MSU LIBRARIES “ mm 719?)? CL“. ’0‘." J (Cu [2&1 J..- —---.-.W RETURNING MATERIALS: P1ace in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. ALLOZYME DIFFERENTIATION BETWEEN GENE POOLS IN COMMON BEAN (Phaseolus vulgaris L.), WITH SPECIAL REFERENCE TO MALAWIAN GERMPLASM BY Susan Louise Sprecher A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Plant Breeding and Genetics Program/Horticulture Department 1988 ABSTRACT ALLOZYME DIFFERENTIATION BETWEEN GENE POOLS IN COMMON BEAN (Phaseolus vulgaris L.), WITH SPECIAL REFERENCE TO MALAWIAN GERMPLASM BY Susan Louise Sprecher Domesticated types of Phaseolus vulgaris L., the common bean, are differentiated into two sub-specific taxa, originating in Central America and the eastern Andes, and differing in quantitative and biochemical characters. They are separated by genetic incompatibility mechanisms. During this study of isozyme variability in Malawian bean germplasm, it was found that alternate alleles at certain isozyme loci are specific to each of the two gene pools. These results from starch gel electrophoresis (SGE) were confirmed for six to eight loci in three populations of domesticated beans: 373 landrace lines sampled from Malawi; improved cultivars of commercial classes from both gene pools; and parental lines known to produce F hybrid weakness or lethality when intercrossed. 1 The allozyme evidence suggests that divergence between the gene pools occurred before domestication. Subsequent recombination has been restricted not by geographical separation but by post-zygotic incompatibility mechanisms, as shown in Malawian lines which lack significant numbers of gene pool recombinants in spite of the inter- planting of both types over several centuries. Although selection by local growers has maintained desirable ideotypes, it has not produced recombinations of the gene pool-specific traits of seed size and growth habit. Isozyme data from $68 of Malawian material indicate that a previously known syndrome producing male sterility is a result of between-pool hybridization. This and other genetic incompatibilities suggest that incipient speciation is occurring between the gene pools. Two tightly linked loci of a tetrameric form of NADH diaphorase were described using SGE, Diap-1 and Diap-Z. Five alleles producing interacting subunits and two null activity alleles were identified. Specific diaphorase alleles occur in the tropical black and tropical red bean types from Central America, both identified as distinct entities within 2; vulgaris from previous work. Another allele was found only in certain cultivars collected in Turkey. Only three out of seven diaphorase alleles were found among Malawian lines, indicating that the entire range of domesticated bean germplasm is not present in that country. ACKNOWLEDGEMENTS I take this opportunity to gratefully acknowledge the many people of Michigan State University who have made it possible for me to pursue and complete my doctoral research. Several of these people are part of the Malawi Project of the Bean/Cowpea Collaborative Research Support Program and the Plant Breeding and Genetics Group, which generously financed my graduate studies. My major professor, Dr. M. Wayne Adams, has supported me unstintingly over five years, with his confidence and with his uncommonly broad knowledge of the common bean. His insight into the possibilities for future applications of my work in the study of plant/pathogen coevolution has been especially encouraging. I have enjoyed having Dr. Anne E. Ferguson as a travelling companion and as a fellow-thinker and writer. Of all the students and faculty whom I have met at Michigan State University, those involved in bean research are the ones I will remember most. Bean researchers are a special breed! Information from the national bean programs of other countries, a wealth of experience in agronomy and breeding, and insight into the conditions of agricultural research in many parts of the world, are the professional gifts the Bean Group students have shared with me. They have provided friendship and support in the Joys and agonies of graduate student life, and broadened my outlook. I hope their careers will be fulfilling and satisfying, and that we will meet again in the years ahead. 11 TABLE OF CONTENTS Page LIST OF TABLESIIOOOOOOOOOOO00....0.00.00.00.00.0..OOOOOOOOOOOOOOOOOOOOVi LIST OF FIGURESOOOOOOOOOOOOOOIOO...00.0.0000...OOOOOOOOOOOOOOOOOOOOIOOix GENERAL INTRODUCTION Isozyme Analysis as a Genetic Tool................................1 Isozyme Analysis in Common Bean...................................2 Isozymes Analyzed in the Current Work.............................6 ISOZYMES AND PROTOCOLS...............................................7 LIST OF REFERENCES..................................................13 CHAPTER ONE: ALLOZYME DIFFERENCES BETWEEN TWO MAJOR GENE POOLS IN Phaseolus vulgaris L. INTRODUCTION........................................................16 MATERIALS AND METRODS...............................................2fl Electrophoretic Methods..........................................2u Plant Materials..................................................26 RESULTS.............................................................30 Isozyme Data.....................................................30 Population Analyses..............................................32 Male Sterility...................................................u6 DISCUSSION..........................................................51 Gene Pool Differentiation and the Sequence of Divergence.........52 Reproductive Barriers............................................53 Intra- and Inter-specific Relationships..........................S9 LIST OF REFERENCESOOOOOOOOOOOOOOOOOOOOOOOOO0.0..0.0.00.000000000000065 iii iv CHAPTER TWO: ISOZYME VARIANTS AT TWO BETA-NADH DIAPHORASE LOCI IN DRY BEANS (Phaseolus vulgaris L.): CORRELATIONS TO GENE POOLS INTRODUCTION........................................................69 The Diaphorase Enzymes...........................................69 DIAP Isozyme Analysis............................................7u Gene Pools in Common Bean........................................77 MATERIALS AND METHODS...............................................80 Plant Materials..................................................80 Electrophoretic Methods..........................................83 RESULTS.............................................................86 Enzyme Activity..................................................86 The Zymograms Seen in DIAP.......................................86 Phaseolin Electromorphs..........................................87 DIAP Patterns in Various Populations.............................90 Populations Segregating for DIAP................................107 Test of Linkage between DIAP and Architecture Genes.............110 DISCUSSION.........................................................115 The DIAP System in Common Bean..................................115 Evidence fer Linkage between DIAP Loci..........................121 Gene Duplication................................................12u Gene Pool Specific DIAP Alleles.................................126 Divergence of DIAP Variants.....................................130 Future Investigations...........................................131 LIST OF REFERENCES.................................................136 CHAPTER THREE: THE MAINTENANCE OF SEED CLASS IDEOTYPES IN VARIETAL MIXTURES IN TRADITIONAL AGRICULTURE IN MALAWI INTRODUCTION.......................................................1u1 MATERIALS ANDMETHODSOOOOOOCOO...0...0...OOOOOOOOOOOOOOOOOOOOOOO0.01149 RESULTS............................................................1S1 Variation among Seed Classes....................................151 Variation within Seed Class.....................................16N Information from Sheila Mkandawire..............................177 DISCUSSION.........................................................180 Comparisons between Northern and Central Malawi Samples.........180 The Maintenance of Ideotypes....................................183 Perceptual Distinctiveness......................................185 LIST OF REFRENCESOCOOOOOOO0.00.00.00.00000000000000000000000000.00.192 APPENDICES APPENDIX A: Data for 375 Bean Lines Sampled from 15 Farms in northern MalawiIOOOOOIIOOOOOOOOIOOOIOOOOOOOOOOOOOOO0.00.195 APPENDIX B: Information on Malawian Beans from an Interview with Sheila mandaWireOOOOOOOOOO...OOOOIOOOOOOI...0.0...00.0.20” 10. 11. 12. 13. 1n. LIST OF TABLES Page CHAPTER ONE Tissues assayed for enzyme activity following SGE................25 Phaseolin and Dwarf Lethal genotypes assayed for isozymes........29 Mobility variants observed at eight isozyme loci.................31 Isozyme genotypes in 373 lines from 15 Malawian farms...........3u Genotype frequencies at six loci in 373 Malawian lines...........35 Allele frequencies at seven loci in 373 Malawian lines...........35 Allele combinations observed in 373 Malawian lines...............36 Seed size traits of allele combinations #1 and #7................38 Differences between 266 Andean and 51 Mesoamerican lines.........u2 Allozymes observed in phaseolin and Dwarf Lethal genotypes.......u3 Genotypes of commercial classes from the two gene pools..........UA Gene pool differences among 116 black seeded lines...............fl7 Allozymes associated with the two major gene pools in bean.......u7 Frequency of predominant allele in samples from gene pools.......u8 Isozyme genotypes of 13 male sterile Malawian lines..............50 CHAPTER TWO Enzymes reported in genetic studies of diaphorase isozyme........71 Seed size differences among Malawian DIAP genotypes..............91 Association between DIAP zymotype and phaseolin type.............91 DIAP pattern in accessions in relation to tropical blacks........9u DIAP pattern and seed size in 126 black seeded accessions........97 vi 10. 11. 12. 13. 14. 15. 16. 10. 11. 12. 13. vii I Association between seed size, DIAP pattern and phaseolin.......101 Red mexican and pink cultivars analyzed for DIAP................10A DIAP genotype of lines from Mesoamerican and Andean pools.......105 DIAP pattern and allele frequencies for gene pool lines.........106 Characters of lines carrying the Diang allele..................106 Segregation ratios for DIAP alleles at two loci.................108 Yield of DIAP isolines and commercial 'C-20'....................112 DIAP type of lines intermated for recurrent selection...........112 DIAP in five selection cycles of a pinto x architype cross......11u Expected isozymes from enzymes of differing subunits............117 Isozymes and ratios expected in segregating tetramers...........123 Known combinations or alleles at DiaE-1 and Diae-Zeooooeoooooooo125 CHAPTER THREE Seed classes among 375 lines collected in northern Malawi.......153 No. and type of seed classes in each 25 seed farm sample........15u Allozyme combinations occurring in each seed class..............155 32 Malawi seed classes sorted by frequency and gene pool........156 Number of lines belonging to each gene pool on 15 farms.........158 Total heterozygosity associated with isozyme heterozygosity.....160 Occurrence of growth habit types among isozyme combinations.....162 Growth habit among gene pools in the Malawian sample............162 Growth habit in 32 seed classes in the Malawi sample............163 Distribution of different growth habit lines on 15 farms.......165 Lines with minimum and maximum distance in 15 farm samples......166 Growth habit differences among three seed classes, by farm......168 Allozyme combinations found in seed class #3....................170 viii 1N. Isozyme-by-growth habit groups in seed class #3.................17O 15. General knowledge about seed classes contributed by Sheila mandawireIOOOO..0...IOOOOIOOOOOOOOOOOOI0.00....0.0.178 16. Planting and seed selection practices from S. Mkandawire........179 LIST OF FIGURES Page CHAPTER ONE 1. Seed size frequency histogram. 373 Malawian landrace lines........39 CHAPTER TWO 1. R values of DIAP isozymes, with corresponding alleles...........88 f 2. Pedigree relationships involving tropical blackgermplasm.........93 CHAPTER THREE 1. Four growth habit and isozyme groupings of seed class #3........172 1a. Determinate growth habit, isozyme combination #1................173 1b. Indeterminate growth habit, isozyme combination #1..............17u 1c. Indeterminate growth habit, isozyme combination #2..............175 1d. Indeterminate growth habit, isozyme combinations #A.............176 ix GENERAL INTRODUCTION Isozyme Analysis as a Genetic Tool Analysis of isozyme variation has been of major importance in plant genetic characterization during the last 25 years (Shaw 1965, Scandalios 1968, Brewer 1970, Tanksley and Orton 1983). The electorphoretic identification of enzyme alleles has simplified genetic mapping by providing codominant, morphologically neutral markers for identification of qualitative and quantitative trait loci and polymorphic restriction fragment lengths (Goodman et a1. 1980, Tanksley and Bernatzky 1987). It has allowed detailed studies of the genetic structures of wild populations and landraces (Brown 1978, Gottlieb 1981, Brown and Munday 1983), and made precise genotype differentiation possible (Seller and Beckmann 1983). Taxonomic and developmental studies have been facilitated (Scandalios 197A, Decker 1985). Variant functional forms of the same enzyme, or isozymes, are produced as a result of changes in DNA sequences, and the differences may occur between or within taxa. Isozymes can be produced by different loci encoding the same enzyme in more than one form, such as genetically independent proteins, or heteromers of two or more polypeptides, non- covalently bound. Isozymes also result from allelic variants at a single locus, and these are termed allozymes. True isoenzymes are produced from specific different DNA sequences. Secondary isoenzymes are the result of modification of a single gene product after protein synthesis, eg., proteins conjugated with other groups or derived from one polypeptide chain; polymers of a single sub-unit; or conformation- ally different forms (Dixon and Webb 1979). Various sub-compartments of the cell may have specific associated isozymes produced from different loci in the nuclear genome. Such isozymes may differ in amino acid sequence, net charge, acidity, and catalytic activity appropriate to the reactions in which they function. In plants, isozymes of the same enzyme may be compartmentalized in the cytosol, the mitochondria, the chloroplast, and glyoxysomes (Newton 1983). In some instances, the mitochondrial enzyme form may be the result of post-translation modification of the original cytosolic sequence, both coming from a single gene (Dixon and webb 1979). The number of electromorphs found at a single enzyme locus depends on the electrochemical properties of the buffer system used. For human haemoglobin protein, 150 variants have been found, and enzyme proteins probably have as much variability (Dixon and Webb 1979). In plants, some highly polyallelic loci are known, such as a peroxidase locus in tomato with over 59 alleles (Rick 1983). The use of a starch gel as an electorphoretic matrix, in which isozymes may be separated on the basis of their net electrical charge, has provided an inexpensive tool for rapidly screening the large numbers of genotypes necessary in many types of genetic studies. Isozyme Analysis in Common Bean Genetic studies of variants in enzymes in the genus Phaseolus began to appear a few years after gel methodologies were introduced. West and Garber (1967a and b) surveyed esterase (EST) and leucine aminopeptidase (LAP) isozyme patterns in 15 species of Phaseolus. To their knowledge, this was the first study of the inheritance of electrophoretically haw .m- D pzp Ian... 'I I tbs . o . o “In. a one. :v; .‘N. q.. I ‘6'“ ”.Ip... Viv. :4 in a..‘= ‘5 n.- ' w v (I: l‘ .‘. R n!- variant enzymes in plant interspecific hybrids and their F progeny. They found that intraspecies zymogram patterns resembled each other more than interspecific patterns, and by comparing both enzymes these authors were able to distinguish each of the 15 species (1967a). Among N9 cultivars and lines within 2; vulgaris three differing anodal EST patterns were seen, and only one cathodal pattern. For LAP a single anodal pattern was produced. Interspecific crosses were produced between P; vulgaris and P; coccineus; and F , F and backcrosses were 2 3 analysed. Amphidiploids containing P; aureus, P; mungo, or P; trilobus were also analysed. By using this interspecific polymorphism the authors were able to show that inheritance of LAP in vulgaris x coccineus was determined by codominant alleles at one locus. At the single EST locus common to both species, it appeared that the vulgaris allele showed dominance, and it was noted that dominance among allozymes was rarely found in previous studies. The authors pointed out that isozyme analysis can be used to investigate introgressive hybridization more easily than the cumbersome methods of measuring herbarium sRegimens, or producing hybrid indices based on diagnostic characters (1967b). Wall (1968) extended the work on vulgaris x coccineus hybridization "ital F , F , F and backcross genertions, and showed that there was 1 2 3 differential elimination of gametes carrying the donor parent LAP allele When the F hybrid was used as the male parent. This author noted that electrophoresis can be used to measure gene flow between species. Wall and Wall (1975) applied starch gel electrophoresis (SGE) to evolutionary questions in the P; vulgaris - _P__._ coccineus complex in Me"ii-<30, and showed evidence from EST and alcohol dehydrogenase (ADH) patterns for introgression of wild P; vulgaris germplasm into cultivated coccineus via cultivated vulgaris. Their data also indicated unexpectedly high levels of outcrossing in vulgaris, which they attributed to favorable conditions present in the area of origin for this species. Bassiri and Adams (1978a) assayed EST, acid phosphatase (ACP) and peroxidase (PRX) in primary leaves, roots and stems of 15 day old seedlings and found unique banding patterns for each of the 13 species of Phaseolus analysed. No inheritance study was done at the time, so that the number of genetic marker loci represented by the numerous bands which were resolved was not determined. The authors discussed the tissue specificity of isozymes seen in this study. While they did not find EST isozymes in root tissue, other studies have found EST activity in bean roots (eg. Weeden, 198A). They drew attention to the noticable similarity between zymograms of cultivated and wild vulgaris and wild coccineus, such that the domesticated vulgaris appeared to incorporate bands of both wild species. They concluded, in a variation on Wall and Wall's (1975) theory, that wild coccineus contributed to the evolution of the domesticated vulgaris. A second paper by Bassiri and Adams (1978b) used the same enzymes to distinguish between 3” vulgaris cultivars belonging to 19 commercial classes grown in the United States. They found that no class was defined by only one isozyme pattern, although the homology of banding patterns for cultivars in the same class was often high. An interesting case is that of the tropical black group of which the four cultivars were identical in leaf, stem and root isozymes, with the exception of leaf EST. By grouping cultivars by number of polymorphic isozyme bands in «v- ' v 15- .ib‘ AI.“ 0... I "I: fI.r .a I:‘. "v' og.‘ ‘ ¢:;‘ . q“ . Pu common, relationships known from past breeding history were evident. Marques Comes et al. 198A found variation in isozymes of EST, LAP, and soluble total seed protein, but not for PRX and ACP, in three cultivars and four breeding lines of snap beans. Weeden (198A, 1986) identified allelic isozyme variants in bean at loci coding for adenylate kinase, shikimic acid dehydrogenase, malic enzyme, for the small subunit of ribulose bisphosphate carboxylase, PRX, glucose phosphate isomerse, and N-acetyl glucoseaminidase, and noted that polymorphism existed for xanthine dehydrogenase, succinate dehydrogenase, malate dehydrogenase, isocitrate dehydrogenase, and amylase (1983). Isozyme analysis has been applied to beans for descriptive purposes, to differentiate between morphologically similar cultivars, and to identify developmental changes in gene expression. Adriaanse et al. (1969) used SGE to study the properties of proteins in beans after processing, because of the resolving power of this type of gel. Although initially pursuing a non-genetic study, these authors were able to show that cultivars could be characterized by densitograms of banding patterns produced by electrophoresis of water soluble seed proteins. Thirty-three European cultivars, from three commercial classes were used: string beans, bacon beans, and green beans. Banding patterns of cultivars of each class showed common features, yet closely related cultivars were distinguishable. It was concluded that growth environment did not affect banding mobility, and that the protein patterns were controlled genetically. Weeden (198A) used isozyme variants for beta-NADH diaphorase, Peptidase, and ACP in addition to the six enzyme loci mentioned above, a... .9, fish .Qeq One I". SP.» ~-..I a I .5.“ I‘I‘ (1 I:. 0-9 Iv to distinguish among white seeded snap bean cultivars. Bassiri and Carlson (1978) identified isozymes of PRX, malate dehydrogenase, and ACP in a single navy bean cultivar, and found several to be organ specific. They showed that isozyme expression in calli derived from different tissues was similar, and differed from the tissue of origin. Isozymes Analyzed in the Current Work Several isozymes were analysed in preliminary investigations, but not all were feund to be equally useful in the analysis of Malawian landrace lines and populations from the two major gene pools of common bean. Several did not give adequate development of bands (adenylate kinase); others gave good results only intermittently (EST). Certain enzymes, including isocitrate dehydrogenase and glucose phosphate isomerase, which have been reported to be polymorphic in bean (Weeden 1983), appeared to be monomorphic in the material examined. Such lack of variation in mobility could be a result of the buffer system used. The enzymes attempted are listed below. Scientific names and Enzyme Commission number are drawn from Dixon and Webb (1979). and Vallejos (1983); references are given for the activity stain protocols used. ISOZYMES AND PROTOCOLS ADENYLATE KINASE: ADK ATP:AMP phosphotransferase EC 2.7.A.3 Inconclusive results; not used in analysis. Stain: Weeden 198A ACID PHOSPHATASE: ACP Orthophosphoric-monoester phosphohydrolase (acid optimum) EC 3.1.3.2 The general non-specific phosphohydrolases function in the hydrolysis of phosphomonoesters such as in the formation of sucrose in photosynthesis (Goodman and Stuber 1983). Found in monomeric and dimeric terms in plants. One polymorphic locus in beans. Stain: Weeden 198A ALCOHOL DEHYDROGENASE: ADH Alcohol:NAD oxidoreductase, EC 1.1.1.1 This enzyme is a dimer in a variety of angiosperms, but can appear as a tetramer in yeast (Dixon and Webb 1979). In wheat Hart (1983) distinguishes between isozymes produced by ADH's with different co- enzyme (NAD or NADP) and substrate (aromatic or aliphatic alcohols) specificities. In the current work no polymorphism was seen. Stain: Rick et al. 1977 ASPARTATE AMINOTRANSFERASE: AAT (GLUTAMATE OXALOACETATE TRANSAMINASE: GOT) L-Aspartate:2-oxoglutarate aminotransferase EC 2.6.1.1 These are reported as dimers in several crop species with the exception of pepper (McLeod et a1. 