Heredity 55 (1985)335—339 The Genetical Society of Great Britain Received 20 March 1985

Genetic variation and selection for fertility in the fungus Cochliobolus heterostrophus

James A. Kolmer and Department of Plant Pathology, North Carolina State Kurt J. Leonard University; United States Department of Agriculture, Raleigh N.C. 27695-7616, U.S.A.

The response to selection for perithecial number in Cochliobolus heterostrophus, a heterothallic ascomycete, is described. Thirteen MATA isolates and 12 MATa isolates of the fungus were mated in all possible combinations and evaluated for fertility. Progeny from some of these matings segregated for a defective perithecial allele. Selection for increased numbers of mature perithecia was carried out for six generations of intermatings among four to six parents in each generation which were selected from approximately 24 progeny from the preceeding generation. After six generations the numbers of perithecia had increased by nearly a factor of four. The proportion of additive genetic variance decreased linearly in the selection generations. The increase in perithecia was confirmed in concurrent mating tests among subpopulations from each selection generation.

INTRODUCTION Selection experiments have been used to demonstrate polygenic variation for various traits Quantitativegenetic variation has received little in fungi. In Aspergillus nidulans, penicillin titer was attention in studies of the genetics of plant doubled after four generations of recurrent selec- pathogenic fungi although it is thought to be tion (Merrick, 1976). Selection in both directions important in determining the parasitic abilities of over several generations for hyphal growth rate these organisms. Cochliobolus heterostrophus was effective in Neurospora crassa (Papa, Srb and Drechsler, the causal agent of southern corn leaf Federer, 1966). Similar results were found for blight, is a heterothallic fungus with two mating growth rate in Schizophyllum commune (Simchen, type alleles, MATA and MA Ta, required for sexual 1966). The response to selection in these studies reproduction in pairings between haploid isolates was evidence of polygenic variation influencing (Nelson, 1957). Nelson (1959a) observed differen- these traits. The present study was undertaken to ces in numbers of perithecia produced in matings determine the nature of the genetic control of of sexually compatible isolates of the fungus, but perithecial number in Cochliobolus heterostrophus did not quantitatively analyse these heritable Drechsler. The extent of polygenic variation pres- differences in fertility. Nelson (1959b) later iden- ent in the fungus which influences fertility was tified a single gene in naturally occurring isolates tested by means of recurrent selection for increased that can eliminate or greatly reduce perithecial perithecial number. production when it is present in both members of a pair of sexually compatible isolates of the fungus. In studies with other species of Cochliobolus, MATERIALS AND METHODS Nelson and others demonstrated that fertility is controlled by genes other than the two alleles at Initial population the mating type locus (Nelson, 1961; 1964; 1970; Thirteen MATA and 12 MATa isolates were Kline and Nelson, 1968), but the role of polygenic chosen from a group of conidial isolates collected variation was not rigorously examined. from diseased corn leaves from fields in North Paper number 9714 of the journal series of the North Carolina Carolina or from first generation ascospore pro- Agricultural Research Service. geny obtained from crosses of such isolates. The 336 J. A. KOLMER AND K. J. LEONARD

25 isolateswere paired in all possible combina- Table 1 North Carolina Design II. Sources of variation and tions. All parental and progeny isolates in this expected mean squares study were stored at 4°C in potato dextrose agar (PDA) slants. The fungal isolates were grown on Source d.f. Expected mean square PDA containing 10 g dextrose per litre for 4 days Isolates MATA A-i V+(4)VA(4a)VMATA after which small (2-4 mm) blocks of agar with Isolates MATa a-i V,+(4)Vfl

(I) Table 2 North Carolina Design II. Analysis of variance for a 180 number of perithecia in initial and selection populations U- 160

UJ 140 GenerationSource -J 120 of of Mean selection variation d.f. square 0Lii 0 Isolates MATA 12 14,084* 0 60 11 LU Isolates MAYa 13,026* 40 MATAXMATa 120 614* I-.I 20 M.S.E.t 454 216 0LU 1 Isolates MATA 13 3,840* 1 2 3 4 5 6 /// Isolates MAYa 11 18,835* GENERATIONSOF SELECTION MATAxMATa 131 1,018* Figure 1The response to selection based on additive genetic M.S.E. 457 436 variance over six generations. 2 Isolates MATA 14 7,196* Isolates MATa 8 9775* MATAxMATa 107 1,347* is presented in table 2. The means, selection differ- M.S.E. 356 748 entials, and components of genetic variation are 3 Isolates MAYA 12 5,084* presented in table 3. With six cycles of selection Isolates MAYa 11 13,628* the mean number of perithecia per leaf disk MATAXMATa 127 1,735* increased from 90 to 165 and the standard errors M.S.E. 449 842 ranged from 024 to 326. 4 Isolates MAYA 10 8,031* The selected isolates in the initial population Isolates MAYa 10 20,712* all had the effective perithecial allele, thus fixing MAYAXMAYa 88 3,712* the allele in the subsequent selection generations. M.S.E. 315 1,467 All selection after the first cycle utilised polygenic 5 Isolates MAYA 12 2,465* variation. The proportion of additive genetic vari- Isolates MAYa 10 6,582* MAYA XMAYa 107 1,791* ance relative to total genetic variance in the initial M.S.E. 394 767 and selected populations decreased linearly from 083, to 023 over the six generations (fig. 2). 6 Isolates MAYA 11 13,689* Isolates MAYa 6 21,573* MAYA XMATa 53 7,962* Evaluationof subpopulations in M.S.E. 210 1,033 common environments * Significantat 005. tMeansquare error. Subpopulationsconsisting of isolates randomly sampled from the initial and selected populations were evaluated in three separate tests shown in fig. confirmed in these tests. In a separate test, sub- 3. The standard errors of mean numbers of pen- populations from generations one and five were thecia in the three tests ranged from 32 to 66. In evaluated in a common environment to estimate general, the trend of increasing numbers of pen- the difference in the number of effective factors thecia with succeeding generations of selection was influencing perithecial number between selected

