TREE vol. 4, no. 1, January 1989

POPULATION GENETICS is basic to an understanding of genetic resource conservation, population biology and evolution. Less certain is its im- portance in practical plant breeding, particularly in the present era of re- combinant DNA technology. At the University of California, Davis, in August 1988 a symposium was held on Population Genetics and Germ- plasm Resources in Crop Improve- ment. This provided an opportunity to re-examine the current input of population genetics to plant breed- ing. The symposium was held, in part, to honour Professor Robert W. Allard, who recently retired from the chair of genetics at Davis. Since the publication in 1960 of his book, Prin- ciples of Plant Breeding’, Allard’s re- search and ideas have significantly influenced those of plant breeders and population geneticists alike. Many of the speakers at the symposium were Allard’s former students or students of former students.

Population Genetics and P lant Breeding: Homage to R.W. A llard

The symposium began with a ses- sion on kinds of genetic diversity in plant species. In turn, J.L. Hamrick (Athens, GA), P. Gepts (Davis) and M.T. Clegg (Riverside, CA) reviewed allozyme, seed storage protein and molecular diversity in plant species. There is now a rich data base for plant allozyme diversity from which it has emerged that long-lived, out- crossed and geographically wide- spread species possess the highest levels of variation within popu- lations, while early successional annuals partition most variation be- tween populations.

In contrast, very little is known about intraspecific molecular diver- sity in plants. The few studies con- ducted indicate that restriction frag- ment length polymorphism (RFLP) detected by nuclear genomic and complementary DNA probes, plus variation in DNA content, is wide- spread, while chloroplast DNA is highly conserved. An exciting recent discovery is that the human mini- satellite probe 33.6 detects several RFLPs in varieties of Asian and Afri- can rice, allowing DNAfingerprinting of rice cultivars; a poster reporting this work was presented by J.F. Dallas (Columbia, MO).

B.S. Weir (Raleigh, NC) opened a discussion on the genetic structure and geographic organization of populations with a summary of re- cent improvements in the statistics used for estimating gene diversity

Richard Abbott is at the Dept of Biology and Preclinical Medicine, University of St Andrews, St Andrews, Fife, KY16 9TH, UK.

Richard Abbott

0 1989, Elsev~er Smnce PublIshem Ltd (UKl0169-5347i891$02 00

based on allozyme and DNA se- quence data. The genetic structure of a species is to a large extent affected by its mating system; in plants, mat- ing systems are very diverse. A.H.D. Brown (Canberra) reviewed the util- ity of the mixed mating model orig- inally developed by him and Allardz for estimating outcrossing rates of plants. The model has been widely used but is not designed to deter- mine the paternity of outcrossed offspring. Methods for this are avail- able but require many more genetic markers than are currently available in most wild populations.

The importance of gene flow and selection in determining patterns of genetic diversity among plant popu- lations was discussed by S.K. Jain (Davis). He emphasized that al- though there are many examples of population differences for single gene and quantitative traits, in only a few instances have such differences been proved to result from selection.

Jain’s paper served as an intro- duction to a session devoted to microevolutionary processes. R.A. Ennos (Edinburgh) presented a com- prehensive analysis of the strengths and weaknesses of population genetic and biometrical genetic approaches to the detection and measurement of selection in plant populations. He argued that such approaches fail to be predictive and that a more ecologically based approach is demanded which recog- nizes the widespread occurrence of genotype X environment interac- tions. Only by focusing attention on an ecological situation, and resolv- ing the relevant genetic variation ex- pressed in the wild, will accurate prediction displace adaptive story- telling.

The population genetics of host- pathogen systems, and of popu- lations formed following coloniz- ation, were discussed in turn by J.J. Burdon (Canberra), and S.C.H. Barrett and 6. Husband (Toronto). Burdon emphasized the need to measure fitness costs associated with resistance and virulence genes in the host and pathogen respectively. Such costs remain virtually un- explored despite their obvious im- portance. Barrett and Husband made clear that an accurate picture of the genetic effects of plant colonization hinges on a detailed knowledge of the colonization event. When suf- ficient initial genetic variability is present in a population, and strong

epistatic effects occur between cer- tain loci, selection might ultimately lead to the evolution of specific multilocus genotypes and therefore the presence of significant l inkage disequilibrium. This possibility was considered in turn by A.M. Hastings (Davis), and J.P. Adams and P. Gross0 (Ann Arbor, Ml). Multi locus associations are common in pre- dominantly self-pollinated plants but not in Drosophila except for loci that are very tightly linked. A simulation model developed by Adams and Gross0 showed that the conditions for the evolution of significant link- age disequilibrium are normally very restricted and, except in predomi- nant selfers, rarely occur.

The final session of the sym- posium was devoted to genetic diversity and its utilization in plant im- provement. D.R. Marshall (Adelaide) emphasized the importance of popu- lation genetics in a discussion of the acquisition, multiplication and re- generation of germplasm for use in crop improvement, and then raised the unanswerable question of where to draw the limits to germplasm conservation in an age when recom- binant DNA technology allows trans- fer of genes from species to crop plants irrespective of relationship. O.A. Muona (Oulu, Finland) followed up with a clear statement that the population genetics of a species is still of crucial importance in forest tree improvement where artificial breeding populations closely re- semble their natural counterparts. Finally, a description by C.W. Stuber (Raleigh) of the recent use of isozyme and RFLP markers for locat- ing and manipulating quantitative trait loci in maize breeding, although not integrated with the population genetics theme, made it very ap- parent how useful such markers will be in examining the population genetics of quantitative traits in natural populations.