1983). This enzyme has a role in the transamination which eliminates N from amino acids, forming the keto acids fer the Krebs cycle and fbr gluconeogenesis (Goodman and Stuber 1983). No polymorphism was seen in the current work. Stain: Weeden and Emma n.d. :._ .. aid» 3 ..' 9 '5. U 5"- CATALASE: CAT Hydrogen-peroxide:hydrogen-peroxide oxidoreductase, EC 1.11.1.6 A tetramer in maize, where cytosolic and mitochondrial forms are found (Goodman and Stuber 1983.) The stain used (Conkle et al. 1982) is a transient negative stain. Rubisco isozymes also appear to develop on gels stained fer CAT. No polymorphism seen in the current work. Stain: Conkle et a1. 1982 DIAPHORASE: DIA A general name for several enzymes which catalyze the oxidation of either beta-NADH or beta-NADPH in the presence of an electron acceptor. Occurs as tetramers and monomers in soybeans where it is reported as EC 1.6.”.3 (Kiang and German 1983). monomers of EC 1.6.99.1 and dimers of EC 1.6.N.3 reported in spinach leaves (Dixon and Webb 1979), monomers of 1.6.99.3 reported in barley (Brown 1983). Two polymorphic loci, producing subunits which interacted to form tetramers, were described in the current work. Stain: Weeden 198A, Harris and Hopkinson 1976. ESTERASE: EST These enzymes have broad specificity in plants, and the carboxylic ester hydolases (EC 3.1.1.-) stained may depend on the ingredients in the staining solution. EST isozymes are numerous in plants, with maize having at least 10 loci with 2 to 9 alleles each (Goodman and Stuber 1983). They occur as monomer and dimers in several crops. Inconsistent development caused this enzyme to be dropped in this study, although on good gels polymorphisms were evident. Polyacrylamide gels appear to give clear, consistent results in beans (Tohme, pers. comm.). Stain: Rick et a1. 1977: Weeden 198A GLUCOSE-6-PHOSPHATE DEHYDROGENASE G-6-PDH D-Glucose-6-phosphate:NADP+ 1-oxidoreductase (Vallejos 1983) 3alpha-Hydroxysteroid:NAD(P)+ oxidoreductase (Dixon and Webb 1979) EC 1.1.1.H9 A dimer in soybeans (Kiang and Gorman 1983). No polymorphisms seen in the current work. Stain: Weeden and Emmo n.d. GLUCOSE PHOSPHATE ISOMERASE: GPI (Phosphoglucoisomerase) D-Glucose-6-phosphate ketol-isomerase EC 5.3.1.9 This glycolytic enzyme is reported by Freeling (1983) to be u w.. Sg': . .q .I n... 0-,. \ 6". ru- H Anaerobic Protein #55 (ANPSS), one of the few enzymes induced during anaerobic stress in maize. It usually occurs as a dimer in crops, but is known as a monomer in petunia (Wijsman 1983), where prolonged staining of the gel causes G-6-PDH activity to appear. No polymorphisms seen in the current work. Stain: Weeden and Emmo n.d. GLUTAMATE DEHYDROGENASE: GDH L-Glutamate:NAD+ oxidoreductase (deaminating) EC 1.9.1.2 The number of subunits is not listed for this enzyme in Dixon and Webb (1979) but the related Glutamate dehydrogenase (NAD(P)+) is reported to be a hexamer. GDH EC 1.A.1.2 is described as a hexameric enzyme in rice (Endo and Morishma 1983) and alfalfa (Quiros 1983); Goodman and Stuber describe it as polymeric in maize (1983). No polymorphism was seen in the current work. Stain: Weeden and Emma n.d. GLUTAMATE OXALOACETATE TRANSAMINASE: GOT see Aspartate aminotransferase: AAT ISOCITRATE DEHYDROGENASE (NADP+): IDH three-Ds-Isocitrate:NADP+ oxidoreductase (decarboxylating) EC 1.1.1.A2 This NADP-specific IDH is usually found at one locus in plants but is known to have two loci in maize (Goodman and Stuber 1983) and soybeans (Kiang and German 1983). It is known as a dimer in maize and as dimers and monomers in barley (Brown 1983). No polymorphism was seen in the current work. Stain: Weeden and Emmo, n.d. LEUCINE AMINOPEPTIDASE: LAP Recommended name: Aminopeptidase(cytosol) Systematic name: alpha-Aminoacyl-peptide hydrolase(cytosol) EC 3.4.11.1 Monomers in crops. No polymorphism seen in the current work. Stain: Weeden and Emmo n.d. MALATE DEHYDROGENASE: MDH L-Malate:NAD+ oxidoreductase EC 1.1.1.37 This NAD-specific malate dehydrogenase has been well-characterized in maize (Goodman and Stuber 1983) where it is a dimeric enzyme. In . . . A. u an :- .I- dl I - . ‘ I :- .|.I ‘ 1 4-. . ‘1 II! I. . 1 . w .I- In. .3 . u in In R. A v a '4 a (v on" . I4 5 n o u 2. us -. - u . .c It. .4“ New 7.5 e u 1. e u . k - P .l A: "lid u'. ‘b I An aw 6 0 v D I . e n v A. . a . P. . m c a e . Him .l. n\uv CI 1 1 OIRU . . an E m. K \ nln ml. nun all 5’. «- ~\~ an. an. ..' Ir of ab. Ha- -\I 1O barley (Brown 1983) it occurs as monomers and dimers; in celery it is dimeric (Orton 1983); in soybean it is thought to be a dimer (Kiang and German 1983). Ne polymorphism was seen in the current work. Stain: Weeden and Emmo n.d. MALIC ENZYME: ME Recommended name: Malate dehydrogenase(oxaloacetate-decarboxylating) (NADP+) Systematic name: L-Malate:NADP+oxidoreductase(oxaloacetate- decarboxylating) EC 1.1.1.A0 This is the NADP specific malate dehydrogenase. Goodman and Stuber (1983) report that in maize it may occur as a tetramer, and that it does go through changes in expression during development. One polymorphic locus was used in the current work. Stain: Weeden and Emmo n.d. N-ACETYL GLUCOSEAMINIDASE: NAG A fluorescent end-product is produced from activity on N- methylumbelliferyl-N-acetyl-beta-D-glucosaminide, and the bands are read under an ultraviolet light source. Polymorphism at one locus is known in beans (Weeden 1986), and was seen in the current work. Some preliminary results suggest that the locus is linked to a gene producing phaseolin variability. Weeden 1986. PEROXIDASE: PRX Donor:hydrogen peroxide oxidoreductase EC 1.11.1.7 A monomer in plants (Dixon and Webb 1979). One cathodal polymorhpic locus known in bean (Weeden 198A) was analyzed in the current work. Stain: Weeden 198A PHASEOLIN/SEED PROTEIN: SDPR This is the major seed storage protein in beans. Although it lacks enzyme activity its forms can be visualized by staining seed extracts with a general protein stain. (The same stain is used to visualize the major protein in leaves, rubisco; see below in this list.) The major band seen in SGE is assumed to be phaseolin, the most abundant seed protein. In beans two mobiity variants are seen on starch gels. These can be correlated to the differences between the major anodal band seen H I 'w' er1 n.- I \ .uvy 3'" ‘OVM qua fawn .‘H 09" I , \O-q ’4‘ ‘05. cu‘. . v ‘. ‘ o... 11 on SDS/PAGE, which differs between genotypes from the two major gene pools in beans (Gepts and Bliss 1986). Stain: general protein stain, Weeden 198A PHOSPHOGLUCOMUTASE: PGM alpha-D-Glucose-1,6—bisphosphate:alpha-D-glucose-1-phosphate phosphotransferase EC 2.7.5.1 This glycolytic enzyme is found as monomers in several crop isozymes. No polymorphism seen in the current work. Stain: Weeden and Emmo n.d. 6-PHOSPHOGLUCONATE DEHYDROGENASE (DECARBOXYLATING): 6-PGD 6-Phespho-D-gluconate:NADP+ 2-oxidoreductase (decarboxylating) EC 1.1.1.3” In maize this is listed as EC 1.1.1.AA (Goodman and Stuber 1983), while in barley ( Brown 1983) and eat (Price and Kahler 1983) it is EC 1.1.1.A3 (6-Phospho-D-gluconate:NAD(P)+ 2-oxidoreductase), probably indicating a difference in developing protocol. It is a dimer in these plants. No polymorphism seen in the current work. Stain: Weeden and Emma n.d. RUBISCO: RBCO Recommended name: Ribulosebisphosphate carboxylase Systematic name: 3-Phospho-D-glycerate carboxyl-lyase (dimerizing) EC A.1.1.39 This is the most common protein in green leaf tissue, often accounting fer 50$ of protein present, and its isozymes are readily developed by a general protein stain. RBCO is the major band which appears. The variation seen in beans has been shown to result from alleles at a single nuclear locus encoding the small subunit (Weeden 198”). This polymorphism was analyzed in the current work. Stain: Weeden 198A SHIKIMATE DEHYDROGENASE: SKDH Shikimate:NADP+ 3-oxidoreductase EC 1.1.1.25 In barley (Brown 1983), tomato (Rick 1983) and petunia (Wijsman 1983) it is described as a monomer; in celery Orton (1983) notes that the pattern found in homozygotes of two co-migrating bands could be explained on the basis of a dimeric or monomeric enzyme. It is possible that there is some modification of the translated peptide; the intermediate band could be invariable and produced by a locus other than the one producing the most anodal and the most cathodal allozymes; and 12 this latter explanation would hold if the protein were dimeric as well. This double banded pattern occurs in beans, and makes heterozygotes difficult to verify. The polymorphism at a single locus (Weeden 198A) was analyzed in the current work. Stain: Weeden 198A ‘22- I n 0‘.) I ‘n I! i LIST OF REFERENCES Adriaanse, A., Klop, W. and Robbers, J. J. E. 1969. Characterisation of Phaseolus vulgaris cultivars by their electrophoretic patterns. J. Bassiri, A. and Adams, M. W. 1978a. An electrophoretic survey of seedling isozymes in several Phaseolus species. Euphytica 27:uu7-u59. Bassiri, A. and Adams, M. W. 1978b. Evaluation of common bean cultivar relationships by means of isozyme electrophoretic patterns. Euphytica 27:707-720. Bassiri, A. and Carlson, P. S. 1978. Isozyme patterns and differences in plant parts and their callus cultures. Crop Science 18:955-958. Brewer, G. J. 1970. An Introduction to Isoenzyme Technique. Academic Press, New York. Brown, A. H. D. 1983. Barley. pp 55-78 in Isozymes in Plant Genetics and Breeding, Part B. edited by S.D. Tanksley and T. J. Orton. Elsevier Science Publishers B. V., Amsterdam. Brown, A. H. D. and Munday, J. 1982. Population-genetic structure and optimal sampling of land races of barley from Iran. Genetica 58:85-96. Conkle, M. T., Hodgekiss, P. D., Nunnally, L. B. and Hunter, S. C. 1982. Starch Gel Electrophoresis of Conifer Seeds: a Laboratory Manual. USDA. Pacific Southwest Forest and Range Experiment Station, Berkeley. Decker, D. S. 1985. Numerical analysis of allozyme varition in Cucurbita pepo. Economic Botany 39:300-309. Dixon, M. and Webb, E. C. 1979. Enzymes. Academic Press, New York. Endo, T. and Morishma, H. Rice. pp 129-1A6 in Isozymes in Plant Genetics and Breeding, Part B. edited by S. D. Tanksley and T. J. Orton. Elsevier Science Publishers B. V., Amsterdam. Gepts, P. and Bliss, F. A. 1986. Phaseolin variability among wild and cultivated common beans (Phaseolus vulgaris) from Colombia. Economic Botany A0:A69-fl78. Goodman, M. M., Stuber, C. W., Newton, K. and Weissinger, H. H. 1980. Linkage relationships of 19 enzyme loci in maize. Genetics 96:697-710. 13 _‘O. u l‘ in“ (vi 3‘»- ..ul 1 " 1 ‘ I 'A J‘. I ’4 C II - :- .inl i . u .- . , l 'u’: . I u ..,‘h I‘m 1N Goodman, M. M. and Stuber, C. W. 1983. Maize. pp 1-3A in Isozyme in Plant Genetics and Breeding, Part B. edited by S. D. Tanksley and T. J. Orton. Elsevier Science Publishers B. V., Amsterdam. Gottlieb, L. D. 1981. Electrophoretic evidence and plant populations. Progress in Phytochemistry 7:1-N6. Hart, G. E. 1983. Hexaploid wheat (Triticum aestivum L. em Thell). in Isozymes in Plant Genetics and Breeding, Part B. edited by S.D. Tanksley and T. J. Orton. Elsevier Science Publishers B. V., Amsterdam. Harris, H. and Hopkinson, D. A. 1976. Handbook of enzyme electrOphoresis in human genetics. Berth-Holland, Amsterdam. Kiang, I. T. and German, M. B. 1983. Soybean. pp 295-328 in Isozymes in Plant Genetics and Breeding, Part B. edited by S. D. Tanksley and T. J. Orton. Elsevier Science Publishers B. V., Amsterdam. Marques Comes, M., Leal, N. R. and Cordeiro, A. R. 1984. Padroes eletrofereticos em progenitores elingagens de feijao-de-vagem (Phaseolus vulgaris L.). Revista Ceres 31:231-237. McLeod, J. F., Guttman, S. I. and Eshbaugh, W. H. 1983. Peppers. pp 189-202 in Isozymes in Plant Genetics and Breeding, Part B. edited by S.D. Tanksley and T. J. Orton. Elsevier Science Publishers B. V., Amsterdam. Newton, K. J. 1983. Genetics of mitochondrial isozymes. in Isozyme in Plant Genetics and Breeding, Part A. edited by S. D. Tanksley and T. J. Orton. Elsevier Science Publishers B. V., Amsterdam. Orton, T. J. 1983. Celery and celeriac. pp 351-368 in Isozymes in Plant Genetics and Breeding, Part B. edited by S.D. Tanksley and T. J. Orton. Elsevier Science Publishers B. V., Amsterdam. Price, S. and Kahler, A. L. Oats. 1983. pp 103-128 in Isozymes in Plant Genetics and Breeding, Part B. edited by S. D. Tanksley and T. J. Orton. Elsevier Science Publishers B. V., Amsterdam. Quiros, C. F. 1983. Alfalfa, luzerne. pp 253-29N in Isozymes in Plant Genetics and Breeding, Part B. edited by S. D. Tanksley and T. J. Orton. Elsevier Science Publishers B. V., Amsterdam. Rick, C. M. 1983. Tomato. pp 1u7-166 in Isozyme in Plant Genetics and Breeding, Part B. edited by S. D. Tanksley and T. J. Orton. Elsevier Science Publishers B. V., Amsterdam. Rick, C. M., Fobes, J. F., and Belle, M. 1977. Genetic variation in Lycopersicon pimpinellifolium: evidence of evolutionary change in mating systems. Plant. Syst. Evo. 127:139-170. Scandalios, J. G. 197“. Isozymes in development and differentiation. Ann. Rev. Plant Physiology 25:225-258. 15 Shaw, C. R. 1965. Electrophoretic variation in enzymes. Science Seller, M. and Beckmann, J. S. 1983. Genetic polymorphism in varietal identification and genetic improvement. Theor. Appl. Genet. 67:25-33. Tanksley, S. D. and Bernatzky, R. 1987. Molecular markers for the nuclear genome of tomato. pp 37-NA in Nevins, D. J. and Jones, R. A. (eds.) Tomato Biotechnology. Alan R. Liss, New York. Tanksley, S. D. and Orton, T. J. 1983. Isozymes in Plant Genetics and Breeding, Part A and B. Elsevier Science Publishers B. V., Amsterdam. Vallejos, C. E. 1983. Enzyme activity staining. pp N69-516 in Isozymes in Plant Genetics and Breeding, Part A. edited by S. D. Tanksley and T. J. Orton. Elsevier Science Publishers B. V., Amsterdam. Wall, J. R. 1968. Leucine aminopepetidase polymorphism in Phaseolus and differential elimination of the donor parent genotype in inter- specific backcrosses. Biochem. Genet. 2:109-118. Wall, J. R. and Wall, S. W. 1975. Isozyme polymorphisms in the study of evolution in the Phaseolus vulgaris x P; coccineus complex of Mexico. pp 287-305 in C. L. Markert (ed.) Isozymes vol. A, Academic Press, New York. Weeden, N. F. 1983. Variation at isozyme loci in Phaseolus vulgaris. Ann Rept. Bean Improv. Coop. 26:102-105. Weeden, N. F. 198A. Distinguishing among white seeded bean cultivars by means of allozyme genotypes. Euphytica 33:199-208. Weeden, N. F. 1986. Genetic confirmation that the variation in the zymograms of 3 enzyme systems is produced by allelic polymorphism. Ann. Rept. Bean Improv. Coop. 29:117-118. Weeden, N. F. and Emmo, A. C. n.d. Horizontal Starch Gel Electrophoresis Laboratory Procedures. NYAES. Geneva, New York. West, N. B. and Garber, E. D. 1967a. Genetic studies of variant enzymes. 1. An electrophoretic survey of esterase and leucine aminopepetidases in the genus Phaseolus. Can J. Genet. Cytol. 9:6A0- 6A5. West, N. B. and Barber, E. D. Genetic studies of variant enzymes. II. The inheritance of esterases and leucine aminopepetidases in Phaseolus vulgaris 5’2; coccineus. Can. J. Genet. Cytol. 9:6A6-655. Wijsman, H. F. W. 1983. Petunia. pp 229-252 in Isozymes in Plant Genetics and Breeding, Part B. edited by S. D. Tanksley and T. J. Orton. Elsevier Science Publishers B. V., Amsterdam. CHAPTER ONE ALLOZYME DIFFERENCES BETWEEN TWO MAJOR GENE POOLS IN Phaseolus vulgaris L. INTRODUCTION The common bean Phaseolus vulgaris L. has been proposed as an ideal model for study of the evolution of crop plants. Its wild progenitor populations remain in existence, and a range of primitive and improved cultivars is present in agriculture (Evans and Walters 1979). Recent studies have produced evidence indicating that cultivated types belong to two separate sub-specific taxa, or gene pools (Evans 1975, Gepts and Bliss 1985, Gepts et al. 1986), and the ramifications of this fact increase the interest of common bean as a model. The presence of evolutionarily differentiated groups within a crop allows detailed study of the extent and variety of reproductive isolating mechanisms operating at the subspecies level, of the coadaptation between the sub-group and its pathogens or symbionts, and of gene flow. This chapter contributes data from isozyme analysis to show that cultivated material from the two major gene pools in common bean is marked by alternate alleles at several loci coding for enzymes, and that gene flow between the pools is restricted by post-zygotic isolating mechanisms. These results support the hypothesis that divergence has occurred during the evolution of P; vulgaris, and suggest that hybridization between groups within the species has the potential to provide the type of novel variation usually found in wider crosses. The wild types of the species 2. vulgaris have a broad natural 16 17 distribution in the New World, appearing in the Mexican plateau and the lowland tropics of Mesoamerica, and in the eastern Andean region of South America from Venezuela to Argentina (Kaplan 1981). They differ in morphological and biochemical traits (Vanderborght 1983, Gepts et a1.‘ 1986, Gepts and Bliss 1986). The most recent revision of the systematics of Phaseolus (Delgado-Salinas 1985) describes the wild taxa as botanical varieties, those from the Andean region being 2; vulgaris var. aborigineus Burkart, and the Mexican being var. mexicanus A. Delgado, var. nov. They are distinguished from the domesticated bean. No wild beans have been identified in the archeological record, but domesticated types have been dated circa 7,000 B.P. in both Mesoamerica and the Andean region, at a time before cultural exchange took place between these areas (Kaplan 1981). Present day evidence for independent domestication in both regions has been summarized by Evans (1980). Gentry (1969) suggested that in Mexico the different wild races had been incorporated into agriculture in numerous events, on the basis of the correspondence between local wild seed types and the cultivated varieties grown in each area. Berglund-Brucher and Brucher (1976) found evidence fer domestication of horticultural and kidney bean types from the wild Andean P; aborigineus, which is still gathered for consumption ' in some areas. The case fer independent domestication events, and sub-species divergence prior to cultivation, is supported by the genetic differences feund between cultivated types originating in Mesoamerica and the Andes, in such characters as seed size and growth habit (Evans 1975). and disease resistance and adaptation (see Gepts and Bliss 1985, Kelly et al. 1987). Reproductive barriers are also present between certain .1. hi 6. ‘fl All q“. v. in: I” 18 cultivars from the two groups, which exhibit hybrid weakness and lethality (Singh and Gutierrez 198A, Gepts and Bliss 1985). Recently, findings on the distribution of seed storage protein variants have suggested a more diffuse domestication history, involving repeated incorporation of wild types into cultivation, and Colombia has been proposed as an additional domestication region (Gepts and Bliss 1986). To date, the differences between the gene pools in seed size, seed protein structure, and alleles at loci producing incompatibility, have been the best characterized means of distinguishing genotypes originating in either the Andean or Mesoamerican centers. The seed size differences are qualitative. Evans (1975) examined A,000 accessions adapted to field conditions in Cambridge, England, and classified both wild and domesticated beans into races, based on seed size and growth habit. According to her analysis the progenitor types were those with climbing growth habit and indeterminate axes, of which the smaller seeded types originated in Central America, and the larger seeded in South America. The small seeded types radiated into three sub-races, and the large seeded produced a single sub-race. These vary in growth habit and axis type, but maintain differences in range of seed size: 15 - 80 g/100 seed in the small seeded races, and A0 - 100 g/100 in the larger seeded. Differentiation between the sub-specific groups of been is also present in the molecular structure of the major seed storage protein, Phaseolin. Strong evidence for two (and possibly multiple non-centric) domestication regions has come from an examination of the geographical distribution of phaseolin variants among wild and cultivated beans. It susgests that the gene pools seen in domesticated types are a result of 19 early intraspecific divergence. Electromorphs of phaseolin,have been examined under one-dimensional sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS/PAGE) and two-dimensional isoelectric focusing SDS/PAGE, and the variation present has been shown to be the result of polymorhpism at a single locus (Brown et al. 1981, Gepts et al. 1986). The wild forms in Mesoamerica carry the phaseolin variants designated '8', or 'M' (a heterogeneous class). In Colombia, '8', 'B' or 'CH' types are present. Among wild material from the eastern Andes south of Colombia the 'T' variant (or rarely 'S') is feund. Cultivated accessions carry the phaseolin type found in local wild material, or a regional variant. In addition to 'T', the electromorphs 'C', 'H', and 'A' have been found in Andean cultivars, which include such classes as the kidney, snap, or ”pop” beans. '8' phaseolin is common to the Mesoamerican navies, pintos, tropical blacks, etc. 'T', 'C', 'S', and '8' occur in the geographically intermediate Colombian region, and this latter finding has been used to support a non-centric domestication scenario (Gepts et al. 1986, Gepts and Bliss 1986). The presence of barriers to free hybridization among cultivated P; vulgaris have been known for some time (Davis and Frazier 196A, Frazier 1967). Recent work, and results from the current study, show that some of these barriers result from genetic incompatibility between Andean and Mesoamerican types. The most fully characterized isolating mechanism produces weakness and lethality in the F when each parent carries a dominant allele at two complementary Dwarf Lethal (DL) loci. The DL and DL alleles 1 2 interact detrimentally in the offspring (Shi et al. 1980, 1981). The a 1i " .O“ 1 ‘0 use to It) {A ow \, 5:; in. v.5 ex; Ia. 