 1989 the population (Fig. 2): (i) VT1984 and VT2006 collected in FIG. 2. Results of somatic incompatibility tests for S. granulatus the same location in 1987 and 1989, respectively; (ii) VT1983; isolates from the Poverty Creek site. Encircled isolates were somat- (iii) VT1996; and (iv) VT1989. However, the lack of transi- ically compatible. The ectomycorrhizal tree associates, P. strobus, tiveness between isolates VT1990, VT1993, VT1994, and are also indicated. Downloaded by guest on October 1, 2021 Population Biology: Jacobson et aL Proc. Natl. Acad. Sci. USA 90 (1993) 9161 Table 1. Data matrix of the RAPD markers identified from population isolates by primers OPA-7 (A7), OPB-18 (B18), OPB-10 (B10), OPA-13 (A13), and OPA-16 (A16). Population isolate Marker 1984 1995 1993 2006 1994 1996 1989 1983 1990 A13:300 0 0 0 0 1 0 0 9 0 A13:350 0 0 0 0 0 1 0 9 0 A13:380 0 0 1 0 1 1 1 9 1 A13:400 0 0 1 0 0 0 0 9 0 A13:430 0 0 1 0 1 0 1 9 0 A13:470 0 0 1 0 0 1 1 9 0 A13:500 1 1 1 1 1 1 1 9 1 A13:630 0 0 0 0 0 0 1 9 0 A13:650 0 0 1 0 0 1 1 9 1 A13:900 1 1 1 1 1 1 1 9 1 A13:1050 1 1 0 0 0 0 1 9 1 A13:1100 1 0 1 1 0 1 1 9 1 A16:420 0 0 0 0 1 0 0 0 0 A16:470 0 0 0 0 0 0 1 0 0 A16:700 1 1 1 1 1 1 1 1 1 A16:850 1 1 1 1 0 0 1 1 1 A16:1050 1 1 1 1 0 0 1 1 1 A16:1400 1 1 0 1 0 1 1 1 0 A16:1500 1 1 0 1 0 0 0 0 0 B18:280 0 0 0 0 1 0 0 0 0 B18:330 0 0 0 0 1 0 0 0 0 B18:350 0 0 0 0 1 0 0 0 0 B18:390 0 0 0 0 1 0 0 0 0 B18:450 1 0 0 1 1 0 0 0 0 B18:480 1 1 1 1 0 1 1 1 1 B18:500 0 1 0 0 1 1 1 0 0 B18:530 0 0 0 0 1 0 0 0 0 B18:580 0 1 0 0 0 0 0 0 0 B18:630 1 0 1 1 1 1 1 1 1 B18:700 0 0 0 0 1 0 0 0 0 B18:760 1 1 1 1 1 0 1 0 0 B18:820 1 1 0 1 0 0 1 0 1 B18:950 0 0 0 0 1 0 0 1 0 B18:1100 1 1 0 1 0 1 0 0 1 B18:1200 1 1 1 1 1 1 1 1 1 B18:1300 1 1 1 1 1 1 1 1 1 A7:500 1 1 0 1 0 0 0 0 1 A7:600 1 1 1 1 0 0 1 1 0 A7:680 1 1 0 1 0 0 0 0 1 A7:720 1 1 1 1 0 1 1 1 1 A7:800 0 0 0 0 1 0 0 0 0 A7:850 0 0 0 0 0 1 0 0 0 B10:520 0 0 0 0 1 0 0 0 0 B10:600 0 0 0 0 1 0 0 0 1 B10:700 1 1 1 1 0 1 1 1 1 B10:820 1 1 1 1 0 1 1 1 1 B10:1300 1 1 1 1 1 1 1 1 1 0, Marker not present; 1, marker present; 9, data missing. Marker numbers refer to the primer used and band mobility (in bp). dice coefficient (equivalent to the band sharing coefficient), ilar); and isolates VT1993 and VT1989 (85% similar). Isolates was used to construct a phenogram via the unweighted pair VT2006 and VT1984 were most similar, with only one dif- group method (23). The phenogram represented an excellent ference at the 47 markers examined (98% similar). fit of the data, as determined by the cophenetic correlation Comparing the Two Methods. The data from the somatic coefficient (r = 0.97078). incompatibility tests were used to construct a similarity RAPD marker analysis indicated that this population was matrix using the simple matching coefficient (23). A pheno- composed of at least eight genotypes exhibiting various gram was constructed using the similarity matrix via the degrees of presumptive genetic relationship. VT1994 and unweighted pair group method of NTSYS (23). To quantify the VT1983 were the most genetically distinct genotypes in the agreement between the somatic incompatibility tests and population, having, respectively, only 39% and 56% similar- RAPD marker analysis, the product-moment correlation for ity with other isolates from the population at the loci exam- the phenograms was computed by using the matrix compar- ined. The remaining isolates were at least 70% similar to one ison option of NTSYS (23). There was no correlation between another and clustered around two regions of high genetic the similarity matrices (Fig. 4) obtained from the two analyses similarity: isolates VT1984, VT2006, and VT1995 (93% sim- (r = 0.01). It is obvious then that the two methods produced Downloaded by guest on October 1, 2021 9162 Population Biology: Jacobson et al. Proc. Natl. Acad Sci. USA 90 (1993)
Primer OPB-18 Primer OPA-13 Primer OPA-07
v- " CD XO 0O COCo CG <0 F- F- CD x0 U O Cd U: Co X w o Co m X C C ~o ~cC co c cD c 4 Kbp 4 CA X O v s X co c vl QCO C;) Cl) bP4 C: cow 4, A0 VI CS a) 4
1.5 1.0
0.6 F