Table 3 Means, selection differentials, and components of genetic variation of initial and selection generations

Additive x Generations Mean number Additive additive of of Selection genetic genetic selection perithecia/leaf diskdifferential variance variance

0 44±0•59 20 539 99 1 90±084 36 376 145 2 71±024 30 392 149 3 100±1•18 12 321 222 4 129±336 10 484 561 5 123+121 24 107 255 6 165±191 — 513 1732 338 J. A. KOLMER AND K. J. LEONARD

0Ui effective factors between isolates of generations z one and five. > 0 I- uJ 1.0 wz DISCUSSION a .90 .98 v = - x .80 a4 .1025 In our tests, the numbers of mature perithecia per 0 I- .70 leaf disk were determined to be under the control 60 of the perithecial gene locus and polygenic effects. I)w z .50 Nelson (1959b) described as recessive an allele 40 that prevented perithecial production in Coch- > 30 liobolus heterostrophus, however he did not quanti- 0 tatively distinguish between the numbers of pen- zUi 20 thecia formed when one or both paired isolates aUi .10 Ui had the effective perithecial allele. We have shown > 0 that the perithecial allele acted additively in deter- a mining numbers of perithecia. GENERATIONSOFSELECTION Thenumber of mature perithecia per leaf disk Figure 2 Change in additive genetic variance relative to total in matings among isolates from the sixth gener- genetic variance over six generations of selection based on ation was nearly four times greater than the additive varEance. numbers in matings among isolates of the initial population. This increase is similar to selection 220 gains in experiments with penicillin production in (I) Aspergillus nidulans (Merrick, 1976) and growth 0200 180 rate in Neurospora crassa .(Papa,Srb, and IL 160 Federer, 1966) and Schizophyllum commune (Sim- IL chen, 1966). A realised heritability estimate of 074 -J140 .—_...._.* for the number of perithecia per leaf disk was Er120 / IL obtainedfrom the slope of the regression of the a-100 / meannumber of perithecia per generation on the 80 / accumulatedselection differential (table 3) (Fal- 060 IL coner, 1980). Because of the small population sizes, I-I 40 random fixation of some alleles affecting the num- Er 20 IL ber of perithecia could have occurred; however a- ______the intensity of selection should have been o 1 2 3 4 5 6 sufficient to transcend the effects of drift in increas- GENERATIONS OF SELECTION ing perithecial number. Perhaps with larger popu- lations the decline in the proportion of additive KEY .—.TEST#1 TEST#3 genetic variance would have been less dramatic. We did not consider reverse selection for lower TEST#2 numbers of perithecia as fertility would be reduced Figure 3 Evaluation of subpopulations in common environ- to a level that would prohibit ascospore formation ments from the initial population, and all selected in the selected crosses. generations in three separate tests. The rate of increase in fertility appeared to be greater between the initial population and fourth isolates in these two generations. Two individual selection generation than between the fourth and isolates from generations five and one had mean sixth generations (fig. 1). Continued selection for numbers of 2035±927 and 9063±684 pen- perithecial number should be possible since there thecia per leaf disk, respectively. The combined is an appreciable proportion of additive genetic estimate of additive genetic variance of 561 was variance remaining in the sixth generation of selec- obtained for the two generations. Burnett's (1975) tion. Our selection scheme, based on selecting the formula for estimating the difference in number of most fertile MATA isolates and the most fertile effective factors in haploids, 1/4 (High extreme— MATa isolates is similar to selection based in Low extreme)2/Additive Genetic Variance, was general combining ability (Falconer 1980). Both used to calculate an estimated difference of 57 of these selection schemes act on additive genetic SELECTION IN C. HETEROSTROPHUS 339