'h 20 two dominant DL loci alleles have been shown to be associated with specific seed sizes and phaseolin types. In a study of material known to produce hybrid weakness, parental lines of the DL DL dl dl genotype were smaller seeded (averaging 9.8 mm in length, and1A.A m: 1: width). and carried the Mesoamerican phaseolin allele '8'. The complementary d1 d1 DL DL genotype was found in medium and large seeded lines (length 1312 mm,2wi§th 5.9 mm), carrying the 'T' and 'C' phaseolins which are associated with Andean origin (Singh and Gutierrez 198A, Gepts and Bliss 1985). Evidence for the presence of two major gene pools in the species has begun to influence genetic investigations of common bean. For example, Kelly et al. (1987) pooled yield data collected over years and locations to suggest that the genetic background of the large seeded and small seeded commercial classes was sufficiently different to bias yield stability analyses in which they were ranked together. However, a complete description of the genetic structure of the species and an exploration of the potential for recombination between the gene pools remain to be done. It is probable that complete panmixis between domesticates of the two pools has not been achieved. The need for uniformity and quality has emphasized mating among related commercial types in most breeding and improvement programs, and this has narrowed the scope of much recombination to the confines of the class, ie., those lines sharing Breen pod or mature seed traits, which are marketed or processed as a unit. The poor combining ability displayed in most crosses between Classes from different gene pools, although incompletely understood in the past, has probably contributed to reducing inter-pool exchange. 21 A restriction in genetic diversity within bean classes has been documented. Adams (1977) pointed out that the high degree of within- and even between-class homogeneity, suggested by biochemical and morphological trait similarities found in navy beans, pintos, great northern types, and other types, made common bean particularly vulnerable to genotype-specific problems. The interaction between genotype and environment (G x E) of improved cultivars within certain commercial classes has been shown to be very similar, emphasizing the genetic relatedness within each one (Ghaderi et al. 1982). In snap beans, Marques Gomes et al. (198A) examined isozyme banding patterns of feur enzymes and total protein for three snap bean lines, and concluded that the lack of differences observed was due to the inbreeding nature of the crop, the narrow genetic base of the lines, and intense selection pressure for type. Starch gel electrophoresis (SGE) has been one of the tools used to address the question of class and cultivar variability within P; vulgaris (see introductory chapter), but until the current study it has not been focused on distinguishing between the major gene pool groups. Bassiri and Adams (1978) demonstrated that homology of total isozyme patterns fer three enzymes was often high among dry bean cultivars belonging to the same commercial class, although no pattern was the exclusive property of any one class. Grouping cultivars by number of polymorphic isozyme bands in common produced clusters whose members were known to share pedigree relationships. Three European commercial classes were found to be distinguishable on the basis of data produced by densitograms of SGE isozymes from water soluble seed proteins, although members of each class had some bands in common (Adriaanse et 22 al. 1969). However, more isozyme variation has been found with the use of additional enzyme assays. Weeden (198A) identified specific allelic differences at ten enzyme loci in 97 white-seeded snap bean genotypes, and found 72 different combinations of allozymes. The most frequently occurring allele combination was present in only six lines, 6.2% of the total, and 58$ of the lines carried unique allozyme fingerprints. Although genetic variation within commercial classes of dry beans was known to be limited for certain traits, it was expected that isozyme analysis would be useful fer quantifying variability among lines from regions where systematic genetic improvement had not narrowed the germplasm base. The diversity for seed type appearance among landraces in Malawi indicated that they were highly variable genetically, and this was confirmed for morphological and phenological traits in a sample population. Therefore, an isozyme study was initiated to further characterize the variation present. However, when the lines were analyzed using SGE, the material was found to be depauperate for allozyme combinations. When it became apparent that the two most frequent allele combinations were correlated to differences in seed size, the study was broadened to investigate the allozyme variation among cultivated types from the Andean and Mesoamerican gene pools. It was hypothesized that the large and small seeded gene pools already differed for allelotype when they were introduced into Eastern Africa, and therefore, to determine whether in fact the gene combinations seen in Malawi were characteristic of the Mesoamerican and Andean gene pools, SGE isozyme analysis was conducted on a variety of lines originating outside of the country. In one group the gene pool affiliation of each line had been defined by phaseolin and DL genotype; 23 another set included a variety of cultivars and accessions thought to be from both domestication regions, but which had not been analyzed for these definitive traits. The results showed a consistent isozyme differentiation between Andean and Mesoamerican types, and gave additional evidence for the existence in common bean of two major gene pools, each "a set of genotypes characterized by similar allele frequencies and allele associations" (Gepts and Bliss 1985). In addition, the results showed that these gene pools are separated by effective post-zygotic incompatibility mechanisms. MATERIALS AND METHODS Electrophoretic Methods Isozyme variation was visualized in extracts of root, seed or leaf tissue, following the SGE protocols and enzyme activity stains Weeden used to zymotype snap bean cultivars (Weeden 198A, 1986; Weeden and Emmo n.d.), with a few modifications. Grinding buffers of .12 M reduced glutathion/tris, pH 7.6 (Rick et al. 1977), or .1 M tris maleate pH 8.0 (Weeden 198A), were found to be suitable for all tissues. Weeden's standard tank/gel buffer system of lithium borate pH 8.1/tris citrate pH 8.” was used routinely. The enzymes assayed, using Weeden's protocols unless otherwise noted, were acid phosphatase (ACP), alcohol dehydrogenase (ADH), aspartate amino transferase (AAT), catalase (CAT) (Conkle et a1 1982), NADH diaphorase (DIAP), glucose phosphate isomerase (GPI), glutamate dehydrogenase (GDH), leucine aminopeptidase (LAP), malic enzyme (ME), peroxidase, (PRX), seed protein (SDPR), rubisco (RBCO), and shikimic acid phosphatase (SKDH). The tissues assayed are listed in Table 1. Electrophoretic variants at the same locus, previously defined or found during this research, were designated as Fast (F). Slow (3). Intermediate (I) or Null (N), depending on their relative mobility or absence of activity. In heterozygous or heterogeneous accessions both alleles present were reported. A simple procedure for visualizing phaseolin variation using SGE was introduced. The general protein stain of naphthol blue black, 2” 25 Table 1. Tissues assayed for enzyme activity following SGE. Enzyme/Protein Leaf Root Seed ACP Acid phosphatase X X ADH Alcohol dehydrogenase X AAT Aspartate aminotransferase X X X CAT Catalase X DIA NADH Diaphorase X GPI Glucose phosphate isomerase X GDH Glutamate dehydrogenase X LAP Leucine aminopeptidase X ME Malic enzyme X X X PRX Peroxidase X X SDPR Seed protein X RBCO Rubisco X SKDH Shikimic acid dehydrogenase X X X 26 dissolved 1 mg/ml in 5:5:1 methanol/water/acetic acid, used to visualize rubisco protein in leaves (Weeden, 198A), was applied to extracts of imbibed seed cotyledon tissue following electrophoresis in the lithium borate/tris citrate buffer system. One major band appeared per zymogram, and this was assumed to be an isozyme of phaseolin, which comprises 36 - A6} of total seed protein in beans (Ma and Bliss 1978). The isozyme is referred to as a seed protein (SDPR) variant in the present study. Plant Materials Three groups of lines were examined: landraces from Malawi; lines known to belong to a specific gene pool on the basis of phaseolin or DL genotype; and diverse cultivars and accessions which had not been tested for phaseolin or DL genotype. Samples of bean landraces grown by 15 small farmers in Malawi were collected in 1983, and a random selection of 25 lines was made from each farm. Measurements of morphological and phenological characters made under experimental field conditions in Malawi showed that significant genetic variation was present among the 375 lines (Martin and Adams 1987a and 1987b). The whole population contained approximately 32 seed types, or classes, distinguishable by seed color, shape, size and pattern. Seed size ranged from 16.7 to 62.9 g/100 seed (Martin 198A). According to the seed size groupings used to classify commercial bean cultivars, the lines were distributed as follows: 5.91 were small seeded, weighing less than 25 g/100 seed; 29.61 were medium seeded, between 25 and A0 g/100; and 6A.5$ were large seeded, more than no g/100. 27 The lines were brought to Michigan State University (MSU) in 198A and increased in the greenhouse or field, so that the material analyzed with SGE was from two to feur presumably selfed generations away from the single progenitor seed collected in the field. An average of six seeds or plants from 373 lines (two lines were lost during increase at MSU) were analyzed fer the enzymes DIAP, ME, PRX, SDPR, RBCO and SKDH. Some of the morphological data collected by Martin (198“) was subsequently analyzed in combination with the isozyme results. The six cultivars in which were first identified the phaseolin types feund most frequently in cultivated beans were chosen for analysis, in the expectation that they would have genetic backgrounds typical of their respective gene pools. These lines were 'Tendergreen' carrying 'T' phaseolin, 'Contender' with 'C', 'Ayacucho' with 'A', and 'Nuna de Huevo de Huanchaco' with 'H', from the Andean pool; and 'Sanilac' with 'S', and 'Boyaca 22' with 'B' from the Mesoamerican (Brown et al. 1981, Gepts et al. 1986). These cultivars were supplied by P. Gepts and F. Bliss from the University of Wisconsin. They had identified the 'Boyaca 22' accession as being a mixture of 'B' and 'S' genotypes. Twenty-seven lines previously identified as to DL genotype by Singh and Gutierrez (198“) were requested from the Centre Internacional de Agricultura Tropical (CIAT), Cali, Colombia. There were six lines of the genotype DL DL dl d1 , associated with the Mesoamerican pool; 15 1 1 2 2 carrying d1 d1 DL DL , from Andean germplasm; and six of the neutral 1 1 2 2 recessive genotype d1 d1 d1 d1 , which is able to cross to either 1 1 2 2 dominant type without producing hybrid weakness. The phaseolin type of 21 of these 27 lines had been published (Brown et a1. 1982, Gepts and 28 Bliss 1985). The phaseolin and DL lines (Table 2) were analyzed for the 13 enzymes listed in Table 1. Fifty-seven cultivars from North and South America were analyzed for one or more of the enzymes DIAP, ME, PRX, SDPR, RBCO, and SKDH. A collection of 116 cultivars and accessions, received from CIAT upon request for a selection of black seeded material of various seed sizes and growth habits, was analyzed for two enzymes, DIAP and SDPR. ht. ”Eu a." nan «vi 9:” MM. Lav. mun u... a...“ find n... “3.. 2“ and .0“ min a." and. .v-w 2!... ur.. ”5.. turn env. a.“ .15.. Mn... .3. 29 Table 2. Phaseolin and Dwarf Lethal genotypes assayed for isozymes. 1 2 Identification Origin CIAT i Phaseolin Seed Protein a. Lines carrying phaseolin types(3) Ayacucho Peru A F PI 313590 Boyaca 22 Colombia 8.8 S Contender U.S.A. C F Nuna Huevo de Huanchaco Peru G12588 H F Sanilac U.S.A. GAA98 S S Tendergreen U.S.A. T F b. Lines carrying Dwarf Lethal alleles(4) Aysekadin - Turkey G153 C F Cali Fasulya Turkey G159 S,C S PI 165A35 Collacatlan Mexico 6278 S 8 PI 176712 Turkey G568 T,C F Barbunya Terkey G623 C F Barbunya Turkey G688 S,C F PI 206983 Turkey G910 - F PI 313755 Mexico G2618 - F Zacaticano PI 319665 Mexico G2858 - S Bolivia 6 (p752) Bolivia o3aou T F Brasil 2 = Pico de Oro Brazil G380? - S Carioca Brazil GA017 S S Cuilapa 72 (P 691) Guatemala GA489 S S Brazil 37A Brazil 65066 T F Sacavem 597 Brazil G5129 T F Brazil 668 Brazil G71A8 S S Tortolas Dianas Chile 67160 C F Maestro U.S.A. G756A T F Aragon W. Germany G7613 T F Coco Bicolore du Pape W. Germany 67633 T F Coco Rose W. Germany 67635 T F A30 Colombia - S BAT 332 Colombia S S BAT 1061 Colombia S S Diacol Calima Colombia - F ICA L23 Colombia T F ICA Pijao Colombia 8 S 1 :From SDS/PAGE analysis, Gepts and Bliss 1985 .2 From SGE analysis . 3 “From Gepts et a1. 1986, Gepts and Bliss 1986 From Singh and Gutierrez, 19811. v,- 0 v. Q. ‘~ 2L: ~ 1"“ MA. Id». at . t) (J f' ‘d RESULTS Isozyme Data The allozymes observed fer the loci Ac , Mg, 255, gng, and S522 corresponded to those in genotypes previously described by Weeden (198A, 1986) which were used as controls in each gel. The DIAP assay of root tissue produced six combinations of bands. Three of these had been reported (Weeden 198A), and three were seen for the first time. The DIAP isozymes were defined as the tetrameric enzyme products of seven alleles at two NADH diaphorase loci, REEEZl with Fast, Slow, Intermediate and Null alleles, and 212222 with Fast, Slow and Null alleles. A discussion of this system is the subject of Chapter Two. Table 3 gives the relative mobility values for isozymes observed at polymorphic loci in the material analyzed. Application of a general protein stain to extracts of seed cotyledons produced one major band which appeared in two mobility variants. Under the electrophoretic conditions used, the leading anodal edge of the more slowly migrating SDPR band moved 0.92 :_.02 as far as the faster band. 'Sanilac' and other lines (eg. 'Bunsi', 'Rufus', 'Jamapa', 'Pinto UI 111') known to carry '3' phaseolin from SDS/PAGE analyses (Brown et al. 1982) produced the slower band, as did the .accession 'Boyaca 22,‘ which is a mixture of 'S' and 'B' phaseolin ‘tYpes. 'Tendergreen', 'Contender', 'Ayacucho' and 'Nuna Huevo de Huanchaco' produced the faster band. The mobility variation of SDPR in SGE appears to represent the 30 h. 5». .1 \P~ 5‘... 31 1 Table 3. Mobility variants observed at eight isozyme loci. Mobility Enzyme/Protein Locus of allelic variant Relative to Fast Slow Intern. Acid phosphatase £22 .25 .21 anodal front NADH Diaphorase 2&2221 .72 .65 .69 anodal frontz " Diap:2 .10 -.20 anodal frontz Malic enzyme Mg .3A .29 anodal front Peroxidase £55 -.56 -.A8 anodal front Seed protein §QEE 1.0 .92 faster band3 Rubisco rch 1.0 .85 faster band3 Shikimic acid Skgh .A9 .A3 anodal front dehydrogenase 1 Following A to 5 hr electrophoresis in lithium borate pH 8.1/tris citrate pH 8." SGE buffer system, as described in text. 2 These are Rf values fer isozymes of homotetramers of the sub-units of the allele. 3 These values measured relative to the faster band after staining, due to deydration of gel during staining. g» L) kl ..' 32 migration difference seen in the most anodal band of the phaseolin zymotypes produced in one-dimensional SDS/PAGE. Among 'T', 'C', 'H', and 'A' phaseolins this band co-migrates, and its position is anodal to the corresponding band which co-migrates in 'S', 'B' and 'CH' electromorphs. (see Gepts et al. 1986, Gepts and Bliss 1986). Thus, this SGE technique discriminates between phaseolin types from the Mesoamerican or Andean centers, but not within types from the same region. A comparison of the results of SDS/PAGE and SGE analysis shows the correlation between the two methods (Table 2). Where the phaseolin type is known to be 'T', 'C', 'H' or 'A', from SDS/PAGE (Gepts and Bliss 1985). SGE of seed protein produces the Fast band. Lines having '8' and 'B' phaseolin produce the Slow seed protein band. In cases where phaseolin from both gene pools occurred in one accession (i.e., Cali Fasulya, Barbunya G688), the isozyme seen on SGE corresponded to only one of the types. This may have been a result of testing only a few seeds per accession. These results show that SGE can provide a quick method of distinguishing phaseolin types between the two major gene pools, and in the following discussion the Fast SDPR variant is assumed to identify Andean phaseolin types, and the Slow variant Mesoamerican. Because phaseolin types are reported to be allelic (Brown et al. 1981, Gepts et al. 1986), the SDPR bands are assumed to represent allozymes of a single locus, designated Sdpr in the following discussion. Population Analyses Appendix A lists isozyme allele data fer the population of 373 Pflilawian landrace lines. In the sample, 317 lines were homozygous at all seven loci examined, Diap-1, Diap-2, Mg, Prx, Sdpr, rch, and Skdh. 33 Diang was monomorphic for the Slow allele in the entire group. Fifty- six lines were heterozygous or segregating at one or more of the six polymorphic loci. They averaged 3.9 heterogeneous loci per line, and had a total heterozygosity of H = .6A3. These lines were assumed to result from selection of a heterozygous seed during field sampling in Malawi, rather from cross pollination during seed increase at MSU. The total average heterozygosity for the population was H : .097. The lines sampled from two farms (#10 and 12) had much higher levels of heterozygosity, H = .577, while on the remaining 13 farms the average was .022, suggesting that some anomalous influence associated with the individual farm or farmers caused the cases (Table Na and b). The gene frequencies for all the loci examined were very similar, with one allozyme at a frequency of about .8, and the other at about .2 (Table 5). Because each enzyme protocol identified one polymorphic locus with two variants, the SGE analysis had the potential to distinguish 6A allele combinations in the population. However, only 12 different combinations were seen in the 317 homozygous lines, and two of these genotypes accounted for 91.2% of the material. These two combinations consisted, respectively, of the six high frequency alleles and the six low frequency alleles. Each allele combination was given a consecutive identification number. The most common, #1, was present in 2A1 lines, .760 of the homozygous population; the second, #7, occurred in N8 lines, a frequency of .151. Ten other combinations were present in a total of 28 lines, 8.8} of the material (Table 6). From an examination of the lines, it appeared that combination #1 was associated with larger seeded types, and #7 with the smaller seeded. 3A 1 Table Aa. Isozyme genotypes in 373 lines from 15 Malawian farms. No. of lines carrying the isozyme genotype at each farm Isozyme —— — ----------------- ——= ------------- genotype 1 2 3 A 5 6 7 8 9 10 11 12 13 1A 15 DIAP-1 Fast 25 2A 2A 13 12 23 2A 3 23 8 25 8 25 22 23 Slow O 1 1 11 12 0 1 22 1 2 0 0 0 0 1 Het O 0 O 1 1 0 0 O 15 0 17 0 3 1 DIAP-2 Slow 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 ME Fast 5 3 2 1A 12 O 1 22 2 2 O O O 1 2 Slow 20 22 23 10 12 23 24 3 21 9 25 8 2a 19 23 Hat 0 O 0 1 1 0 O O 2 1A 0 17 1 5 O PRX Fast 3 1 1 1O 16 O 1 22 1 2 O O O O 1 Slow 22 2H 2H 13 8 23 24 3 23 9 25 11 2” 23 23 Hat 0 0 0 2 1H 1 2 1 _s O C O _a _s .B O SDPR Fast 2" 2A 2” 12 10 23 2“ 3 2H 10 25 13 2” 2A 23 Slow 1 1 1 12 15 0 1 22 1 3 0 O O O 1 Hat 0 O O 1 O 0 O O O 12 O 12 1 1 1 R300 _- O O C O _. Fast 0 1 1 12 13 0 2 22 1 Slow 25 23 2" 12 12 23 23 3 23 9 25 11 23 23 23 Het O 0 O 1 0 0 O 0 1 15 0 1n 2 2 1 SKDH Fast 3 5 1 11 13 O 1 22 1 O O O O 0 1 Slow 22 20 2A 12 12 23 2A 3 2H 11 25 1O 24 22 23 Hat 0 0 0 2 O 0 0 O 0 1H 0 15 1 3 1 1 Each farm sample consisted of 25 lines. Accessions heterozygous or segregating at each locus are designated "Het". 35 Table Ab. Genotype frequencies at six loci in 373 Malawian lines. Genotype frequencies Genotype ------------------------------------------------------- Diap-1 Mg 255 Sdpr rch Skdh FAST .756 .177 .156 .769 .1A5 .156 SLOW .139 .713 .7A8 .156 .759 .7A8 HET .105 .110 .097 .075 .097 .097 Table 5. Allele frequencies at seven loci in 373 Malawian lines. Enzyme Allele frequency locus ........................... Fast Slow M .826 .1711 233222 0 1.0 Mg .232 .768 £55 .20A .796 __2g .807 .193 rch .193 .807 Skdh .20” .796 36 1 Table 6. Allele combinations observed in 373 Malawian lines. Allele at Enzyme Locus Allozyme Gene ---------------------------------------- Combination No.of Pool Diap-1 Diap-2 Mg Prx Sdpr rch Skdh # Lines Freq. 1 2A1 A .760 F S S S F S S A 11 A .035 F S F S F S S 2 7 A .022 F S S S F S F 3 2 A .006 F S F F F S S 11 2 A .006 F S S F F S S 9 1 A .003 F S S S S S S 10 1 A .003 F S S F S S S 12 1 A .003 F S S S F F S [3:266 7 A8 M .151 S S F F S F F 5 1 M .003 F S F F S S S 6 1 M .003 S S S F S F F 8 1 M .003 F S F S S F F n: 51 0 n = 56 .150 Segregating at one or more loci Sorted according to gene pool on basis of similarity to combination #1 and #7. A = Andean, M = Mesoamerican. 37 This association was verified when seed character data collected in Malawi (Martin 198A) were analyzed by allele combination. Lines carrying combination #1 and #7 were significantly different for seed length, width, and weight (Table 7). Length and width differences were comparable to those found to occur in bean cultivars belonging to the different gene pools (Gepts and Bliss 1985). These findings, along with the occurrence of the Fast SDPR electromorph in allele combination #1, and the Slow in combination #7, supported the division of the Malawi sample into Andean and Mesoamerican types for further analysis. Therefere, the remaining 28 homozygous lines were grouped with combinations #1 or #7 on the basis of similarity in allozyme genotype. Twenty-five lines differing for one to two alleles from combination #1 were associated with it to produce a sub-population of 266 Andean lines; the three lines most similar to combination #7 were grouped to form a total of 51 Mesoamerican lines (Table 6). A histogram of seed weights found in these two sub-populations and in the 56 segregating lines (Figure 1) showed that the heterozygous material was intermediate in seed size between the divergent mean seed weights feund in combination #1 (A7.A g/100) and #7 (26.5 g/100) (Table 7). This is consistent with the expectation that, since the most frequent allozyme combinations in the population differ at all loci, the heterozygotes identifed by isozyme analysis would be predominantly those resulting from crosses between combinations #1 and #7. Consequently, the segregating lines were considered to be hybrids between Andean and Mesoamerican types. Significant differences between the gene pool sub-populations were also observed in 18 other morphological and phenological traits for 38 1 Table 7. Seed size traits of allele combinations #1 and #7. Combin.