FIG. 3. RAPD markers identified with primers OPA-07, OPA-13, and OPB-18 from nine isolates from the Poverty Creek site. RAPD reaction mixtures of25 y4 contained 7-9 ng ofgenomic DNA; 250 ,uM each dATP, TTP, dGTP, and dCTP; 47 ng ofrandom l1-mers (Operon Technologies, Alameda, CA); and 0.5 unit of Taq DNA polymerase (Promega). Amplifications were conducted in a Perkin-Elmer thermocycler with an initial 5-min denaturation cycle at 95°C, followed by 45 cycles of 37°C for 1 min (annealing), 36°C for 1 min (polymerization), and 93°C for 1 min (denaturation). Electrophoresis of the 25-pl reaction volumes was in 0.7% agarose with 1% Synergel (Diversified Biotech) and 1x TAE buffer. Gels were run for 5 hr at 100 mV and stained with ethidium bromide. significantly different distributions of presumptive individu- somatic incompatibility tests with ectomycorrhizal fungi is als in the population. necessary. Somatic incompatibility tests are known to be Somatically compatible isolates VT1984 and VT2006, from ineffectual for determining genetic identities of closely re- the same location, were genetically most similar, with only lated individuals (2, 6, 7, 10, 16, 24, 25). Dikaryons synthe- one RAPD marker difference. None ofthe other isolates that sized from monokaryotic spores from a single basidiocarp were somatically compatible was genetically similar to one (sib-composed dikaryons) may or may not be compatible with another. Significantly, VT1994, which was compatible with each other and the parent dikaryon. Levels ofincompatibility VT1993 and VT1995, was least similar to all other isolates are apparently species specific but were as high as 50%o in the examined in the population. Thus, while these isolates are heterothallic, tetrapolar Basidiomycetes examined (inbreed- somatically compatible, they are clearly not of the same ing potential of 25%). In most of these studies, populations genotype. were composed of numerous different individuals as deter- mined by high levels of somatic incompatibility between isolates. However, based on the backcrosses conducted with DISCUSSION sib-composed and parent dikaryons, these authors cautioned Results from the somatic incompatibility tests with S. gran- that somatic incompatibility might erroneously indicate ge- ulatus were easily scored as the presence or absence of a netic identity of closely related but genetically different barrage zone and were completely reproducible. Somatic individuals if they had occurred in the natural populations. incompatibility, as determined by the presence of a barrage Mating systems in ectomycorrhizal fungi have not been zone, is thus a real physiological response occurring between well studied primarily because of difficulties encountered different S. granulatus individuals in vitro. with spore germination. However, there are numerous rea- Somatically incompatible isolates were accurately identi- sons to suspect that Suillus populations, if not all ectomy- fied as genetically different individuals. However, our study corrhizal fungi, may be composed of closely related individ- also showed that while compatible isolates may have high uals, which would render somatic incompatibility testing genetic similarity as determined by RAPD marker analysis invalid for distinguishing different genotypes. To date, the (isolates VT1984 and VT2006, 98% similarity), they may be mating systems of S. granulatus (Swedish isolates only), S. genetically quite distinct as well (isolates VT1994, VT1995, luteus, and S. bovinus are known to be heterothallic and VT1993, and VT1990, 39-70%o similar). Somatically compat- bipolar (26, 27), resulting in an inbreeding potential of 50%. ible isolates of S. granulatus are thus not necessarily genet- In addition, it is unknown but suspected that ectomycorrhizal ically identical or similar individuals, clones, or genets. Basidiomycete individuals are perennial, living for many An examination of the studies in which somatic incompat- years in association with the roots of the host plant. An old ibility tests were used to describe root-pathogenic and wood- ectomycorrhizal population may thus be composed of nu- decomposing Basidiomycete populations suggests that a re- merous sexual generations of individuals coexisting at one evaluation of the basic premise for the appropriate use of time, exhibiting various levels of genetic relatedness. RAPD markers VT 1984 Somatic Suillus species are also known to possess dikaryotic in ncompatibility addition to monokaryotic spores (28). Studies with S. gran- ,