variance affecting the trait of interest. As the num- some genetic recombination occurs in natural ber of perithecia per leaf disk increased through populations (Leonard, 1973) but other mechan- the generations (fig. 1), the proportion of additive isms such as heterokaryosis (Leach and Yoder, genetic variance decreased (fig. 2). The reduction 1982) might account for it. The low fertility of in the proportion of additive genetic variance can wild-type isolates relative to that attained in our be attributed to fixation of alleles affecting num- selected populations may reflect the limited oppor- bers of mature perithecia. In the later selection tunities for sexual reproduction to occur in nature. generations, additive x additive variance accoun- Perithecial production and maturation of asco- ted for nearly all of the genetic variance present. spores are inhibited by exposure to light at even This agrees with Simchen (1966) who observed a low intensities. Because it takes 3 weeks for pro- similar reduction in additive genetic variance duction of mature ascospores, there are likely to among haploid lines of Schizophyllum commune in be few situations in nature in which the process a population subjected to recurrent selection for would not be interrupted before the perithecia or rate of mycelial growth. In the population selected ascospores matured. for high rate of mycelial growth, additive genetic variance decreased from 25 to zero within 10 gener- ations. In the population selected for slow growth rate,additivegeneticvariance remained REFERENCES unchanged over the 10 generations. Fluctuations in the apparent rates of increase BuRNETr,j.H.1975. Mycogenetics. John Wiley and Sons, of fertility (fig. 1) in our experiments as well as London. EBERHART, S. A., MOLL, R. H. AND COCKERHAM, C. C. 1966. differences between different tests of isolates from Epistaticand other genetic variances in two varieties of the same generation (fig. 3) can be attributed to maize. CropScience,6, 275—280. environmental variation. Specifically, different FALCONER, D. s. 1980. Introductionto Quantitative Genetics. batches of Sach's agar and leaf disks used in differ- LonghamInc. New York. ent tests could have affected the expression of KLINE, D. M. AND NELSON, R. i. 1968. Variation in mating capacities among 103 isolates of Cochliobolussativus. perithecial numbers. The three tests using sub- Phytopathology, 58, 1055. populations of different generations evaluated in LEACH,J. AND YODER. 0. c. 1982. Heterokaryosis in Coch- common environments served as a control measure liobolusheterostrophus. Experimental Mycology, 6, 364-374. for the selection experiment. These experiments LEONARD, K. .j.1973.Association of mating types and virulence in Helminthosporium maydis, and observations on the origin confirmed the trend of increasing numbers of pen- of the race T population in the United States. thecia with succeeding generations. The estimated Phytopathology, 63, 112-115. difference of 57 effective factors influencing num- LEONARD, K .t. 1978. Polymorphisms for lesion type, fungicide bers of perithecia between generations one and tolerance, and mating capacity in Cochliobolus carbonum isolates pathogenic to corn. Can. J. Bot., 56,1809—1815. five is consistent with the rapid increase in pen- MERRICK, M. i. 1976. Hybridisation and selection for penicillin thecial number due to selection, and the rapid productioninAspergillus nidulans—a biometrical decrease in additive genetic variance, both of approach to strain improvement. In: Genetics of Industrial which indicate that a relatively small number of Microorganisms.Editedby K. D. MacDonald. Academic Press, New York. polygenic units influence numbers of perithecia. NELSON, R. R 1957. Heterothallism in Helminthosporium Our estimate is similar to Simchen's (1966) esti- maydis. Phytopathology, 47,191-192. mates of effective factors for differences in growth NELSON, R. R. 1959a. Genetics of Cochliobolus heterostrophus rates between high and low selection lines of S. I. Variability in degree of compatibility. Mycologia Vol. Ll commune. His estimates ranged from 55 to 929 No. 1, pp. 18—23. effective factors. NELSON, R. R. 1959b. Genetics of Cochliobolusheterostrophus. IV.A mutant gene that prevents perithecial formation. The sexual stage of C. heterostrophus can be Phytopathology,49,384-386. induced readily in the laboratory, but it has rarely NELSON, R. R. 1961. Evolution of sexuality and pathogenicity. been found in nature. Reproduction of the fungus II. A comparison of the pattern of sexuality in Cochlioblus in nature is primarily asexual by wind-dispersed victoriaeand related species. Phytopathology,51, 222-223. NELSON, R. R 1964. Inhibition of perithecial formation in conidia. Although sexual reproduction seems to Cochlioboluscarbonum. Phytopathology, 54, 876-877. be unimportant in the life cycle and the levels of NELSON, R. R. 1970. Variation in mating capacities among fertility among field isolates of C. heterostrophus isolates of Cochliobolus carbonum, Can. J. Bot., 48,261-263. are low, the fungus has retained polymorphisms PAPA, K. E, SRB, A. M. ANI) FEDERER, W. T. 1966. Selection for increased growth rate in inter- and intra-strain crosses for mating type and the major genes affecting pen- of Neurospora. Heredity, 21, 595—613. thecial production as well as quantitative variabil- SIMCHEN, G. 1966. Monokaryotic variation and haploid selec- ity. Circumstantial evidence indicates that at least tion in Schizophyllum commune.Heredity, 21, 241-263.