# no. Seed length (cm) Seed width (cm) Seed weight (3/100) 1 241 1.56 + 7 A8 1.0A + .2 .77 + .1 117.1: 1 7.2 .1 .66 + .01 26.5 I 8.6 t test sig. at P 1 .01 .05 .01 1 Seed size measurements from Martin (198A). 39 Figure 1. Seed size frequency histogram, 373 Malawian landrace lines. The population consisted of 266 Andean lines, 51 Meso- american lines, and 56 lines heterozygous at one or more enzyme loci. See Table 6. 40 .H shaman Goo. Sis Eons, Bum we 3 Jaqwnu Jaqwnu quwnu A1 which data had been collected in Malawi (Martin 198A) (Table 8). Analysis of the 33 lines carrying gene pool specific phaseolin and DL genotypes showed substantially the same allozyme differentiation between domestication centers. Although polymorphism had been identified at a GPI locus (Epizgl) (Weeden 1986), variation was not seen in this material under the buffer systems used here; ADH, AAT, CAT, GDH, and LAP also appeared to be monomorphic. Isozyme data from eight polymorphic loci, Acp, Diap31, Diap-Z, Mg, Prx, Sdpr, rch and Skdh, were summarized on the basis of the gene pool affiliation of the lines in which they were observed (Table 9). In the case of the five lines carrying the neutral recessive DL genotype, whose phaseolin type was unknown, SDPR genotype was used to place them into the Mesoamerican or Andean pool. The frequency of the predominant allele among lines from each pool was calculated (Table 13). Among the seven polymorphic loci analyzed in common with the Malawian sample, the most frequent alleles in the Andean and Mesoamerican groups were those found in allozyme combinations #1 and #7, respectively. Isozymes specific to sub-groups within the gene pools appeared in this material. While the Diang S allele was most frequent in both pools, the 2i22:§.§ allele was feund only in the Mesoamerican, and Diap: 2‘! occurred only in the Andean. These associations were confirmed in later research on DIAP (see Chapter Two). Table 10 summarizes the allozymes found in 57 North and South American cultivars. A full analysis of each line for every locus was not done, but all the data produced is reported. Homogeneity across commercial classes within the more numerous Mesoamerican gene pool is A2 1 Table 8. Differences between 266 Andean and 51 Mesoamerican lines. Quantitative Gene Pool Means Sig. diff. Trait Andean Mesoamer. P<.01 Days to emergence 12.9 12.A *' Days to first flower 50.0 53.6 ** Days to end of flowering 68.2 73.5 " Duration of flowering 18.3 20.0 *' Days to maturity 95.A 98.0 ** Leaflet area cm2 A9.0 3A.3 ** Leaflet length cm 10.8 8.8 '* Leaflet width cm 7.0 6.1 ** Leaflet shape 1.6 1.A '* Petiole length cm 5.8 5.2 ** Hypocotyl length cm 6.8 6.A '* Hypocotyl diameter 1/32" 8.3 7.A " Seedling weight gm 1.3 0.82 '* Nodes on the main stem 1A.8 19.2 " No. of seeds per pod A.A 5.2 " No. of pods per plant 27.2 A7.7 *' Seed yield/plant gm 55.3 63.5 " No. of seeds per plant 19.3 251.5 ** 1 Quantitative data was measured in replicated field plots in Malawi by Martin (198A). The 317 homozygous Malawian lines were assigned to the Andean or Mesoamerican gene pools on the basis of allele combinations. See Table 6. A3 Table 9. Allozymes observed in phaseolin and Dwarf Lethal genotypes. Gene Pool Enzyme Allele w/ phaseolin type Acp Diap-1 Diap-Z Mg Prx Sdpr rch Skdh I. Mesoamerican a. Lines carrying phaseolin types B Boyaca S S S F F S F S Sanilac S S S F F S F F b. Lines carrying Dwarf Lethal genotype DL DL d1 d1 1 1 2 2 S 6278 S S S F F S F F S 6A017 S S S F S S F F S 6AA89 S S F F F S F F S G71A8 S S S F S S F F S BAT 332 S S S F F S F F S BAT 1061 S S F S F S F F c. Lines carrying neutral Dwarf Lethal genotype d1 d1 d1 d1 1 1 2 2 - 62858 S S S F F S F F - 63807 S S S S F S F F - A30 S S F S F S F S/F S ICA Pijao S S F F F S F F II. Andean a. Lines carrying phaseolin types A Ayacucho F F S S S F S S C Contender S F S S S F S S H Huanchaco F F S S S F S S T Tendergreen S F S S S F S S b. Lines carrying Dwarf Lethal genotype d1 d1 DL DL 1 1 2 2 C G153 F F S S S F S S S,C 6159 S S N F F S S F T,C 6568 F F S S S F S S C 6623 F S N S S F S S S,C 6688 F S S/N S S F S S - 6910 F S N S S F S S T 6380A S S S F S F F S T 65066 F F S S S F S S T 65129 F F S S S F S S C 67160 F F S S S F S S T 6756A F/S F S S S F S S T 67613 F F S S S F S S T 67633 F F S S S F S S T 67635 F F S S S F S S T ICA L23 F F S S S F F S c. Lines carrying Dwarf Lethal genotype dl1dl1d12d12 - 62618 S S F F S F F F - Diacol Calima F F S S S F F S 1 Phaseolin type data from Gepts and Bliss (1985). AA Table 10. Genotypes of commercial classes from the two gene pools. Gene Pool Enzyme Alleles Class Cultivar Diap-1 Diap-2 Mg Prx Sdpr rch Skdh I. Mesoamerican a. Central American Blacks 28-5-1 S F F F F F B-190 S F F F F F Black Magic S F F F F Black Turtle Soup S F/S F F/S F Domino S F F Jalpatagua 72 S S S F F F Jamapa S F S F S F F La Vega F F S F Midnight S F S F F Porillo Syntetico S F F F S F F Puebla 152 S S F F S F San Fernando S F F S F F T39 S F F S F F b. Great Northern 6N 11A0 S S S F S F Harris S S S F F JM 2A S F S F F KOAAO S S F F F 01 59 S S F F F Valley S S F F F c. Navy, Bush Laker S S F F F Sanilac S S F F S F F Seafarer S S F F F F d. Navy, Upright C-20 S F F F F F Neptune S F S F F F Swan Valley S F F F F F e. Small White Aurora S F F F F F Bunsi S F/S F F S F Hyden S S S S F Monument S F F F F F A5 Table 10 (cont'd.) f. Pinto Agate S S Amber S S Cahone S S Carioca S S Chiapas 7 S S Fiesta S S JM 126 I S Olathe S S Ouray S S Pindak S S UI 111 S S UI 11A S S g. Red Mexican and Pinks Big Bend I S Desarrural-1 S S GR 760 S S Garnet I S K0228 S S NW 59 I S NW 63 I S Rufus I S UI 36 I S Victor I S Viva I/S S II. Andean a. Cranberry Cardinal F 3 A23 F S b. Kidney Isabella F S Mecosta F S Montcalm F S cn'n'n'u 'u-m'n'n 'u-u'u'n'n 'n'u'n'n'n'u'n 'n'n'n'u'n'n'n 'fl'fl'llm "llUJUJUJ (0'31'9'31 A6 apparent, and the most common alleles in this group are those of gene combination #7, found in small seeded Malawian material. The Andean cultivars carry the alleles of combination #1. In the population of 116 black seeded lines a significant association between SDPR and DIAP alleles was observed (Table 11). Andean SDPR was found in 2A of the 25 lines carrying Diap-1 Fast and Diap-1 Null, and 95.61 of the lines carrying Mesoamerican SDPR carried the Diap-1 Slow or Diap-1 Intermediate alleles. These DIAP and SDPR combinations correspond to those found among the 33 lines carrying gene pool specific phaseolin and DL genotypes. Seed size differences between groups carrying the DIAP alleles characteristic of the Mesoamerican and Andean pools were significant. A compilation of the allozymes found to be associated with each pool in these populations of beans is given in Table 12. A summary of the predominant allele frequencies appears in Table 13. Male Sterility A syndrome producing near sterility was found in 13 lines from the Malawi population during seed increase in the greenhouses at MSU. Vegetative growth of the plants was normal but pollen did not dehisce, and unlike what usually occurs in P; vulgaris, the stigmas of opened flowers were not pollen-covered. Numerous small pods, 3 to 5 cm long, were produced on racemes and remained on the plant for A to 6 weeks, but ovules were not present. Flowering continued over a much longer time than expected, and senescence was delayed. On most plants one or two seeds were produced, usually as single seeds in small pods. When isozyme data for these lines were examined, 10 of the 13 lines were A7 Table 11. Gene pool differences among 116 black seeded lines. Seed Protein No. of lines carrying Diap-1 alleles Type --------------------------------------- Fast or Null Slow or Interm. Total Andean 2A A 28 Mesoamerican 1 87 88 Total 25 91 166 1 Mean Seed Size 38.8 23.5 (g/100 seed) a b 1 Different letters indicate significant differences, LSD P<.01. Table 12. Allozymes associated with the two major gene pools in bean. Enzyme Allele Gene Pool Acp Diap-1 Diap-2 Mg Prx Sdpr rch Skdh Mesoamerican S S,I S,F F F S F Andean F F,N S,N S S F S ".2"- A8 1 Table 13. Frequency of predominant allele in samples from gene pools. Gene Pool Frequency of predominant allele in each pool and Population Acp Diap-1 Diap-2 Mg Prx Sdpr rch Skdh I. Mesoamerican S S S F F S F F Phaseolin/DL (12) 1.0 1.0 .66 .75 .83 1.0 1.0 .96 Cultivars (57) .83 .69 .73 .88 1.0 1.0 .98 Black seeded (88) .93 .73 - - - - - II. Andean F F S S S F S S Malawian (266) 1.0 1.0 .95 .98 .99 .99 .97 Phaseolin/BL (21) .7A .71 .79 .86 .95 .95 .81 .90 Cultivars (5) - 1.0 1.0 1.0 1.0 1.0 1.0 1.0 Black seeded (28) - .82 .93 - - - - - 1 Number of lines in each sample population shown in parentheses. A9 found to be heterozygous or segregating for one or more of the six polymorphic loci analyzed in this population (Table 1A). This suggests that the lines were the result of crosses between Andean and Mesoamerican types, and indicates that the sterility is a result of a post-F incompatibility mechanism between the gene pools. 1 50 1 Table 1A. Isozyme genotypes of 13 male sterile Malawian lines. Enzyme Allele Allozyme Farm/Line combination Diap-1 Mg Prx Sdpr rch Skdh # 3 13 1 F S S F S S 7 6 12 F S S F F S 10 2 O F S - - - S 10 3 O - S F - - - 10 A 0 - S S F F 10 16 0 - - - S - S 11 12 1 F S S F S S 12 3 0 - - - F - S 12 7 0 F - - - s s 12 16 O - - S F S - 12 19 O - - - - - - 13 A O F - - - - - 1A 16 O F - S F S S 1 Allozyme combinations detailed in Table 6. DISCUSSION Combined information from unimproved Malawian landraces, improved cultivars, and lines varying for phaseolin and DL alleles, shows that domesticates of the two major gene pools of Phaseolus vulgaris are characterized by the presence of specific alleles at eight enzyme loci (Table 12). Homogeneity for these electromorphs is found within gene pool across diverse commercial classes and morphological variants (Tables 9 and 10). Tests using SGE analysis of segregating populations have shown no linkage among loci coding fer SKDH, GPI, RBCO, DIAP, ME and PRX in common bean (Weeden 1986), and thus linkage is not a plausible explanation for the association between these alleles and the respective gene pools. Rather, the presence of this differentiation is suggested to result from pre-domestication divergence and natural selection, and the subsequent generation of coadapted linkage groups and reproductive isolating mechanisms. These isozyme results contribute to an understanding of the structure of P; vulgaris in two ways. They confirm and expand the genetic definition of the domesticated sub-specific taxa by identifying additional allele associations in each pool. They also provide evidence for strong reproductive barriers operating by genetic mechanisms rather than geographic isolation between the pools. In addition, these results generate speculation on relationships within the species and the genus. 51 52 Gene Pool Differentiation and the Sequence of Divergence While it is not yet known whether the pool-specific allozymes found in cultivated types are an artifact resulting from selection during domestication, or whether they reflect restricted variation in the wild, the isozyme data can be used to confirm that the wild progenitors had diverged into two gene pools prior to domestication, and to infer that they did not contain common alleles at the loci examined. Analysis of wild material may reveal other polymorphic loci and additional allelic variants, which were excluded during selection for cultivated types. The example of the highly variable class of 'M' phaseolin, found only in the wild (Gepts et al 1986), supports this expectation. However, it is unlikely that only dissimilar isozyme alleles would have been selected during domestication in each center if the same adapted alleles were present in both of the wild pools. The similarities fer isozyme genotype, among lines differing for phaseolin type within the same region, suggest that certain seed protein variants appeared after the enzyme loci diverged. The type cultivars carrying 'B' and 'S' phaseolin are the same at eight enzyme loci; those carrying 'A' and 'H' differ from 'C' and 'T' only for ACP (Table 9). A combination of phaseolin and additional isozyme data from cultivated and wild types will be useful in defining the sequence of divergence in the species. The hypothesis that the isozyme differentiaton seen in this study predates domestication is also supported by allozyme homogeneity between other sub-groups within the same gene pool. The Mesoamerican pintos and tropical blacks and reds are very similar in their allele complements (Table 10), in spite of their respective adaptations to dry upland and 53 humid tropic environments, and divergence for other traits. This suggests early differentiation from the Andean pool, and but it also indicates retention of certain coadapted allelic complexes during the process of further allopatric adaptation. In the case of the locus 2322:3' the same allele, Signg S, is most common in both gene pools, but the variants 2132:; E and 222223.! are specific to the Mesoamerican and Andean pools respectively (Table 12). This suggests that variation arose at this locus after divergence into the two groups. The 232223.: allele appears to be endemic in the tropical black seed class, and it may be possible to trace the emergence of this group within the Mesoamerican gene pool by finding an intra- or extra-specific source of the allele. The 2£2£:§.! allele has been found in several seed classes, but only in cultivars originating in Turkey. This suggests that it is of modern origin. Reproductive Barriers The Malawian material examined in this study reveals the strength of the reproductive barriers which operate between gene pools in g; vuggaris. Isozyme data showed that 15$ of the lines in the sample were heterozygous, probably the result of inter-pool crosses. But only 8.8% of the morphologically diverse population consisted of homozygous recombinations of the suite of alleles specific to each gene pool. It appears that post-zygotic reproductive isolating mechanisms are best able to account for this low level of successful inter-pool gene flow. Domesticated beans, presumably from both gene pools, were brought to Africa approximatley 350 years ago, and found suitable natural and agricultural environments. In Central and Eastern Africa they are a 5A major feed crop, and in some areas attain the status of a dietary staple (Nyabyenda et al. 1981). A diverse array of different seed classes is grown in the area, and the full range of seed sizes in domesticated types can be feund. Although large seeded classes are more popular and more common (the lower frequency of small seeded lines may reflect a slightly poorer adaptation of this germplasm, in addition to consumer preference), both large and small seeded types are usually maintained on each farm in Malawi. Landrace lines differing for seed type, plant habit, maturity date, etc. are traditionally planted within the same plot; and numerous small plots belonging to different farmers lie close to each other in the agricultural landscape. Beans are predominantly a self-pollinating crop, but approximately 1.01 outcrossing has been measured in Malawi (Martin and Adams 1987b). Malawi stands in contrast to most parts of Central or Andean America, where the traditional varietal mixtures being planted consist only of types belonging to the regional gene pool (Gentry 1969, Berglund-Brucher and Brucher 1976, Kaplan 1981). Thus, opportunities for gene flow between the domesticated pools have probably been less common in the crop's centers of origin than in some areas where it has been introduced. While no wild or weedy forms of g; vulgaris are known in Malawi, and the related 2; coccineus, g; lunatus and g; acutifolius are rare, Malawi appears to be a "microcenter" of crop variation, where introgression among domestic races may be expected to occur (Harlan 1970). There does not appear to be a pre-zygotic barrier due to reduced length of flowering coincidence between the two predominant allozyme genotypes in the Malawi sample. Analysis of data collected locally 55 (Martin 198A) showed that the homozygous Andean and Mesoamerican lines had an average overlap in flowering time of 1A.6 days, when planted on the same date (Table 8). Differential pollinator visitation between gene pool types was not studied. The lower frequency of small seeded lines in most farmers' fields means that crossing between pools is less common than crossing within the large seeded pool. Based on the frequency of allele combinations #1 (.760) and #7 (.151) among the homozygous lines in the sample, 231 of the hybridizations are expected to result in genetic exchange between the pools. The presence of isozyme heterozygotes confirms that such matings occur, but the lack of homozygous recombinants suggests that there is selection against later-generation segregants. It can be hypothesized that growers apply divergent selection for traits (such as those in Table 8) associated with large and small seeded types, and thus indirectly affect the frequency of allozyme combinations. But the continuous variation characteristic of so many agronomic traits, combined with the wide range of locally acceptable phenotypes, probably prevent humans from being an effective source of selection against recombinant types. The possibility of farmer selection against lines segregating fer seed coat patterns is discussed in Chapter Three. The activity of natural selection on inter-pool recombinants has not been studied. While genotypes from both pools are suited to Malawi's central plateau, it is possible that recombinant types are not seen in the population because they are maladapted. The geographical separation between the gene pools in bean and the evidence fer genetic divergence suggest that different coadapted complexes may be present. 56 The observed low frequency of allozyme recombinants could result from the linkage of the enzyme alleles studied to internally coadapted gene blocks, of the kind described by Allard et al. (1972). Homozygous segregants would then be maladapted or less fit than parental types. This mechanism presents a type of post-zygotic incompatibility barrier. The presence of genetic incompatibility produced by DL loci has not been verified in Malawian material, but sub-vitality and lethality is known to occur in the F and later generations of intermatings between certain local large and1small seeded types (E. Ayeh, C. Madata, personal communication). Based on the presence and distribution of dominant DL alleles in New Werld beans (Singh and Gutierrez 198A), it would be expected that the majority of beans introduced into Malawi carried the dominant DL genotypes appropriate to their respective gene pool. However, offspring from controlled hybridizations of seven pairs of Malawian lines carrying isozyme combinations #1 and #7 did not produce sub-vital F offspring. Hybrid F seed were set on vigorous plants in the field, glthough careful obsersations were not made fer lethality at that stage (Sprecher and Elizondo-Barron, unpublished data). However, an additional inter-pool barrier does appear to be operating. A syndrome consisting of indehiscent anthers, sterile pods, and extended vegetative growth was observed among Malawian lines, most of which resulted from gene pool crosses. Of the 13 sterile lines, ten were heterozygous for an average of 3.5 of the isozyme loci studied (Table 1A). 