During the symposium a special session was held to honour Profes- sor Allard. First F.A. Bliss (Wiscon- sin) described the outstanding work of his group that has led to signifi- cant improvement in the cultivated common bean through the use of genetic resources present in wild re- lated material. This was a fitting con- tribution given Allard’s long stand- ing interest in beans. Next, Allard himself reviewed his work on plant population genetics over the past 30 years. Here a direct link between

TREE vol. 4, no. 7, January 1989

plant population genetics and plant breeding was made obvious. How- ever, it was plant breeding in a pre- recombinant DNA era, with empha- sis on composite cross populations and a belief that natural selection might produce a balanced popu- lation of genotypes which, in terms of disease resistance and yield, would surpass that of specific varieties produced by man. In reality,

this never happened. It remains true that an understand-

ing of the population genetics of a species is fundamental to efficient germplasm conservation, which in turn will continue to underpin crop improvement even in the current era of recombinant DNA technology. Mainly due to Allard and his associ- ates, approaches to investigating the population genetics of plants are

now well defined. In future, such approaches will be further integrated with ecology and widely applied in studies of plant population biology.

References 1 Allard, R.W. (1960) Principles ofPlant Breeding, Wiley 2 Brown,A.H.D. andAllard,R.W. (1970) Genetics66,133-145

Animal M itochondrial DNA as a Genetic Marker in Population and Evolutionary Biology Richard G. Harrison

Animal mitockondrial DNA (&DNA) is playing an iflcreasingly important role us a genetic marker in population and evol- utionary biology. The popularity of this molecule derives, in part, from the relative euse with which clearly homologous se- quences can be isolated and compared. Simple sequence organization, maternal inheritance and absence of recombinatioti make mtDNA an ideal marker for tracing maternal genealogies. Rapid rate of se- quence divergence (at least in vertebrates) allows discrimination of recently diverged lineages. Studies of mtDNAs from u diver- sity of animal groups have revealed signifi- cant variation among taxa in mtDNA sequence dynamics, gene order and genome size. They have also provided import& insights into population struc- ture, geographic variation, zoogeography and phylogeny.

The ability to assess DNA se- quence divergence among indi- viduals, populations, species or higher taxa has revolutionized sys- tematic and evolutionary biology. Currently available techniques dif- fer in the nature of the information they provide and in the proportion of the genome that can be assayed. Complete sequencing of homolo- gous DNA fragments from different organisms provides the most powerful and direct method for obtaining information on amount of

Richard Harrison is at the Section of Ecology and Systematics, Corson Hall, Cornell University, Ithaca. NY 14853, USA.

6

genetic variation or extent of gen- etic divergence. Each nucleotide base pair represents a character that can exist in four discrete character states, and the number of characters is effectively limited only by the time and money avail- able for sequence analysis. Use of the polymerase chain reaction (PCR) to amplify homologous DNA sequences from many individuals or taxa should lead to a rapid ex- pansion of the DNA sequence data base.

At the other end of the spectrum, DNA-DNA hybridization gives a single estimate of genetic distance averaged over the entire genome. This approach may avoid problems that arise from examining only a small portion of the genome, but it is limited in other ways (e.g. it does not produce discrete character state data).

information about mtDNA vari- ation in natural populations has come principally from comparisons of restriction enzyme fragment pat- terns and site mapsi-4. Restriction endonucleases cleave double- stranded DNA at specific recog- nition sequences, producing a series of fragments, the number and sizes of which vary depending on where within the DNA the recog- nition sequence occurs. Thus frag- ment pattern differences can conveniently serve as markers of genetically distinct l ineages. Simple measures of genetic dis- tance can be derived from com-

parisons of restriction site maps or restriction fragment patterns. Alternatively, each site (or frag- ment) can be treated as a charac- ter having two states (present or absent), thereby providing a data set appropriate for phylogenetic analysis5.

Characteristics of animal mitochondrial DNA

The small, closed-circular mito- chondrial genomes of animals are clearly homologous DNA se- quences3,6. Since pure samples of mtDNA can be prepared from small amounts of tissue with relative ease, it is straightforward to com- pare homologous mtDNA se- quences from a wide variety of organisms. Although total nuclear DNA is easy to prepare, isolating hon-4ogous nuclear sequences is more difficult and has typically necessitated constructing and screening genomic libraries for each individual or species in- volved. (Selective amplification using PCR provides an important new shortcut.)

Gene content and genome organization Animal mtDNA exhibits remark-

able conservation of gene content. MtDNA molecules from ver- tebrates, insects and echinoderms include two ribosomal RNA (rRNAI genes, 22 transfer RNA (tRNA) genes, and I3 genes that code for proteins involved in electron trans- port or ATP synthesis3.6, (Fig. 1 I. Each mtDNA molecule also has a control region containing se- quences that function in initiation of replication and transcription. Gene order is conserved among vertebrates, but comparisons across phyla indicate that major rearrangements have occurred’. The mtDNA molecule in animals is exceptionally compact, with few in-