Isozymes and the Genetic Resources of Forest Trees

Genetic data are an essential prerequisite for analysing the genetic structure of tree populations. The isozyme technique is the best currently available method for obtaining such data. Despite several shortcomings, isozyme data directly evaluate the genetic resources of forest trees, and can thus be used to monitor and manipulate these resources. For example, preliminary isozyme data indicate that domestication of cultivated species has generally reduced variation within populations. Geographic differences among natural populations are less evident in trees than in herbaceous plants, but need to be considered in sampling and conservation strategies. In dealing with remnant population size (Nr) and distribution, management programs should recognize that the half-life of heterozygosity is about 1.4 Nr generations, whereas the number of trees required to retain half the current alleles after a bottleneck is about the square root of the original population size. T growing awareness of two major theses provide the occasion of this opening contribution to the symposium. The first is that the genetic resources of economic plant species, and in particular of forest trees species, are finite. Yet, these resources have been managed without concern for their conservation or renewal. It is now clear that our remaining forest resources must be used prudently to guarantee a long future for managed species. The second thesis is that the isozyme technique is well suited for monitoring and assisting the manipulation of our genetic heritage, at least in part (Brown 1978). The last decade has witnessed the growth and impact of this technique in population biology. The ensuing contribu­ tions to this symposium will amply demonstrate the ways in which electrophoretic data can contribute to our genetic knowledge of populations of forest trees. This paper first considers the properties of isozyme data and then discusses how the data can be used to monitor and conserve the genetic resources of forest trees. ISOZYMES: BONANZA OR BLIND ALLEY IN FOREST GENETICS? Phenotypic as Opposed to Genetic Evaluation Two radically different approaches can be taken when measuring differences among individuals, populations or species. The first we define as the "phenotypic" approach. It is motivated by the argument that it deals directly with characters of economic or biological importance. The ideal of this approach is to know the phenotypic expression of 1 Presented at the Symposium on Isozymes of North American Forest Trees and Forest Insects, July 27, 1979, Berkeley, Calif. Principal Research Scientist, Commonwealth Scientific and Industrial Research Organization (CSIRO), Division of Plant Industry, Canberra, Australia. Experimental Officer, CSIRO, Division of Forest Research, Canberra, Australia. yield, quality, and pest and disease resistance in all possible environments. Clearly, this ideal is unattainable, because the need for decisions limits, in space and time, the number of environments that may be tested. Quantitative genetics attempts to overcome these limitations by using a predictive model. The model assumes that unspecified polygenes can simulate the genetic and developmental complexities that underly phenotypic measurements and their variation. Various classes of effects on phenotypic variation—additive, dominance, epistasis, genotype x environment interaction—are defined (Libby and others 1969). The predictive power, however, is severely limited to the genotypes tested, the environments experienced, and the mode of measurement. Of course phenotypic (agronomic or silvicultural) assessment and phenotypic selection are essential in any plant breeding program. It is less certain whether such data provide valid measures of genetic variation and genetic similarity among popula­ tions. The second approach to evaluation, the "genetic" approach, is entirely different in emphasis. Its aim is to detect genetic differences as close as possible to the DNA (deoxyribonucleic acid) level. The ideal of this approach is to know all the DNA sequences. Such basic data would form direct measures of how genetically variable is one population, and how similar it is to another, entirely free of environmental effects. Data from different species would be directly comparable. Again, this ideal is far from being a reality. Perhaps it may be attainable for very restricted classes of DNA (for example, organellar DNA, highly repeated DNA) or in very simple organisms (Smith 1979). But the need for "genetic," as distinct from "phenotypic," evaluation is apparent. Because we cannot presently measure variation directly at the DNA level, we must lay down criteria for the choice of techniques that are currently feasible. Table 1 summarizes four criteria first suggested by Hubby and Lewontin (1966; see also Lewontin 1974). These criteria require that the effect of an allelic substitution at a locus be detectable (C), and distinguishable from the substitution of any other allele at the same locus (A), or other loci (B). Table 1—The Hubby-Lewontin criteria for genetic markers, and the properties and pitfalls of the isozyme technique with respect to each criterion Criteria to be met Properties of isozymes Problems occurring A. Alleles are distinct in individuals Codominance; free of Genetic or environ epistatic or environmental mobility; mental effects mobility modification B. Effects are locusspecific Enzyme specificity Nonspecific assays; duplication or polyploid C. All substitutions are detectable Charge-state separaDetectable class of tion, irrespective of base substitutions is function or variation restricted and biased in variation D. Loci are sampled randomly Assay availability is Loci assayed are a independent of limited class and variation differ in variation Further, the sample of loci studied should be random, irrespective of function, or likely level of polymorphism (D).