0f the three homozygous lines, two carried the Andean allozyme combination #1, and one differed from it only at the RBCO locus. These results suggest that this syndrome results from incompatibilities between the genomes of the two pools, appearing later 57 than the F , and reducing gene flow. Simil;r male sterility had been described previously but had not been associated with gene pool or seed size differences. The syndrome of extended flowering time, and numerous small sterile pods, was described by Frazier in selections of OSU 9161 and OSU 172 (1967). Male sterility caused by non-dehiscence of anthers and low pollen viability was feund in a population derived from snap bean lines by Ibrahim et al. (197A). Agbo and Weed found reduced pollen viability in F progeny of the cross 'Ouray' x 'Aurora'. Both of these lines are fro: the Mesoamerican gene pool, but 'Aurora' is closely related to the tropical black group, and 'Ouray' is an upland-adapted pinto. Van Rheenen et al. (1979) noticed F plants with an extended flowering period, nearly empty anthers, and polIenless stigmas, which produced only a single seed each. These authors showed that the progeny segregated for sterility in a manner consistent with control by a single recessive gene. However, crosses between the male fertile and infertile segregants in the F were successful only 2.61 of the time. The parents of the cross in whiZh the steriles occurred were 6LP-2, 'Roko' (680A3), a large seeded rose coco, plant habit Type I, and GLP-16, (68050) a small round cream/beige seeded bean, plant habit Type I (Hidalgo 198A). In this case the parents may belong to the Andean and Mesoamerican gene pools. Finding sterility and semi-sterility in the offspring of a cross between WI 7A-20A7 and 'Swedish Brown', Mutschler and Bliss (1980) postulated that the effects were produced by two unlinked loci and a reciprocal translocation. Wyatt (198A) described the shrunken appearance of anthers from indehiscent anther (IA) mutants produced by crosses between white and colored-seeded snap beans. Manual self-pollination of the steriles 58 produced A8 - 811 successful fertilizations, while crosses with a black- seeded snap bean were 53 - 931 successful, so that female fertility also appeared to be affected. Under field conditions only 231 of the IA types reproduced. These symptoms may not all result from the same cause, and examination of the parental lines will be required to show whether they are the result of between-pool incompatibilities. However, some reported differences in seed size, and the occurrence of male sterility in the more diverse snap bean hybrid populations, suggest that such a mechanism is operating. Additional recombination barriers which occur in the F and F are known in breeding and improvement work in bean (eg. E. Ayehf in MaIawi; M. W. Adams and J. D. Kelly, unpublished data), and may contribute to lack of recombination in Malawi. These involve recessive mechanisms, different from the dominant DL gene system, but have not been completely described. The lack of recombinant types among the dry bean lines studied here contrasts with the numerous combinations of allozymes at ten loci found in snap beans by Weeden (198A). Among the 97 lines zymotyped by Weeden, the most common allele combination for the ACP, DIAP-1, ME, PRX, RBCO, and SKDH loci common to these two studies, was the one feund here to predominate in the Andean gene pool (Table 12). It occurred in 12.A1 of the lines. The Mesoamerican combination of alleles was not present. This finding is consistent with the predominance of 'T' and 'C' phaseolin in snap beans, and their association with the Andean pool (Brown et al. 1982, Gepts and Bliss 1985). During the intense breeding and improvement which has occurred in this group, recombination barriers 59 between gene pools may have been overcome. The 'C' type of phaseolin shares polypeptides with the 'S' and 'T' types (Brown et al. 1981), and cultivars carrying it may represent an intermediate position between gene pools. A thorough characterization of the barriers separating the major gene pools of g; vulgaris remains to be done. The information produced will clarify phylogenetic relationships in the species, and facilitate recombination. The allozyme associations observed in this study provide a means of identifying specific recombinations of linkage blocks resulting from inter-pool hybridization. Some of these differentiated alleles may mark coadapted gene complexes, and it is possible that certain combinations of linkage groups may be feund to produce novel or transgressive ferms. Breeding and improvement programs have feund that intermating between certain genotypes of beans produces commercially worthless types (eg. Kelly and Adams 1987), and with isozymes it may be possible to identify specific incompatible complexes. Intra- and Inter-specific Relationships As an example of the complexity of genetic structures arising during crop evolution, common bean can be assessed in light of the ideas of Harland and de Wet concerning classification at the infraspecific level in cultivated plants, and the effects of hybridization after divergence (Harlan 1971, Harlan and de Wet 1971). These authors discuss the relationships among the groups which can undergo progressively less successful genetic exchange with a crop's primary gene pool. The evidence fer isolating mechanisms between material originating in Mesoamerican and Andean domestication centers suggests that each of 60 these two sub-specific taxa in beans is a separate primary gene pool, and that each constitutes the other's secondary gene pool. This analysis is consistent with Harlan and de Wet's dictum that crossing between the primary and secondary gene pools of a crop is neither free nor simple, and that recovery of recombinant types in advanced generations is difficult. Restricted exchange may occur between the pools in bean, mediated by such mechanisms as the DL neutral genotype, d1 d1 dl dl . The tertiary gene pool fer each of the primary pools wodld1thin gonsist of other species in Phaseolus which can be crossed to g; vulgaris with varying degrees of difficulty. These include 2; coccineus, which hybridizes reciprocally with g; vulgaris in the wild, but under cultivation appears to have developed a unilateral incompatibility; and E; acutifolius, whose hybrid offspring with g; vuigaris require embryo rescue (Evans 1980). The relationships among wild and domesticated types of g; vulgaris need to be re-examined, but it is suggested here that the cultivated ferms making up each primary pool are more closely related to the wild types of the same gene pool than to the domesticated forms in the alternate primary gene pool. Reports of free recombination between wild and cultivated forms werre made before the dual structure of the species was evident, and comparisons were not made of compatibility of wild x wild, or wild x cultivated, crosses between gene pools. Gentry (1969) reported wild and cultivated g; vulgaris to be conspecific, lacking barriers to crossing, but his opinion was based on careful investigation of only Mesoamerican material. Smartt (1969) saw no evidence of fertility barriers between the primitive and cultivated forms in the species, and on the basis of the distribution of wild forms 61 in both Mexico and Argentina, proposed that large gene pools existed in cultivated g; vulgaris, produced by unrestricted gene flow between all wild and domesticated types (1978). Berglund-Brucher and Brucher (1976) reported numerous successful manipulated recombinations between domesticated g; vulgaris and the wild South American 2; aborigineus. However, the domesticated parents were not named, and it is unclear whether the wild Andeans were found to be equally interfertile with cultivars from both gene pools. Because these authors found no spontaneous hybrids in nature they concluded that there are definite genetic barriers between the wild and domesticated bean in South America. Miranda-Colin and Evans (1973) evaluated reciprocal crosses among all combinations of certain Mexican wild populations and temperate region cultivars of g; vulgaris and g; coccineus. The results suggested that the intraspecific wild and cultivated materials were compatible, but the gene pool source of the cultivars used in this work were not identified. The wild types of the two species were compatible, but unilateral incompatibility was found to occur between wild and cultivated when 2; vulgaris was the pollen parent, suggesting that cultivated types in both species have diverged from the wild. The assessment of g; vulgaris made in the current study departs from a model of free wild/cultivated exchange, on the expectations that wild material from the two pools is separated by isolating mechanisms similar to those operating between domesticated types. Wild/cultivated exchange within each taxa is anticipated to to be relatively free because of the occurrence of divergence before domestication, and the literature appears to support this expectation (Berlund-Brucher and Brucher 1976, Gentry 1969). 62 It may also be hypothesized that each of the two primary gene pools of g; vulgaris has a different complement of taxa in its tertiary pool (32222 Harlan and de Wet 1971), with which gene exchange occurs rarely. Gene flow between the Phaseolus originating in Mesoamerica (eg. 2; acutifelius, g; coccineus) and the Mesoamerican g; vulgaris may be freer than with the Andean. The occurrence of introgression between 2; coccineus and the g; vulgaris races of the Mesoamerican tropics or highlands was suggested by Freytag (1955). A line of wild Mesoamerican 1g; vuggaris has been shown to have isozyme band similarities to a wild E; coccineus as well as to domesticated vulgaris types (Bassiri and Adams 1978). Kaplan (1981) suggests that g; coccineus influenced seed size in Mesoamerican g; vulgaris domesticates, but would not have done so in Andean South America where there is no record of g; coccineus. Thus the Andean and Mesoamerican common bean may differ in their interspecific relationships. It is noted that definite descriptions of such relationships will be greatly facilitated by the use of species- and gene pool-specific isozyme markers. In light of the marked differences between the gene pools of bean, it is proposed that the conclusions from past genetic studies involving the species be re-evaluated from the perspective of the sub-specific taxa. Because many studies have focused only on cultivar or class differences, additional infermation may be produced if results are compared by gene pool categories. For example, Berglund-Brucher and Brucher (1976) recount several instances of disease resistance being transferred from accessions of their wild Andean collections into European and United States cultivars. On the basis of the hypothesis that co-evolution between plant and pathogen has occurred within bean 63 gene pools (Gepts and Bliss 1985), it would be of interest to know whether the Andean germplasm was most successful in conferring resistance on Mesoamerican cultivars or Andean ones, ie. whether the resistance genes mediated against co-adapted or unadapted pathogens. Recent isozyme analysis of the pathogen producing angular leaf spot in beans, Phaeoisariopsis griseola (Sacc.)Ferr., revealed the presence of only two allozyme combinations at four polymorphic loci among A2 isolates from 1A pathogenicity groups. The two differing genotypes were associated initially with place of isolate collection, either in Africa or the New World. However, when the gene pool structure in the host was taken into account, it was found that the pathogen zymotypes were more closely correlated to the Andean or Mesoamerican origin of the cultivars on which the isolates were found (Correa-Victoria 1987). This suggests that the angular leaf spot pathogen has coevolved differentially in each center of origin of the crop. Berglund-Brucher and Brucher (1976) point out that there are differences among bean accessions in phytohemagglutinin (PHA) reaction, and mention that Mesoamerican, Venezuelan [probably small seeded types] and central Andean types produce a strong reaction, in contrast to the weak reaction shown by wild and cultivated types of Chile and Argentina. This does not appear to present a close correlation between PHA type and gene pools, but a re-evaluation of the details of the study may indicate otherwise. The isozyme polymorphism in g; vulgaris is not expected to be entirely associated with gene pool differences. Although these loci which are nearly monomorphic within a pool or sub-group have reduced usefulness as markers within that taxa, additional allelic variants are 6A expected to be found in cultivars and related wild material. These may increase efficiency in genetic marking and mapping by avoiding the crossing barriers between the pools. SGE isozyme analysis is expected to continue to produce fruitful results in the study of common bean and of its evolution as a crop. LIST OF REFERENCES Adams, M. W. 1977. An estimation of homogeneity in crop plants, with special reference to genetic vulnerability in the dry bean, Phaseolus vulgaris L. Euphytica 26:665-679. Adriaanse, A., Klop, W. and Robbers, J, J. E. 1969. Characterisation of Phaseolus vulgaris cultivars by their electrophoretic patterns. J. Allard, R. W., Kahler, A. L. and Weir, B. S. 1972. The effect of selection on esterase allozymes in a barley population. Genetics Bassiri, A. and Adams, M. W. 1978. Evaluation of common bean cultivar relationships by means of isozyme electrophoretic patterns. Euphytica 27:707-720. Berglund-Brucher, O. and Brucher, Heinz. 1976. The South American wild bean (Phaseolus aborigineus Burk.) as ancestor of the common bean. Economic Botany 30:257-272. Brown, J. W. S., Ma, Y., Bliss, F. A., and Hall, T. C. 1981. Genetic variation in the subunits of globulin-1 storage protein of French bean. Theoretical and Applied Genetics 60:251-259. Brown, J. W. S., McTerson, J. R., Bliss, F. A., and Hall, T. C. 1982. Genetic divergence among commercial classes of Phaseolus vulgaris in relation to phaseolin pattern. HortScience 17:752-75A. Conkle, M. T., Hodgskiss, P. D., Nunnally, L. B. and Hunter, S. C. 1982. Starch Gel Electrophoresis of Conifer Seeds: A Laboratory Manual. USDA, Pacific Southwest Forest and Range Exp. Station. Correa-Victoria, F. J. 1987. Pathogenic variation, production of toxic metabolites, and isoenzyme analysis in Phaeoisariopsis griseola (Sacc.) Ferr. Ph.D. dissertation. Michigan State University, East Lansing. Delgado-Salinas, A. 0. 1985. Systematics of the genus Phaseolus (Leguminosae) in North and Central America. Ph.D. dissertation, University of Texas, Austin. Evans, A. M. 1975. Genetic improvement of Phaseolus vulgaris. In Nutritional improvement of food legumes by breeding. PAG Symposium, FAQ, Home (1972) pp. 107-115. Evans, A. M. 1976. Beans. In N. W. Simmonds, Evolution of Crop 65 66 Plants, pp. 168-17A. Longman, London & NY. Evans, A. M. 1980. Structure, variation, evolution, and classification in Phaseolus. DP 337-3A7 in Summerfield, R. J. and Bunting, A. H. (eds.) Advances in Legume Science, Kew. Evans, A. M. and Walters, D. E. 1979. Describing, evaluating and utilizing a germplasm collection of Phaseolus vulgaris beans. pp 127- 132 in Proc. Conf. Broadening Genet. Base Crops, Wageningen. Freytag, George. 1955. Variation of the common bean (Phaseolus vulgaris 2;) in Central America. Ph.D. dissertation, Washington University. Gentry, H. S. 1969. Origin of the common bean, Phaeolus vulgaris. Economic Botany 23:55-69. Gepts, P. and Bliss, F. A. 1985. F1 hybrid weakness in the common bean. J. of Heredity 76:AA7-A50. Gepts, P. and Bliss, F. A. 1986. Phaseolin variability among wild and cultivated common beans (Phaseolus vulgaris) from Colombia. Economic Botany AO:A69-A78. Gepts, P., Osborn, T. C., Rashka, K. and Bliss, F. A. 1986. Phaseolin- protein variability in wild forms and landraces of the common bean (Phaseolus vulgaris): evidence for multiple centers of domestication. Economic Botany AO:A51-A68. Ghaderi, A., Adams, M. W. and Saettler, A. W. 1982. Environmental response patterns in commercial classes of common bean (Phaseolus vulgaris L.). Theor. and Appl. Genet. 63:17-22. Harlan, J. R. 1970 Evolution of cultivated plants. pp 19-32 in Frankel, O. H and Bennett, E. (eds.) Genetic Resources in Plants--Their Exploration and Conservation. Blackwell, Oxford. Harlan, J. R. 1971. Agricultural origins: centers and noncenters. Science 17A:A68-A7A. Harlan, J. R. and de Wet. 1971. Toward a rational classification of cultivated plants. Taxon 20:509-517. Heiser, C. B. 1965. Cultivated plants and cultural diffusion in nuclear America. American Anthropology 67:930-9A9. Kaplan, L. 1965. Archeology and domestication in American Phaseolus (Beans). Economic Botany 19:359-368. Kaplan, L. 1981. What is the origin of the common bean? Economic Botany 35:2A0-25A. Kelly, J. D., Adams, M. W. and Varner, G. V. 1987. Yield stability of determinate and indeterminate dry bean cutivars. Theor. and Appl. 67 Genetics 7A:516-521. Ma, Y. and Bliss, F. A. 1978. Seed proteins of common bean. Crop Science 18:A31-A37. Marques Gomes, M., Rocha Leal, N. and Rodriques Cordeiro, A. 198A. Padroes eletrofereticos em progenitores e linghagens de feijao-de-vagem (Phaseolus vulgaris E;)° Revista Ceres 31:231-237. Martin G. B. 198A. Genetic diversity of bean landraces in northern Malawi. M.S. thesis, Michigan State University, Michigan. Martin, 6. B. and Adams, M. W. 1987a. Landraces of Phaseolus vulgaris (Fabaceae) in northern Malawi. I. Regional variation. Economic Botany A1:190 -203. Martin, G. B. and Adams, M. W. 1987b. Landraces of Phaseolus vulgaris (Fabaceae) in northern Malawi. II. Generation and maintenance of variability. Economic Botany A1:20A-215. Miranda-Colin, S. and Evans, A. M. 1973. Exploring the genetic isolating mechansims between Phaseolus vulgaris L. and g; coccineus Lam. Ann. Rep. Bean Improvement Cooperative 16:39-A1. Mutschler, M. A. and Bliss, F. A. 1980. Genic male sterility in the common bean (Phaseolus vulgaris E;)° J. Amer. Soc. Hort. Sci. 105:202- 205. Nyabyenda, P., Sekanabanga, C. and Nyangurundi, L. 1981. Bean production in Rwanda. pp 99-121 in Potential fer Field Beans in Eastern Africa. CIAT, Cali, Colombia. Rheenen, H. A. van., Muigai, S. G. S., and Kitivo, D. K. 1979. Male sterility in beans (Phaseolus vulgaris L.). Euphytica 28:761-763. Rick, 6. M., Fobes, J. F., and Holle, M. 1977. Genetic variation in Lycopersicon pimpinellifolium: evidence of evolutionary change in mating systems. Plant. Syst. Evo. 127:139-170. Shi, C. T., M. C. Mok, and D. W. Mok. 1981. Deveopmental controls of morphological mutants of Phaseolus vulgaris L.: Differential expression of mutant loci in plant organs. Developmental Genetics 2:279-290. Shi, C. T., S. R. Temple, and D. W. Mok, 1980. Expression of developmental abnormalities in hybrids of Phaseolus vulgaris L.: Interaction between temperature and alleleic dosage. Journal of Heredity 71:219-222. Singh, S. P. and Gutierrez, J. A. 198A. Geographical distribution of the DL1 and DL2 genes causing hybrid dwarfism in Phaseolus vulgaris L., their association with seed size, and their significance to breeding. Euphytica 33:337-3A5. Smartt, J. 1969. Evolution of American Phaseolus beans under 68 domestication. In The Domestication and Exploitation of Plants and Animals. E. by P. J. Ucko and G. W. Dimbleby, pp. A51-A62. London. Smartt, J. 1978. The evolution of pulse crops. Economic Botany Vanderborght, T. 1983. Evaluation of Phaseolus vulgaris wild types and weedy forms. Plant Genetics Research Newsletter 5A:18-2A. Weeden, N. F. 198A. Distinguising among white seeded bean cultivars by means of allozyme genotypes. Euphytica 33:199-208. Weeden, N. F. 1986. Genetic confirmation that the variation in the zymograms of 3 enzme systems is produced by allelic polymorphism. Annual Report of the Bean Improvement Cooperative 29:117-118. Wyatt, J. E. 198A. An indehiscent anther mutant in the common bean. J. Amer. Soc. Hort. Sci. 109:A8A-A87. CHAPTER TWO ISOZYME VARIANTS AT TWO BETA-NADH DIAPHORASE LOCI IN DRY BEANS (Phaseolus vulgaris L.): CORRELATIONS TO GENE POOLS INTRODUCTION The use of isozymes to describe and elucidate evolutionary relationships is most powerful when data for multiple polymorphic loci are available, but it is possible for genetic variation in a single enzyme to reveal significant connections among groups. To utilize isoenzymes effectively in taxonomic characterization it is fundamental to determine which loci produce the variant enzyme bands visualized via electrophoresis. Once bands are identified as alleles of specific loci, then variation may be assessed for taxonomic significance. This paper focuses on one enzyme system of common bean (Phaseolus vulgaris L.) which shows taxonomic significance. Isozymes of beta-NADH Diaphorase were identified as gene products of two loci, and several alleles were found to be associated with specific gene pools of the species. The Diaphorase Enzymes Diaphorase (DIAP) is a general term for a variety of oxidoreductase flavoproteins, present in animals and plants, which can be identified on electrophoretic gels via their ability to oxidize either beta-NADH or beta-NADPH in the presence of certain artificial dyes which act as hydrogen acceptors (Worthington Manual 1968, Anonymous 1987). DIAP which are specific or non-specific for either co-factor are known. In isozyme work it is the specificity of the histochemical activity stain 69 70 used to develop the gel which determines the enzymes visualized. Because the enzymes known as "diaphorases" in most plant and animal electrophoretic work are developed through a reaction with an artificial dye rather than a natural and specific substrate, bands on the same gel may represent enzymes which differ distinctly in their structure and activity $911112, although all are able to react with the dye involved (German 1983; Harris and Hopkinson 1976). Similarly, the same protocol used with-different organisms develops various enzymes of the DIAP group. Electrophoretic studies of the genetics of "diaphorase" isozymes, using essentially the same staining reaction, have reported several enzymes which vary in systematic name, Enzyme Commission (EC) number, and quaternary structure. The "diaphorase” isoenzymes in red blood cells are reported as EC 1.6.2.2, (Harris and Hopkinson 1976), in soybeans as EC 1.6.A.3 (Kiang and German 1983), and in barley as EC 1.6.99.3 (Brown 1983). Thus, the term "diaphorase" is an inspecific one, and does not refer to the catalyzer of a unique enzymatic reaction. However, all the DIAP mentioned belong to a family of NAD(P)H dehydrogenases, EC 1.6.-.-., which comprise oxidoreductases acting upon NADH or NADPH as electron donors. Table 1 lists the enzymes commonly reported in genetic studies of DIAP isozymes. Specific identification is aided if the quaternary structure of the enzyme, that is, the number of subunit polypeptides making up the holoenzyme, is evident from isozyme behavior. The number of isozymes of multimeric enzymes expected in polymorphic material containing varying numbers of loci has been determined (Shaw 196A, Brewer 1970, Dixon and Webb 1979). 71 1 Table 1. Enzyses reported in genetic studies oi diaphorase isoayees . El‘. luster Subunits loses and traits Cofactors and Occurrence £12 1.6.2.2. 1 818: IAONZHerricytochroae g; oxidoreductase 1 FAOIsolecule REC: cytochroee 1:5 reductase asieal Other: 111111112 cytochroee 1:5 reductase erythrocytes setheeoplotin reductase liver 40,000 NW in cali liver Not fluorescent EC 1.6.4.! 2 818: IAON:lipoaaide oxidoreductase 2 FAOIeolecale (in spinach) REC: Oihydrolipoaaide dehydrogenase(llAO: anieal and plant Other: lipoaeide hehydrogenasa 100,000 ill in pig heart ”PM cytochroae greductase 102,000 III in spinach diaphorase Shares hoeolooies with EC 1.6.4.2, glutathione disulphide reductase Highly fluorescent when oxidized, non-fluorescent when reduced EC 1.6.99.1 1 818: Wmucceptorl oxidoreductase FM in yeast (in spinach) Other: MOP" diaphorase FAO in plants NADPH dehydrogenase 'old yellow enzyse' in yeast EC 1.6.99.3 1 818: IAONlacceptorl oxidoreductase anieal and bacteria (Foraerly REC: MON dehydrogenase E1: 1.6.2.11 1 Dixon and lath 19M; Dixon and Iehh 1979; Harris and lioptinson 1976: Vallejos 1983. 72 Two standard protocols for staining DIAP are described in electorphoretic studies. They produce negative or positive zymograms depending on whether the reaction end products are soluble or insoluble, respectively. Both are based on the use of the redox indicator 2,6- dichlorophenol indephenol (DCPIP) which changes from blue to colorless when reduced (Vallejos 1983). Brewer et al. (1967) developed a negative stain using the oxidized form of DCPIP which is reduced by DIAP in the presence of NADH. The original oxidized DCPIP stains the whole gel blue, and the presence of active DIAP is indicated by unstained white areas where the dye is reduced. This staining procedure uses an agar overlay to retard diffusion of the DCPIP and subsequent blurring of the bands. Harris and Hopkinson (1976) described a positive stain protocol in which NADH is oxidized by DIAP, DCPIP is reduced, and reduces MTT (thiazolyl blue; 3-[A,S-Dimethylthiazol-Zyl]-2,5-diphenyl-tetrazolium bromide) to its insoluble form which precipitates as formazan at the position of DIAP in the gel. They note that either stain produces the same isozyme patterns. An alternate method of staining for DIAP uses the defluorescence of NADH, but lacks clarity of the bands (Harris and Hopkinson 1976). This method presumably works only for the DIAP EC 1.6.A.3 which is fluorescent when oxidized, and non-fluorescent when reduced (Table 1). Ninety percent of the proteins found in plastids, mitochondria and microbodies are encoded and produced outside of the compartments themselves (Hanson and Day 1980; Newton, 1983). There is evidence that certain DIAP enzymes belong to the group of gene products which are found in more than one subcellular compartment in the cell, although nuclear encoded and translated in the cytosol. NADH and NADPH 73 dehydrogenases (EC 1.6.99.3 and EC 1.6.99.1) are found in the outer and inner membrane of mitochondria (Hanson and Day 1980); mitochondrially- associated DIAP isozymes (EC 1.6.A.3) are reported in soybean (Kiang and German 1983). The cellular location of the enzyme is important in isozyme analysis because it is commonly found that peptides with different subcellular destinations are produced at different loci, sometimes on oeparate chromosomes, and have distinct catalytic and physical/chemical properties which cause them to be distinguishable as isozyme variants (Dixon and Webb 1979, Weeden 1985). Sub-units of multimeric enzymes destined for different locations in the cell may be so immunologically distinct that they will not interact to form holoenzymes during their common formation in the cytosol, and even after processing and maturation at their appropriate destination, will not interact when brought together in dissociation-reassociation experiments. In instances where subcompartmentally-located isozymes are produced by more than one nuclear gene, the gene products are functionally equivalent and compatible, and subunits can form multimers. In addition, organellar isozymes may differ in cofactor specificity, tending to require NAD in the mitochondria and NADP in the chloroplast (Newton 1983). Examples of isozyme systems with subcellular compartmentalization in animals include MDH, IDH, ME and AAT, (Dixon and Webb 1979). In maize, two cytosolic CAT loci do not interact with a mitochondrial locus 32 XEZE' but will £3 112223 intralocus, but no interlocus, heterodimers occur in AAT, among the three loci expressed separately in glyoxysomes, mitochondria and cytosol; and MDH heterodimers only occur between loci expressed in the same compartment (Goodman and Stuber 1983). In tomato 7A cytosolic and plastid 6-phosphogluconate dehydrogenase isozymes, produced at different loci, have been described (Tanksley and Bernatzky 1987). Weeden and Marx (198A) found that in pea, the nuclear-encoded subunits of the cytosolic and plastid forms of this same enzyme did not interact to form dimers. DIAP Isozyme Analysis Discussions of DIAP isozoyme analysis in plant genetics are not common, and it is unclear whether DIAP has been little used or whether the results have been disappointing and thus gone unreported. In a recent overview of isozyme analysis in plant genetics (Tanksley and Orton 1983) DIAP analysis was mentioned in only two out of 21 major crops. The detailed tomato, maize, and wheat linkage maps do not yet report any DIAP loci among their isozyme markers (Tanksley and Bernatzky 1987; Hart 198A; Coe et al. 198A;). Two unlinked DIAP loci in pea (Pisum sativum) are known (Weeden 1983), but have not yet been associated with any other marker or chromosome, and do not appear on the pea linkage map (Weeden 1985). DIAP has been studied in barley (Hordeum spp.) and the work provides an example of the non-specificity of enzymes termed "diaphorases'. In the initial description of the work Brown et al. (1978; Brown and Munday 1982) reported assaying NADH diaphorase EC 1.6.A.3. Two loci, designated 32229, were defined from patterns which did not show any heteroallelic multimers, suggesting that the enzyme being assayed was a monomer. Both of these loci were found to be polymorphic in wild barley, S; spontaneum, and Nadhd-1 was located to chromosome A of the barley genome. After further work (Brown 1983). the 75 loci were re-defined as genes of the enzyme NADH dehydrogenase, EC 1.6.99.3, and given the symbol Egg. The isozymes were found in both leaf and root tissue, and were resolved by the same activity stain based on reduction of DCPIP used in the original work. While S2221 also had four alleles, S992; had only one in domesticated barley, H; vulgare. Again, no hybrid bands were known in heterozygous material. This evidence of a monomeric enzyme may have been the basis for the redefinition of the enzyme from EC 1.6.A.3, which is described as a dimer, to EC 1.6.99.3, a monomer (Dixon and Webb 1979) (Table 1). In wild and cultivated soybean (Glycine soja Sieb. & Zucc. andSl 955 (L.) Merr.), the presence of 5 to 7 DIAP loci has been hypothesized from electorphoresis of seed cotyledon tissue on polyacrylamide and/or starch gels (German 1983, German and Kiang 1983). The DIAP enzymes were found to be of two types, differing in structure, site of activity and in co-dominant versus dominant expression. Two or three loci produce subunits which migrate to the mitochondria where they interact as tetramers, while three or four genes produce monomeric enzymes which remain in the cytosol. These latter have alleles which can function as dominants (producing a band on a gel) or recessives (producing no band). Band intensity was found to differ with tissue and growth stage, and with DIAP activity level, enzyme production level, and departures from co-dominance. Incompletely dominant weak alleles at a locus caused low intensity of staining; production levels of gene product from other loci changed with development stage. Evidence for the presence of monomeric forms of DIAP came from the lack of hybrid or heteromeric isozymes in heterozygous plants. The presence of an association of five bands suggested the presence of two or more loci producing subunits which 76 interacted as tetramers, although only one locus was defined as such. At least one monomorphic band was found to be present in all material. In spite of the evidence that structurally different DIAP enzymes were present, they were all described as being NADH diaphorase, EC 1.6.A.3 (Broue et al 1977; Kiang and German 1983). Structural differences in enzymes are important in zymogram interpretation. Monomeric, dimeric and tetrameric DIAP are reported, and the latter present complex isozyme patterns for analysis. The tetrameric system which has been most closely studied to date is that of lactate dehydrogenase (LDH) in animals (Dixon and Webb 1979). In each tissue the production of isozymes depends on which of the five loci occur, and are expressed, and on the amount of gene product from each. In many tissues two loci are expressed and five isoenzymes are seen, generated from polypeptides termed A and B. The relative amount of the five isozymes is distributed in the ratio 1:A:6:A:1, which indicates that the associaton of the A and B sub-units is random. In certain tissues other ratios are seen, suggesting an imbalance in sub-unit production, or other occurrence, such as more rapid degradation of less stable hybrid isozymes. In LDH the heteromers have properties intermediate to the homomers, often functioning at the mathematical average of the two parental subunits, although cooperative and synergistic interactions between sub-units may occur in other multimeric enzymes (Dixon and Webb 1979). DIAP analysis has been carried out in common been by Weeden (198A). Assays of leaf tissue from over 90 white seeded cultivars of the snap bean type of Phaseolus vulgaris produced three different multi-banded zymotypes. The positive stain, based on Harris and Hopkinson's formazan 77 precipitate protocol (1976), gave the same results when NADPH was substituted fer NADH in the staining solution, showing that this DIAP was apparently not cofactor specific. Two of the zymotypes found, designated Slow and Fast by Weeden, were made up of four bands, each of which showed a mobility difference between the two patterns, and two monomorphic bands, one of which occurred in the cathodal section of the gel. The third zymotype, Unique, was 5-banded and lacked the cathodal band present in the other two. Of the 93 homogeneous cultivars tested, A1.91 were the Slow pattern, and 57.01 were Fast, a non-significant departure from a 1:1 distribution. The Unique pattern was found only in the cultivar 'Aurora'. Weeden suggested that of the unknown number of loci encoding DIAP in bean, one was polymorphic. Alleles were not defined at that time. Gene Pools £2 Common Bean The genus Phaseolus originated in the New Werld, and contains four cultivated species. The most important agronomically is g; vulgaris, the common bean, which is a diploid annual, predominantly self- pollinated, and adapted to sub-tropical or warm temperate summer climates of the world, where it is grown primarily for its dry seeds or green pods. It varies in many characters, the most conspcicuous being seed traits. There is rapidly accumulating evidence, combining archaeological, morphological and genetic data, that domestication of the common bean took place in two centers, the Mesoamerican and the Andean regions. Several traits are differentiated in cultivated material originating in these two centers, suggesting divergence, 78 probably before domestication. These traits include plant habit (Evans and Davis 1978), seed storage protein [phaseolin] type (Gepts et al. 1986), genes conferring lethal response (Singh and Gutierrez 198A), and seed size (Evans 1976). Based on average differences in seed size, the two groups are often referred to as the small seeded (Mesoamerican) and the large seeded (Andean) gene pools. Specific allelic differences between the pools have been identified through electrophoresis of phaseolin, the major seed storage protein. The variants 'S' and 'B' are found in domesticated types from the Mesoamerican pool, and 'T', 'C', 'H', and 'A' have been identified in cultivated Andean types (Gepts et al. 1986, Gepts and Bliss 1986). Isozyme studies of common bean also show that several enzyme loci are differentiated between the pools, so that certain alleles predominate in populations of each group (see Chapter One). DIAP is one of these gene pool-differentiated enzymes, and the current study shows that it also displays allelic variants specific to sub-groups within the two domestication centers. The DIAP isozyme analysis of this study was initially undertaken to characterize a population of common beans from Malawi, and was subsequently broadened to include other accessions when it became apparent that the DIAP system in beans is complex and genetically informative. The results of this research show that in Phaseolus vulgaris a tetrameric DIAP enzyme is produced by a combination of seven alleles, including two nulls, at two closely linked loci. Certain alleles appear to occur only in specific genetic sub-groups or classes within the species, and can be placed in a putative evolutionary sequence. The number of DIAP alleles is unusually high for bean isozyme 79 systems, suggesting that variants have been retained through direct or indirect selection. DIAP appears to be one of the more informative isozyme gene families reported in common bean. MATERIALS AND METHODS Plant Materials Samples of common bean Phaseolus vulgaris L. from several sources were analyzed with starch gel electrophoresis (SGE). Malawian landrace llggg In 1983, samples of the bean stocks of small farmers were collected in northern Malawi by researchers associated with Bunda College of Agriculture, Malawi, and Michigan State University (MSU), as part of a survey of local bean germplasm. A population of 375 lines was grown in replicated field plots at Bunda College, and measurements made on them there showed that significant genetic variation was present for a range of morphological and phenological characters (Martin and Adams 1987). These lines were brought to Michigan in 198A, increased in the field and greenhouse, and were part of an electrophoretic study carried out to determine genetic variability in g; vulgaris at eight polymorphic enzyme loci (see Chapter One). Two lines were lost during increase due to photoperiod effects. From the remaining 373 lines, 2 to 11 seeds were analyzed for DIAP. Subsequently, the phenological and morphological data collected from Malawian field plantings were analyzed on the basis of the DIAP genotype differences of the lines. Architectural ideotypes The Michigan State University Dry Bean Breeding Program has been involved in producing cultivars with an indeterminate but upright narrow profile, pod placement in the upper two-thirds of the canopy, and increased yield stability (Adams 1982). These defined 80 81 architectural traits have been developed from Central American tropical black bean germplasm, such as 'San Fernando' and its EMS mutant NEP-2, and new occur in several MSU architype navy bean cultivars. The seed size of this navy class is approximately 18 - 20 g/100 seed (Kelly and Adams 1987). Progress in the program to transfer these traits into medium to large seeded commercial classes (pintos, red mexican, great northern, kidneys, etc.) has been slowed due to an apparent genetic association or linkage between the small seed size of the Central American blacks and the suite of genes controlling the desired architecture (Kelly and Adams 1987). DIAP analysis was carried out in architype lines, tropical black cultivars and related breeding lines in the MSU program. The phenotypic recurrent selection scheme used at MSU to transfer upright architecture into pintos has produced five cycles of intermating (Kelly and Adams 1987). From each of these cycles the 10 lines scoring highest and lowest for total desirable architecture were analyzed for DIAP. The architecture scores were formulated by G. Acquaah during the course of doctoral research at MSU (Acquaah 1987). The original intermated population was produced from crosses among 31 parents made in 1980. Remnant seed of cycles 2, 3, and A were extant, and cycles 0 and 1 were reconstituted in 198A from crosses among 17 parents (Kelly and Adams 1987; Acquaah 1987). The lines analysed in this study had been increased to the 8 generation in each cycle. 2 Black seeded material A diverse collection of 130 black seeded beans were received from Centre Internacional de Agricultura Tropical (CIAT) in Cali, Colombia upon the author's request for a selection of large and small seeded black beans from each of the feur growth habit groups 82 recognized in common bean (Type I, determinate; Type II, upright indeterminate; Type III, prostrate indeterminate; and Type IV, climbing; Singh 1982). The CIAT collection of common bean germplasm, which includes cultivars, breeding lines and accessions from around the world, is cataloged by seed color and size. Of the 382A black seeded accessions listed in the current germplasm catalog (CIAT, 198A), 151 lines are identified as large seeded (>AO g per 100 seeds), 919 as medium seeded (>2530 g/100 005017 A u1.59 Brazil Manteigao Preto II 00508” A “7.89 Brazil Sacavem 1009 II 005152 A flu.h8 Brazil Manteigao Preto III 005168 A ”9.39 Brazil I 005308 A “H.30 Brazil Sacavem 627 II 005311 A "1.78 Brazil 1896 EBAL Mouro I 012567 A 60.38 Peru Nuna Azul Grande IV 01fl6u2 A A3.78 Brazil Preto I 015568 A 41.52 Turkey TR 32728 I 015897 A 52.0A Turkey TR 38319 II 98 Table 5a, cont'd. Slow Pattern: Diape1 S/Diap-2 S Small seeded <25 g/100 El Salvador 000086 M 23.81 Tinecon No. 1“ III 000095 M 18.0“ Mexico Negro III 000166 M 13.96 Mexico IV 000276 M 13.“1 Mexico Acotlanero Negro IV 000277 M 1“.85 Mexico IV 000730 M 23.67 Guatemala Omon IV 000732 M 1“.22 Guatemala Kax Omon O'colorado III 0007“7 M 2“.“1 Guatemala III 0007“9 M 12.“8 Guatemala K'OS I 000750 M 13.19 Guatemala Kupal I 000782 M 18.7“ Guatemala III 000791 M 20.93 Guatemala IV 000797 M 21.15 Guatemala III 000826 M 21.22 El Salvador IV 000831 M 20.37 Guatemala II 000899 M 21.30 Guatemala II 000938 M 17.19 Colombia Oaxaca 5-1 IV 001165 M 23.0“ Brazil Preto IV 001168 M 23.22 El Salvador III 001187 M 19.19 Guatemala III 001693 M 2“.93 El Salvador 5-67-N II 001732 M 21.0“ El Salvador III 001960 M 20.85 Guatemala III 002“37 M 23.56 Mexico Bombom II 002963 M 1“.67 Guatemala Cerezo 7 Caldos I 00317“ M 20.“1 Guatemala De Suelo Media Guia I 003530 M 17.96 Mexico Veracruz 155 III 003680 M 23.33 Honduras Honduras 35 II 003878 M 19.“1 Venezuela Sucre 5 II 00“187 M 23.37 El Salvador SAL-219-N III 00““79 M 23.07 Honduras CNA 1215 IV 00“708 M 23.22 Colombia Val 2 = Pirolo I 00527“ M 21.59 U.S.A. No. 22 I 006018 M 1“.07 Guatemala Guatemala 396 I 0060“7 M 20.70 Guatemala Guatemala 506 IV 011900 M 18.““ Colombia “0600 Radical II 01627“ M 23.56 Mexico De Suelo II Medium seeded 25 - “0 g/100 000019 M 31.78 China IV 000080 M 25.11 El Salvador Tinecon No. 5 III 000088 M 28.19 El Salvador Pardo Tineco No. 19 II 000725 M 27.56 Nicaragua IV 000825 M 26.93 Zaire WUlma IV GO10“0 M 37.“8 Mexico III 001251 M 26.50 El Salvador II 601681 M 29.26 Brazil Preto I Table 5a (cont'd.) 002170 M 25.96 002283 M 26.92 002290 M 28.81 002291 M 36.11 002860 M 29.“1 003320 M 29.52 003631 A 26.11 003772 A 28.29 009“60 M 30.81 010506 M 27.67 010557 M 39.63 016077 M 25.89 El Salvador Mexico Mexico Mexico Mexico Mexico Brazil France Guatemala Guatemala Guatemala Mexico Large seeded >“0 g/100 008235 M 61.89 00892“ M “1.52 010580 M “1.93 013870 M “7.19 6 Unique Pattern: Diap-1 S/Diap-2 F Mexico Mexico Guatemala Mexico Small seeded <25 g/100 000012 M 17.37 000110 M 15.56 000261 A 22.33 000860 M 1“.00 001308 M 16.70 001757 M 21.93 00186“ M 18.00 001995 M 22.96 0021“6 M 18.7“ 002588 M 17.56 002676 M 15.56 0030“6 M 18.22 003080 M 20.26 003531 M 18.89 00383“ M 20.81 0038“9 M 21.22 003870 M 20.00 00““5“ M 17.25 005773 M 23.92 011921 M 21.19 0119“8 M 18.70 016“16 M 21.10 016562 M 15.10 Venezuela Ecuador Turkey Mexico Colombia Granada Honduras Guatemala Nicaragua Mexico Mexico Guatemala Guatemala Mexico Costa Rica Venezuela Venezuela Colombia Colombia Mexico Mexico Mexico Zimbabwe 99 Tineco Negro Abolado Aribenyo Abolado Negro Puebla 30-A2 Mogul Super Violeta Guatemala 209 Guatemala 923 Guatemala 997 Huasteco Cafe Guatemala 1000 M7971-1 Abolado Caraotas Negros Negro Redondo Negro P1308913 Mono COL No. 20 COL. No. “86 Negro Piloy Blanco Veracruz 157 51051 Miranda 1 Lara 3 ICA Tui ICA Pijao M 7““3-3-1-1“ M7513ABCD Revuelto Tlanchete Rhodesia 596 II III IV IV IV II IV IV IV II IV III IV III II II III II II II II II IV II II II II II II IV II II II II II II 100 Table 5a (cont'd.) Medium seeded 25 - “0 g/100 00173“ M 25.93 El Salvador Total = 2“ Rare Pattern: Diap-1 I/Diap-Z S Small seeded <25 g/100 011105 M 23.89 Mexico Xmejenbul Negro 011119 M 21.00 Mexico Xmejenbul Negro Opac Medium seeded 25 - “0 g/100 001710 M 25.52 El Salvador 002293 M 26.7“ Mexico Bapalena 011356 M 30.89 Mexico Durango 2-2 Total = 5 Null-1 Pattern: Diap-1 N/Diap-2 S Medium seeded 25 - “0 g/100 001688 A 29.15 Brazil Preto Forro Null-2 Pattern: Diap-1 S/Diap-2 u: Medium seeded 25 - “0 g/100 000262 A 3“.9“ Turkey Gelin Ayse III III III III Table 5a (cont'd.) 101 Mixed or segregating genotypes2 Small 001850 002213 00276“ 003386 005937 seeded <25 g/100 18.26 20.59 2“.78 2“.89 21.22 33:23:! Costa Rica Guatemala Guatemala Mexico Guatemala Medium seeded 25 - “0 g/100 00079“ 0008“0 010301 010593 Total = 26.7“ 32.36 30.22 27.59 39,3: Guatemala Mexico Portugal Guatemala Negro Negro Puebla “17 Guatemala 92 Barrio Pole Bl. Manteigao F+S U+S R/S U+S U/S R+U U/S U+S Guatemala 1013 R/S II III III IV IV IV III IV 1 The phaseolin information is M = Mesoamerican; A = Andean 2 based on electrophoresis of seed extracts. U+S indicates a mixture of the two homozygous zymotypes; U/S indicates heterozygotes present. Table 5b. Association between seed size, DIAP pattern and phaseolin. Phaseolin Andean Mesoamerican NADH Diaphorase Allele Combination Seed size S/S S/F 2 1 59 23 61 2“ 2“.8 19.3 b c 1 For the loci Diap-1 and Diap-2, where F=Fast, S:Slow, I=Intermediate, and N=Null alleles. a,b,c, : Seed size signif. diff., LSD at P<.01 102 architecture characters in order to determine whether recombination has occurred. The 61 lines carrying Slow DIAP were homogeneous for phaseolin type; all but two were Mesoamerican, and seed size was predominantly in the small and medium range. The average size was 2“.8 1 8.76 g/100, which by convention is small seeded. 32.3 S were >25g <“0g, and 6.5% were large, over “0 g/100. The 2“ Unique DIAP lines were distinctly homogeneous. Twenty-three of 2“ lines carried Mesoamerican phaseolin. All were less than 26 g/100, with the average seed size being 19.“9 ‘3 2.91 g/100. This material was collected mainly from Caribbean and Central American countries. Nearly 80$ of the Unique pattern lines were Type 11, defined by CIAT. This plant habit designation does not differentiate between upright narrow profile architypes and other indeterminate bush types. These DIAP/phaseolin associations are summarized in Table 3b. Five of these lines carried the Rare pattern, which had previously been seen only in red mexican types, pinks, and in a single line derived from them. They all carried Mesoamerican phaseolin, were small or medium in seed size and averaged 25.61 1 3.3 g/100. These lines were collected from Mexico or El Salvador, and included matte and shiny blacks, and one black and cream mottled line. Ten small and medium lines were either mixed or segregating for DIAP zymotype. Those with Mesoamerican phaseolin carried combinations of Slow, Unique or Rare DIAP. The line which included Fast and Slow DIAP carried Andean phaseolin. Red mexicans and pinks The pinto cultivar 'JM 126', in which the Rare pattern was first seen, contained NEP-Z and the commercial pink cultivar 103 'Sutter Pink' in its pedigree. The former was known to be DIAP Unique, and the latter was found to carry the Rare pattern. Two other cultivars of the pink class and nine of the related red mexican class were analyzed fer DIAP; seven of these carried the Rare zymotype, three were Slow, and one segregated for Slow/Rare (Table 6). Both these groups are small seeded material which comes from Mexico and Central America. Germplasm from the Andean and Mesoamerican Gene Pools The DIAP patterns and phaseolin isozyme type of 33 lines carrying defined phaseolin and DL genotype are shown in Table 7. A summary by gene pool is in Table 8. In the 21 lines associated with the Andean domestication center the most frequent DIAP pattern was Fast. All four phaseolin type lines were Fast, as were 10 out of 15 d1 d1 DL DL lines. Four lines produced a single banded null activity p;tt;rnf disignated Null-2 (Figure 1). These lines varied in color, shape and size of seed but all originated in Turkey (Table 9). While they carried the DL genotype associated with the large seeded material, the three lines analyzed fer phaseolin by Gepts and Bliss (1985) are combinations of 'S' and 'C', or 'C' alone, rather than the more common 'T'. The lines from the Mesoamerican gene pool carried only Slow and Unique DIAP in the two lines carrying 'B' and 'B' phaseolin and in six DL DL d1 d1 lines. 1 1 2 2 The neutral DL genotypes showed association between DIAP and phaseolin type. 10“ Table 6. Red mexican and pink cultivars analyzed for DIAP. Cultivar/Line Class DIAP Pattern Alleles Diap-1/Diap-2 Big Bend Red mexican Rare I/S GB 760 Red mexican Slow S/S Garnet Red mexican Rare I/S K0228 Red mexican Slow S/S NW 59 Red mexican Rare I/S NW 63 Red mexican Rare I/S Rufus Red mexican Rare I/S Sutter Pink Pink Rare I/S 01 36 Red mexican Rare I/S Victor Pink Slow,Rare S/S,I/S Viva Pink Slow,Rare S/S,I/S Other lines found £2 carry Diap-1 I JM 126 Pinto Rare I/S 01710 Rare I/S 02293 Bapalena Rare 1/S 011105 Xmejenbul Negro Rare 1/8 011119 " " Opaco Rare I/S 011356 Durango 2-2 Rare I/S I3 9 “fl—"3 — n 105 Table 7. DIAP genotype of lines from Mesoamerican and Andean pools. Gene Pool/Line Diap-1/Diap-2 Phaseolin I. Mesoamerican a. Phaseolin type lines Boyaca 22 P1 313590 8 S 8.3 Sanilac S S S b. DL DL d1 d1 1 1 2 2 000278 Callacatlan P1 165“35 S S S 00“017 Carioca P-15“ 8 S - 00““89 Cuilapa 72 P-691 S F S 0071“8 Brazil 668 S S S BAT 332 S S S BAT 1061 S F S c. d1 d1 dl d1 1 1 2 2 002858 Zacaticano PI 319665 S S Mesoamerican 003807 Pico de Oro I-1098 S S " A 30 S F " 005773 ICA Pijao S F " II. Andean a. Phaseolin Type Lines Ayacucho F S A Contender F S C Nuna Huevo de Huanchaco F S H Tendergreen F S T b. dl1dl1DLZDL2 000153 Aysekadin PI 16“930 F S C 000159 Cali Fasulya P1 165078 S N S,C 000568 PI 176712 F S T,C 000623 Barbunya PI 179“21 S N C 000688 Barbunya PI 182268 8 S/N S,C 000910 PI 206983 S N - 00380“ Bolivia 6 I-1095 S S T 005066 BZL-37“ F S T 005129 Sacavem 597 BZL-735 F S T 007160 Tortolas x Diana F S C 00756“ Maestro 1VT-72216 F S T 007613 Aragon F S T 007633 Coco Bicolore du Pape F S T 007635 Coco Rose F S T ICA Linea 23 F S T c. d1 d1 d1 d1 1 1 2 2 002618 Col No 263 PI 313755 S S Andean Diacol Calima F S " 1 Phaseolin data from Gepts and Bliss 1985, or from SGE; DL genotypes from Singh and Gutierrez (198“). 106 Table 8. DIAP pattern and allele frequencies for gene pool lines. DIAP Pattern Allele Gene Pool ------------------------------ Diap-1 Diap-Z Fast Slow Unique Null-2 F S F S N Mesoamerican - .66 .33 - - 1.0 .33 .66 - x = 12 Andean .71 001 - 019 .71 029 " 081 .19 x = 21 Table 9. Characters of lines carrying the Diap-Z S allele. 1 Line Origin Color Phaseolin G159 Cali Fasulya Turkey white flat kidney S,C 0262 Gelin Ayse Turkey white/black bicolor Andean 0623 Barbunya Turkey cream oval striped C 0688 Barbunya Turkey " S,C 0910 PI 206983 Turkey " Andean 1 Phaseolin data from Gepts and Bliss 1985, or from SGE. 107 Populations Segregating for DIAP The two populations involved in segregation ratio tests for DIAP produced progeny which did not depart from the expectation of 1:2:1 proportions for alleles at a single locus (Table 10). It was feund that the cultivar 'Swan Valley' and the Malawian accession 077-1 carried DIAP Unique and Fast, respectively. When remnant F seed from a cross between these lines was tested, eighteen seedlingszwere identified as either hybrid or parental DIAP zymotypes. The 1“ resulting F families were then analyzed for DIAP as separate progenies (Table 13, Population II). The eight lines identified as either Fast or Unique produced 39 progeny which did not deviate from parental types. The six F individuals shown to carry the hybrid DIAP pattern Fast/Uniqu: produced a total of 116 offspring, and each progeny segregated into Fast, Fast/Unique and Unique patterns. No other patterns were seen in these populations. Data fer segregation were pooled for all 1““ offspring of heterozygotes, whether F or F . The chi-square analysis, to test goodness of fit of the pgoled segregation ratio of 39:6“:“1 to 1:2:1, was 1.833, which indicated an insignificant difference from the prediction. During DIAP analysis of the architype pinto selection cycles, lines which contained Slow/Rare and Unique/Rare hybrid patterns were identified and selfed. Five Slow/Rare types produced progenies which each segregated into Slow, Slow/Rare and Rare types, and the offspring of two Unique/Rare hybrids segregated into Unique:Unique/Rare:Rare types (Table 10, Population 1). Families were pooled for chi-square analysis, which showed no significant difference from 1:2:1. Pooling the 108 1 Table 10. Segregation ratios for DIAP alleles at two loci. # of Parental Offspring segregating Total x2 families genotype to each class plants 2 df. Population I. Cycles of recurrent selection of pinto x architype QT'QQEQQQEISQ'SE'QE'EJEEQERSGSEQ """"""""""""" II/SS IS/SS 88/88 5 IS/SS 37 82 28 1“7 3.068 .25>P>.10 3' MMQEIS. Jfi‘iéfigg’éé‘QEQS'égé; """""""""""" II/SS IS/FS SS/FF 2 IS/FS -15 ----------- 28 --------- “-1“-- 57 0.875 .75>P>.50 2T';;;;;;EIS;'SE'§;;:E_§’;SOled .CJQQ‘Z and b. """""""""" II/-- IS/-- SS/-- 7 m. '32"""""??6 """""" 15" 20!: 2.235 .50>P>.25 Population II. 'Swan Valley' x 077-1 5325:5533?E'B_I;;:’E_§3fi;§;§'§§';§§3§§;é§ """""""""""" FF/SS FS/FS SS/FF F2 FS/FS --8-----------12------------8-- 28 .“285 .75>P>.50 F “ FF/SS 18 - - 18 3 “ SS/FF - - 21 21 6 FS/FS 31 52 33 116 .3793 109 Table 10 (cont'd.) b. Segregation of Diap-1 ES/Diap-Z SS in heterozygous F and F FF/SS FS/FS SS/FF FS/FS 39 6“ “1 1““ 1.833 . .25>P>.10 III. Segregation of Diap-1 SS/Diap-Z ES or Diap-1 ES/Diap-Z ES pooled across 1.b. and II.b. --/SS -S/FS SS/FF __ _— — —— — FS or IS/FS 5“ 92 55 201 1.““78 .25>P>.10 1 Where P a probability of getting a larger value of chi-square. 110 progenies which shared the Rare pattern gave a total of 20“ plants, and produced a ratio of Rare:Hybrid:Slow-or-Unique, of 52:110:“2. The chi- square test result of 2.235 shows that the ratio is not significantly different from 1:2:1. Pooling progenies from both tests which shared Unique produced 201 progeny, in the ratio of 5“:92:55, not significantly different from 1:2:1 (Table 10, 111). These results show that the isozyme patterns are segregating as if they were allelic at a single locus. However, to account for all the isozymes in each pattern requires a more elaborate model, examined in detail in the Discusssion section. Test 25 Linkage between DIAP and Architecture Genes It became clear that the Unique DIAP pattern was present at a high frequency in the tropical black germplasm and the architype cultivars and breeding lines derived from it. The possibility that Unique was associated with loci controlling upright architecture was explored by analysing two types of material available in the MSU architype breeding program: isolines of an architype cultivar found to be heterogeneous for DIAP; and samples from cycles of a recurrent selection program for introducing architype traits into pintos. If such an association were present, it would allow indirect selection for recombinations between architecture and seed size traits, which were known to be linked. The recently released MSU upright navy bean cultivar, 'C-20', was found to consist of homozygotes of two different patterns of DIAP, Slow and Unique, in equal proportion. This cultivar has the tropical black 'Jamapa' as well as 'NEP-Z' among its parents, contributing Unique (Figure 2). The source of Slow DIAP may have been 'Seafarer', 'W-20' or 111 'Kentwood', etc. A test of 62 seeds of 'C-20' identified 31 individuals as Slow and 31 as Unique. There were no heterozygotes, suggesting that '0-20' consists of a 1:1 mixture of homozygous DIAP genotypes. If the last single seed selection made during the breeding process were heterozygous, and subsequently both groups of homozygous offspring were retained, it suggests that the DIAP genotype is unrelated to architecture. In order to test an association between DIAP type and architype traits in this cultivar, the 62 tested lines were planted and selfed in the greenhouse. The offspring were bulked to produce two isolines differing for DIAP type, which were planted in the field in 1987 in a randomized complete block pattern as two treatments, with unsorted commercial 'C-20' as a control. During the growing season the plots were inspected for visible morphological differences. The middle two rows of each plot were harvested and yield data were taken. No visible difference in upright architecture, length of indeterminate Vining stem, root size, hypocotyl thickness, growing season, etc. was observed by Dr. James Kelly, the breeder of this cultivar (Kelly, personal communication). No significant differences were feund between the two isolines and the commercial 'C-20' in yield (Table 11). In another test for association between architype characters and Uniqe DIAP, samples from five cycles of recurrent selection for an architype pinto were analyzed. It was expected that if Unique were being selected indirectly along with architecture traits due to linkage association, it would increase in frequency over the cycles in those lines rated highest fer architype traits, even though divergent selection was not taking place, and heterozygosity could be expected to 112 1 Table 11. Yield of DIAP isolines and commercial 'C-20' . Seed yield (lb/a) Commercial 'C-20' Diap-2 S. Diap-Z S 3“13 3193 3387 Mean 3331 LSD NS cv 14.9% 1 Data from J. D. Kelly, 1987 Table 12. DIAP type of lines intermated for recurrent selection. Cultivar/Line DIAP Zymotype Genotype Notes Diap-1/Diap-2 Pinto Agate Slow S/S Amber Slow S/S JM 126 Rare I/S Olathe Slow S/S Ouray Slow S/S 01 111 Slow S/S 01 11“ Slow S/S Upright architype C-20 Unique:Slow S/F:S/S 1:1 mixture Midnight Unique S/F tropical black 7602“-E3 Unique S/F HEP-2 derived 77018-810 Unique S/F " 77020-86 Unique S/F " 77026-036 Slow S/S " 77026-D63 Unique S/F " Other Laker Slow S/S Single stem navy 79-1 Fast F/S 79-8 Fast F/S from Cornell Univ. 791583 Unique F/S Single stem navy 1 Parental lines from Kelly and Adams (1987), Acquaah (1987) 113 persist in the tested S generation of each cycle. From preliminary afialyses it was expected that pinto germplasm would contribute the Slow pattern, and architype navy germplasm the Unique pattern. In addition to these zymotypes, however, Fast and a new DIAP pattern given the name Rare, and all combinations of hybrid patterns, were found among the material sampled from the five cycles. These patterns were traced back to specific contributing parental lines used to constitute Cycle 0 (Table 12). The Rare pattern was found in the parental pinto cultivar 'JM 126'. In this pattern (Figure 1), the four most anodal bands migrate between those of the Fast and Slow patterns while the most cathodal band remains the same as that in Fast and Slow. The presence of two additional zymotypes increased the possibility of heterozygotes, and of less directional frequency changes in DIAP type across selection cycles, but a distinct increase in Unique among high scoring lines summed across cycles was seen (Table 13). Lines scoring low in each cycle on a 1 to 5 scale fer architype traits were evenly divided between Slow and Unique (1“ and 15), but in the high scoring lines the homozygous Unique genotype was almost three times as frequent as the Slow (17 to 6). However, there was not a progressive increase in the frequency of Unique from Cycle 0 to Cycle “. The Slow/Unique heterozygotes were in high frequency in both high and low scoring lines. The low frequency of Fast and Rare in both groups gave no indication of their association with either high or low scoring traits. 11“ Table 13. DIAP in five selection cycles of a pinto x architype cross. 1 DIAP Zymotype S U S/U F F/U F/S R R/U R/F R/S The letters refer to the DIAP patterns Slow, Unique, Fast and Rare. Both patterns are present in heterozygotes, eg. S/U. DISCUSSION The DIAP System $3 Common Bean The complexity of the DIAP isozyme patterns found in beans, particularly in heterozygous material, can be explained by the hypothesis that the major bands are produced as isozymes of a tetrameric form of the enzyme, and that two loci contribute protein subunits which interact in all possible combinations to form the holoenzyme. Shaw (196“) and Brewer (1970) discuss in some detail the structure of isozymes expected to result from multimeric enzymes. Shaw gives a fbrmula to predict the number of isozymes of polymeric enzymes which can be expected when variant subunits are produced by coding differences in the structural genes, when the conditions of diploidy and random combination of subunits are met. (s + p - 1)! i: p! (s-1)! where i = number of isozymes p = polymer number s = number of different subunits Here, p refers to the number of subunits making up the polymeric holoenzyme; 3 refers to the variant polypeptides or subunits present in the individual which can interact to ferm the holomer; and i is the number of isozymes expected to be present in the organism and visible on 115 116 an electrophoretic gel. They can originate at the same or different loci, and correspond to the number of non-identical enzyme alleles occurring in the individual. In the case of an individual carrying two-loci which encode the peptide subunits for a tetrameric holoenzyme, when both loci are homozygous, s = 2, p = “, and i = 5. Five isozymes will be fermed, representing holoenzymes made up of the two subunit variants in the ratios of “:0, 3:1, 2:2, 1:3, and 0:“. When one locus is homozygous and the other heterozygous, there are 3 different alleles or subunits in the pool from which the holoenzyme is assembled. Then p = “ and s = 3, and the expected isozyme number is 15. When both loci are heterozygous, p = “, s = “, and the expected number of isozymes, i, is 35 (Table 1“). It may be noted that in 196“ Shaw reported that he was not aware of the occurrence of a 35-banded zymotype occurring in nature. If a single homozygous locus is present, only one isozyme is produced even when the enzyme is tetrameric. Where subunits do not combine at random, and only form holoenzymes with identical polypeptides or only with subunits in the same subcellular compartment, the number of isozymes and the resulting band patterns will depart from Shaw's predictions. For example, an individual with two homozygous loci whose products do not interact will produce only two bands, rather than five, both made up of homomeric isozymes. In this study of DIAP, material assumed to be homozygous produced four different patterns made up of five bands (Fast, Slow, Unique and Rare), and two made up of one band (Null-1 and Null-2). Lines which were known to result from crosses, or material which segregated, 117 1 Table 1“. Expected isozymes from enzymes of differing subunits. s = no. of different interacting subunits 1 2 3 “ P = DOlymer number ---------------------------- 2 1 3 6 1o 3 1 “ 10 12 “ 1 5 15 35 1 Shaw 196“. 118 produced zymograms made up of more numerous bands. One group of zymograms, which included the hybrid patterns Fast/Slow, Fast/Rare, Slow/Rare, was made up of 15 isozymes. The second group, comprising Fast/Unique and Rare/Unique, displayed in excess of 28 identifiable bands. A complete summary could not be made because of faint and indistinct bands, but it is proposed that this type of zymogram in fact carries 35 bands. Thus, DIAP in beans conforms to Shaw's (196“) predictions for a tetrameric enzyme encoded by two polymorphic loci, and __—-__.. <— 1. . _ ‘ __ . _ each of the DIAP zymotypes seen to date can be explained on this basis. In the proposed model two loci, 233221 and Ségng, produce subunits which interact to form five banded patterns when both loci are homozygous. The isozyme which migrates most anodally in each pattern is presumed to be a homotetramer made up of the polypeptide produced by an allele of 2&32213 the most cathodal isozyme in each pattern is a homomer of a 2&222g allele. The three intermediate bands in each pattern are heterotetramers made up of subunits produced by both loci, combined in the proportions 1:3, 2:2, and 3:1. The active alleles present at each locus have been named according to the mobility under SGE of the homotetramers produced by them, relative to the anodal front. Alleles identifed to date are relatively fast (F), slow (S), and intermediate (1), at Ségp:_, and F and S at Siggzg. When it appears that no active peptide is being produced the allele has been termed a null (N), and one has been feund at each locus. Each DIAP pattern seen to date is a different combination of these alleles. For example, the Fast pattern is made up of five isozymes resulting from the combination of 233221 {_and EEEE:§.§ polypeptides, Null-2 results from Diap-1 S and Diap-2 S alleles (Figure 1). 119 This two-locus, seven-allele model accounts for the banding patterns seen in heterozygous and segregating material. When Fast, Slow, and Rare lines are intermated their hybrid offspring have a DIAP pattern made up of 15 isozymes. These types differ for allele at 232221 (F, S, and 1 respectively) but all carry the same S allele at Slang. A heterozygote resulting from any combination of them carries three different DIAP alleles, two at 232221 and one at Signg, and 15 different combinations of the allelic products are produced. When both loci are heterozygous the pattern is more complex. Unique individuals, 2i22:l.§’2l22:§.§' differ from Fast and Rare at two loci. Thus, Fast or Rare x Unique heterozygotes carry “ different subunits (two alleles at two loci) and the expected number of isozymes is 35. The zymograms seen in samples from such double heterozygotes contain over 28 bands, but it has not been possible to identify 35 isozymes. The single-banded zymotypes, Null-1 and Null-2, also fit the two loci model and lend support to the tetrameric nature of the DIAP enzyme. Production of one band instead of five suggests that only one type of subunit is present and only homotetramers are formed. The Null-1 isozyme co-migrates with the most cathodal band fbund in Fast, Slow and Rare, and therefbre has been designated 33.232221 S/Signg'S. The single band of Null-2 co-migrates with the most anodal band in Slow and Unique, and its genotype presumably is 2$2221.§/EEEE:§.!° The null activity alleles may have arisen in several ways, although none were investigated in this study. Subunit polypeptides may be produced at the locus, but may have been altered in such a way that they lack the ability to form a holoenzyme with each other or with the subunit from the active locus. The null locus may produce no peptide at 120 all, due to changes in the structural gene itself or in regulatory genes. It is possible that the allele is active but has evolVed to produce the identical polypeptide as the allele at the other locus, so that only one homomer type is formed, but this probability is small. There is a report in soybean DIAP of an incompletely dominant weak type of null allele which allows the production of heteromeric bands with a second locus, but not homomers of its own allele (Kiang and Gorman 1983). This situation results in “ isozymes of diminishing intensity on the gel. No hybrids with either Null pattern were identified in common bean, and no evidence fer this type of incomplete activity was observed. There are two monomorphic bands on the DIAP gel which apparently do not interact with subunits produced by 232221 and Signg, but which are present in all analyses. They are visible in root samples but can be most clearly observed in leaf samples and in the two Null patterns when the distinct bands seen in roots are fainter or absent. The first band migrates slightly anodally to 2$2221.5' with Rf of .73, and the second is intermediate to it and the origin, with Rf of .““, and may correspond to the invariant band recorded by Weeden (198“) at approximately Rf 0.3 in his Fast, Slow and Unique types. No closely migrating variants which could be identified as possible polymorphisms of these bands have been seen. They do not seem to be interacting with each other or with the known alleles to form heteromers, as evidenced by the lack of intermediate bands between them in leaf tissue, or in zymograms produced by null alleles. These two isozymes may consist of subunits that are unable to interact with the peptides produced by Siapzl or Signg. They may be ferms of structurally different DIAP enzymes which are developed by the non-specific staining protocol used in this study, such as 7.3 121 monomers or dimers of related enzymes reported from other DIAP analysis in plants (Table 1). These may be enzymes present in subcellular regions different from those where 232221 and Sggng occur. The allelic products may be prevented from interacting initially when immature polypeptides by sequence variation, or subsequently when mature by isolation from each other. The five bands in the DIAP patterns Fast, Slow, Unique and Rare do not migrate equidistantly from each other (Figure 1). In dimeric enzymes it is expected that the heterodimer band will migrate to a position half-way between the two homodimers produced by the contributing loci. This situation has been reported in six dimeric enzymes in maize (Schwartz 196“; Goodman and Stuber 1983). However, some rare alleles of cytosol active maize MDH loci produce heterodimers which migrate asymmetrically, even to the point of lying outside of the bracketing parental homodimers in certain cases (Goodman and Stuber, 1983). It may be that migration rates are not additive in common bean DIAP, because of the combination of two subunits, Signg'S and EEEE:3.§ whose homotetramers have net positive and negative charge, and migrate cathodally and anodally. However, in those plants heterozygous for anodally migrating alleles of 232221: the intralocus interaction bands are intermediate to the parental isozyme positions. It appears that interaction between anodally and cathodally migrating subunits produces non-additive mobility. Evidence fer Linkage between DIAP Loci Once the existence of two DIAP loci is postulated, the question arises of whether the genes are linked. If crossing over were to occur 122 it could only be verified in the offspring of hybrids heterozygous at both DIAP loci, where it would be evidenced by specific segregaton ratios and novel combinations of alleles not feund in the parental types or the F . Such double heterozygote hybrids would only be produced by the pareAtal pairs Fast x Unique, and Unique x Rare, whose F 's would be ‘Sigpzl‘gS/Sggng SS, and Sigpzl £§/2l2223.§§' respectively. 1Intermating any other non-null types would produce F progeny which would contain only parental and F genotypes in a monoiybrid ratio whether the loci were linked or indegendent. If there were no linkage between the DIAP loci, and free recombination of alleles were to occur, in the F of each double heterozygote a di-hybrid ratio of 1:1:2:2:“:2:2:3:1 for nine different zymotypes would result. In that case, 25$ of the individuals would have the same genotype as the double heterozygote seen in the F and produce a zymogram made up of 35 bands; 50$ of the offspring would1be heterozygous at one locus and homozygous at the other, and among them produce a variety of 15-banded zymotypes; and 25$ would be homozygous at both loci and have a simple 5 isozyme pattern. These double homozygotes would consist of the two parental types and two recombinant types. If linkage were complete, a 1:2:1 monohybrid ratio would be produced, of genotypes consisting of the parental homozygotes and F heterozygote. The ratio would be 1:1 for 5-banded to 35-banded zymogiams. These outcomes are summarized in Table 15. Because of the co-dominance of alleles each genotype would be visually distinct in SGE analysis. Although a full linkage study was not conducted, it was possible to estimate the distance between the two DIAP loci from the behavior of the segregating material of the 'Swan Valley' x C77-1 cross, and the 123 Table 15. Isozymes and ratios expected in segregating tetramers. Parents Unique x Fast Diap-1 S/Diap-2 Diap-1 E/Diap-z F2 with free recombinatio segregation ratio: F with complete linkage 2 segregatio ratio: Unique x Rare Diap-1‘S/Diap-2 Diap-1 £lDiap-2 naps Hahn n 25$ 35-band: heterozygous at 2 loci heterozygous at 1 locus 50$ 15-band: homozygous at 1 locus 25$ 5-band: homozygous at 2 loci “:2:2:2:2:1:1:1:1 for 9 zymotypes 25$ 5-band: parental, homozygous at 2 loci 50$ 35-band: heterozygous at 2 loci 25$ 5-band: parental, homozygous at 2 loci 1:2:1 for 3 zymotypes 12“ architype pinto recurrent selection populations. As described in the results from analyses of segregating lines, ratios in both populations were monohybrid, 1:2:1 (Table 13). No 15-banded zymotypes or novel zymotypes were seen. There was no evidence for independent assortment of the DIAP alleles or crossing over between loci, and thus the genes are probably quite tightly linked. The lack of crossing over among the combined 201 progeny of double heterozygotes in these two populations (Table 13, III), suggests that the probability of this occurrence is less than 1 in 201, or P <.005, and thus the interlocus distance can be estimated as being less than 0.5 centimorgans. Not every combination of DIAP alleles is known in beans. Table 16 shows that only six of the 12 possible combinations of the seven known alleles were observed in over 700 accessions examined in this study. This lack of recombination may be due to the close linkage between loci, and to the genetic and geographical barriers to recombination between the gene pools. The genotypes which are combinations of alleles specific to different gene pools, Diap-1 S/Diap-Z E, Diap-1 E/Diap-Z S, and Diap-1 S/Diap-2 S, have not been seen. However, the combination of the alleles from different sub-groups within the Mesoamerican pool, Diap-1 l/Diap-Z S, has also not been seen. Gene Duplication The close linkage between 2&3221 and Signg suggests that one locus results from gene duplication, produced by a mechanism such as unequal crossing over. Isozyme studies indicate that there are gene families fer other enzymes in bean, eg. for peroxidase. In pea there are known to be two DIAP loci, but they are not located on the same chromosome. 125 Table 16. Known combinations of alleles at Diap-1 and Diap-2. Diap-1 Alleles Diap-2 ----------- — _ ------------------ Alleles Fast Slow Intermediate Null Fast Unique Slow Fast Slow Rare Null-1 Null Null-2 126 Gene Pool Specific DIAP Alleles Work on allozyme distribution in the two major gene pools of beans has shown that different alleles at eight enzyme loci predominate in each (Chapter One). The Qggpzl 5 allele is most often present in accessions of the large seeded Andean germplasm, and 2222:1.§ is associated with the majority of small seeded Mesoamerican material. Diap-z S is the most common allele at the second locus in both gene pools. F Lines carrying null alleles have characteristics of both Andean and Mesoamerican germplasm. No conclusions about the distribution of the L. 232221.! allele can be drawn from the single 'Preto Forro' accession carying the Null-1 pattern. Its active allele, 2i22:3.§t is common to both gene pools. The intermediate seed size, 29.15 g/100, Type II habit, and black seed color of this line suggest a relationship to the tropical black pool, but the phaseolin type is Andean. The accessions in which 222222.! is found are distinguished by a common collection site, Turkey, although not all lines from Turkey carry this pattern. Their seed characters are diverse and they do not all belong to the same commercial class (Table 12). The four accessions of the Turkish Barbunya seed type (oval cream seed with brown stripes) which were not all carry Null-2: one was Slow and one was a mixture of Null-2 and Slow genotypes. The three Null-2 lines typed by Gepts and Bliss (1985) for phaseolin are 'C' (Andean, or snap bean type phaseolin) or mixtures of 'S' (Mesoamerican) and 'C'. The one non-Null-Z Barbunya examined is a mixture of 'T' (Andean) and 'C' phaseolin. Seed size, DL genotype and other isozyme alleles (except 0159: see Chapter One) of Null-2 lines place them in the large Andean gene pool, but the fact that their 127 genotype at the 232221 locus is the Mesoamerican allele suggests that the occurrence of 2132:§.§ may be associated with introgression from the latter gene pool. A survey to determine the extent and origin of this 212222.! allele remains to be done. A survey of morphological characters of bean genotypes collected in different parts of the world showed that Turkish lines are differentiated from accessions from several Latin American countries, and group with lines from the U.S., Ethiopia, Argentina and F Chile (Evans and Walters 1979). It is possible that the allele is “ associated with a locally desirable trait in Turkey, such as av physiological response, taste, etc., or may be an artifact of intro- duction or subsequent local breeding. The Mesoamerican sub-groups with which specific DIAP alleles are associated are the tropical blacks (233232 E) and small reds (2132;; S). The presence of the Unique pattern appears to be endemic in tropical blacks. The occurrene of 232221 2 is not as well-defined as RESP £45. The Rare pattern is not found in all reds and pinks, and occurs in lines of other seed colors and patterns. A wider survey will be needed to map its extent in the Mesoamerican pool. Both the tropical reds and blacks have been identified as being genetically differentiated groups within the species in previous studies. Although in each group there is only one DIAP allele change from the Mesoamerican gentoype 2&3221 S/Signg S, and this alone would not be strong evidence for differentiation, the DIAP variation appears to be part of a complex of traits which set these types apart. In 1955 Freytag concluded from scatter diagram analysis of agronomic and morphological measurements that variation in common bean 128 originated from two main sources. One he termed the g; vulgaris complex, comprising elongate seeded beans of determinate plant type. The other was an 'unknown species' designated as Tropical Black, which corresponds to the commercial class of that type. He found the Honduran Red Beans and Black seeded beans of the Caribbean notable in their distinctness. DIAP allele differences supports this grouping. Freytag (1965) subsequently published a classification scheme whereby tropical blacks were placed within the race of common bean as Group C, in the cultivated species 2; vulgari . The black (sometimes white, red or brown) seeded types of Group C were described as being adapted to a warm, humid climate. Gentry (1969) notes that the black seeded wild S; vu aris are limited to the tropical south and east of Central America, and do not occur among the wild types in the northwest. He suggests that the small tropical reds may be a genetic derivative of the tropical blacks. Hernandez X. (in Vieira 1973) differentiates between "upland black" and "lowland black" populations in Central America. 'Puebla 152' is an example of the former (Table “). It is Type III in growth habit, highland adapted, with a shiny seed. It carries EEEE:§.§' unlike the tropical lowland Type II blacks with matte seed which carry Signg S. Singh (unpublished) has formulated a scheme of gene groups in beans based on a wide range of cultivated material, and has defined one group which appears to be co-extensive with the Central American black types, and the upright white seeded material derived from them, which have been found to carry DIAP Unique. Examples of Singh's group include the cultivars 'Jamapa', 'Porrillo', 'Midnight', 'ICA Pijao', and '0-20'. Two cultivars in which Rare was found ('Sutter Pink' and 'Red Mexican') 129 are placed into Singh's medium seed size gene pool 5, but in general Central American small reds are placed into his gene pool 3. Central American tropical blacks share another molecular characteristic of taxonomic significance. Preliminary findings show that in healthy specimens of nine cultivars of the tropical black, or black turtle soup (BTS) class, a significant fraction of the RNA was high molecular weight double stranded RNA not found in the apparently healthy 2; vulgaris cultivars and types 'Red Kidney', 'Top Crop', F i . 'Bountiful', 'Kentucky Wonder Wax', and 'Pinto' (2&2) (Wakarchuk and : Hamilton, 1985). This dsRNA was not homologous to any single stranded 8 RNAs extracted from the same plants, but was homologous to BTS types, and to total DNA of these and 3 non-BTS cultivars. There was no homology to two viruses tested, but similarity to a 3 kilobase pair restriction fragment of Echerischia coli was found. The authors concluded that the dsRNA fragment in the bean was not viral in origin. While this characteristic has not been shown for the whole class, it may be an additional example of a trait which is endemic in this gene pool. 0f the nine BTS cultivars, 'Domino' and 'Midnight' have been analyzed in the current study and shown to carry 2322:17§(2$EE:§ S, the Unique pattern. In Malawi, large seeded types are favored by consumer and growers, and predominate throughout the region (Edje et al. 1982). This correlates with the high frequency of the Andean DIAP allele, PiéE:l.§- It can be assumed that both alleles were introduced to eastern Africa, and that one did not evolve from the other within the past “00 years. Only the two patterns Fast and Slow were seen in the lines sampled from this area. This suggests that germplasm carrying the alleles Diap-1 S, 130 Diap-1 S, Diap-Z S and Diap-Z S may not have been among the original or subsequent bean introductions. Further work is needed to confirm this, but it indicates that the Central American black and red germplasm, and the gene pool associated with Diap-Z S may be used in eastern Africa to widen the local genetic base and contribute specific desirable traits. It is interesting to note that the lack of the DIAP pattern Unique, associated wtih in tropical black bean varieties, in light of a frequently stated dislike of black beans by consumers in Malawi (Edje et al. 1981). This germplasm may be used to contribute such traits as yield stability, disease resistance and architecture (Kelly and Adams 1987) without transferring black seed coat color. Divergence 25 DIAP Variants The fact that certain alleles occur in specific gene pools poses the question of how they diverged within the species. It is possible to suggest an evolutionary sequence. At the DIAP-1 locus 2i22:§.§ is the most common allele in both gene pools, and may be the original progenitor from which the duplicate locus and variants arose. The Slow pattern (222231 S/Qigng‘S) appears to be the ancestral type within the Mesoamerican gene pool. The Mesoamerican Rare (Sgggzlel/Qigng‘S) and Unique (giggzl SIQigng‘F) differ from each other in alleles at both loci, but each differ from Slow by only one allele at one locus, and probably both were derived from it. It has been suggested that there is a closer relationship between the tropcial blacks and reds and g; coccineus (Freytag 1955), and this may mean that Eléflzl I or Sggp gig alleles were introduced from related species, but this remains to be investigated. 131 Null-1 (Diap-1 S/Diap-Z S) could be derived by loss of DIAP-1 activity from either the Mesoamerican Slow or the Andean Fast. Its intermediate nature allows little speculation about its origin. Null-2 (Qigpzl §/gggg;g S) is also anomalous. It carries the Mesoamerican allele as 2&22:l.§' yet is found in accessions with phaseolin, isozyme and seed size traits which place them closer to the Andean pool. Future Investigations pi Several directions for fUrther research are suggested by the DIAP E system in common bean. A determination of the distribution of each of II the variant DIAP alleles would reveal the extent to which they represent gene pool and sub-group specific alleles, and may clarify the relationships of these groups within the species. Linkage between Type 11 architype growth habit and the DIAP loci remains to be tested accurately. Two things would need to be determined in order for any such a linkage to be used in an indirect selection program fer architecture. First, how close the linkage is, and secondly, whether DIAP is linked to the whole suite of genes producing the architype, or only to a portion. The present survey, limited to about 700 lines, has not revealed all combinations of 232221 and Signg alleles. Producing all recombinations of DIAP alleles may answer some questions about the compatibility of allelic variants developed in different gene pool backgrounds, and may give rise to novel types. Not all combinations may be viable, particularly the one containing both Null alleles. Nulls at these two loci would leave only the monomorphic bands 'A' and 'B', and indistinct bands which are seen from leaf tissue. These latter 132 consistently appear to be less concentrated than root isozymes. It may be that the loss of these isozymes will not affect leaf physiology, but may affect root activity. On the basis of the high amount of variability already seen in the DIAP system, new alleles and new combinations may be expected to be found in surveys of wild material. The different levels of DIAP production noted in separate tissues indicates that it would be possible to describe a developmental sequence of DIAP induction, based on the lack of DIAP in bean seed cotyledons, its subsequent appearance in leaf, and its full production in roots. A developmental profile of the activation of this enzyme in each plant part of the normal, the single null, and the double null genotypes may reveal information about the activity of the enzyme. Infbrmation about the sub-cellular location where 235221 and EEEEZE holoenzymes are active would also aid in characterizing the enzyme. The use of the fluorescent stain protocol described by Harris and Hopkinson (1976) could identify which DIAP is being stained. If the DIAP is EC 1.6.“.3, this stain should be successful, and presumably other forms of DIAP would not be activated. The bean DIAP system could then be more accurately compared to the DIAP loci in soybean and pea. An investigation of the amount of enzyme product in each band would confirm whether there is in fact equal production of heteromers and random interaction of subunits. Investigations using_DIAP null alleles Numerous null alleles are known from isozyme work. In maize, null activity alleles have been found or induced at ten isozyme loci (Goodman and Stuber 1983), and the range of their expression includes normal plant function in the homozygous state, 133 disfunction of pollen carrying null genotype, the conferring of sensitivity to low night temperature, and "activity nulls" which lack activity only in the homomeric state. Malic dehydrogenase (MDH) nulls have been used to show that some cytosolic MDH's, once thought to be part of the malate shuttle, are dispensable, but that at least one active mitochondrial MDH is required fer normal development. The use of null activity example is the use of alcohol dehydrogenase (ADH) null isozyme alleles in maize have produced an understanding of the role of each during development under normal conditions and under anaerobic ‘ ‘ a- ‘-”‘l'~ I: 'l stress (Freeling 1983). Nulls can be naturally present or induced in a mutagenic process, such as ADH in barley (Brown 1983). In S; cheesmanii, a wild relative of the tomato, nulls constitute 62$ of the isozyme polymorphisms found in this highly invariant species (Rick 1983). Quiros (1983) points out that in alfalfa a null allele at a PRX locus may be due to deletion of the pertinent chromosome segment rather than simply a lack of activity of the gene itself. Endo and Morishima (1983), discussing rice isozymes, differentiate between temperature sensitive variants which either fail to produce enzyme or produce isozymes which are inactive at certain temperatures, and true inactive nulls which never show enzyme activity. In general, nulls allow exploration of the degree to which an enzyme ferm is dispensable in the plant. If multiple loci are present, as in the DIAP gene family, the lack of activity at one locus may not cause any change in metabolism if its product is compensated fer by other loci. Null activity gene products destined fer different sub- cellular compartments may affect the whole cell. 13“ Kiang and Gorman report null DIAP alleles in soybean (1983). At one locus the null appears to be similar to the activity null of maize MDH, because it produces no homotetramer band and progressively weaker heterotetramer bands as its dosage in the holoenzyme increases. At the other DIAP locus a complete null is known. The series of five progressively shorter and weaker bands is known in bean DIAP, most clearly seen above the most cathodal band in the anodal slice, and this may be caused by interaction with subunits from loci other than 232221 and 2322:2- The identification of null alleles allows certain investigations. In dry beans, with nulls at two loci, these are as follows. By putting both Diap1-N and 212223.! in the same line it will be possible to test whether absence of active product at these two loci causes lethality. Both null alleles are found in viable cultivars, and this suggests that absence of DIAP activity at one locus is compensated fer by gene product from another. However, lack of activity at both these loci may incapacitate the plant by the complete absence of the specific type of DIAP enzyme produced by these loci, or by the absence of DIAP within one particular cellular compartment. By examination of a double null plant, and the stage and conditions at which lethality may occur, it may be possible to characterize the role of these two DIAP loci in bean development. This would be especially useful because DIAP is not a substrate-specific enzyme, and it is difficult to predict where its activity is important in the plant. Once a line is produced with both nulls, the association of the remaining isozymes with specific subcellular compartment may be easier. The two monomorphic bands 'A' and 'B', which do not appear to take part 135 in heterotetramer formation, may be present in a different sub-cellular compartment from that in which 213221 and Sggng occur, or may be a different form of DIAP. Electrophoresis of cell fractions may show which isozymes are associated with the cytosol, and which with the subcellular compartments such as the mitochondria. The polymorphism present in the DIAP system thus provides a mechanism for investigation of taxonomic and developmental questions. LIST OF REFERENCES Acquaah, G. 1987. Genetic and multivariate statistical evaluation of a phenotypic recurrent selection program for recombining erect architecture and large seed size in Phaseolus vulgaris L. Ph. D. dissertation. Michigan State University, East Lansing. Adams, M. W. 1982. Plant architecture and yield breeding in Phaseolus vulgaris. Iowa State J. Res. 56:225-25“. Anonymous. 1987. Biochemical and organic compounds for research. Sigma Chemical Co., St. Louis. Brewer, G. J. 1970. An Introduction to Isoenzyme Technique. Academic Press, New York. Brewer, G. J., Eaton, J. W., Knutsen, C. A., and Beck, 0. C. 1967. A starch gel electrophoretic method for the study of diaphorase isozymes and preliminary results with sheep and human erythrocytes. Biochem. Biophys. Res. Commun. 29:198. Broue, P., D. R. Marshall and W. J. Muller. 1977. Biosystematics of subgenus Glycine (Verdc.): isoenzymatic data. Australian Journal of Botany 25:555-66. Brown, A. H. D. 1983. Barley. pp 55-78 in Tanksley, S. D. and Orton, T. J. (eds.) Isozymes in Plant Genetics and Breeding, Part B. Elsevier, Amsterdam. Brown, A. H. D. and Munday, J. 1982. Population-genetic structure and optimal sampling of land races of barley from Iran. Genetica 58:85-96. Brown, A. H. D., Nevo, E., Zohary, D. and Dagan, O. 1978. Genetic variation in natural populations of wild barley (Hordeum spontaneum). Genetica “9:97-108. Coe, E. H. Jr., Soisington, D. A., and Neuffer, M. G. 198“. Zea mays (corn). pp “91-507 in Genetic Maps 198“. S. J. O'Brien, ed. Cold Spring Harbor Laboratory. Coyne, D. P. 1965. A genetic study of 'crippled' morphology resembling virus symptoms in Phaseolus vulgaris L. J. Heredity 56:162-176. Davis, D. H. and Frazier, w. A. 196“. The incidence of three abnormalities in F2 progeny of crosses between true bushes and Blue Lake derived bush snap beans. Ann. Rept. Bean Improvement Cooperative 7:1“- 16. 136 137 Dixon, Malcolm and Webb, E. C. 196“. Enzymes. Academic Press, N.Y. Dixon, M., and Webb, E. C. 1979. Enzymes. Third edition. Academic Press, New York. Edje, O. T., Mughogho, L. K., Rao, Y. P., and Msuku, W. A. B. 1981. Bean production in Malawi. PP 5“-97 In Potential for Field Beans in Eastern Africa. Cent. Intern. Agric. Tropical, Cali, Colombia. Endo, T. and Morishma, H. 1983 Rice. pp 129-1“6 in Tanksley, S. D. and Orton, T. J. (eds.) Isozymes in Plant Genetics and Breeding, Part B. Elsevier, Amsterdam. Evans, A. M. 1976. Beans. in N. W. Simmonds, ed. Evolution of Crop Plants, pp. 168-172. Longman, London. Evans, A. M. and Davis, J. H. C. 1978. Breeding Phaseolus beans as grain legumes fer Britain. Applied Biology 3:1-“2 Evans, A. M. and Walters, D. E. 1979. Describing, evaluating and utilizing a germplasm collection of Phaseolus vulggaris beans. pp 127- 132 In Proc. Cong. Broadening Genet. Base Crops, Wageningen, 1978. PUDOC, Wageningen. Freeling, M. 1983. Isozyme systems to study gene regulation during development: a lecture. In Isozymes in Plant Genetics and Bredding, Part A. S. D. Tanksley and T. J. Orton, editors. Elsevier Science Publishers B.V., Amsterdam. Freytag, George F. 1955. Variation of the common bean (Phaseolus vulgaris L.) in Central America. Ph.D. Dissertation. Washington University, St. Louis, Missouri. Freytag, George F. 1965. Clasificacion del frijol comun: (Phaseolus vulgaris L. y especies afines). CEIBA 11:51-6“. Gentry, H. S. 1969. Origin of the common bean, PHaseolus vulgaris. Economic Botany 23:55-69. Gepts, P. and Bliss, F. A. 1985. F1 hybrid weakness in the common bean. F. Heredity 76:““7-“50. Gepts, P. and Bliss, F. A. 1986. Phaseolin variability among wild and cultivated common beans (Phaseolus vulgaris) from Colombia. Economic Botany “0:“69-“78. Gepts, P. Osborn, T. C., Rashka, R., and Bliss, F. A. 1986 Phaseolin- protein variability in wild ferms and landraces of the common bean (Phaseolus vulgaris): evidence for multiple centers of domestication. Economic Botany “0:“51-“68. Goodman, M. M. and Stuber, C. W. 1983. Maize. pp 1-3“ in Tanksley, S. D. and Orton, T. J. (eds.) Isozymes in Plant Genetics and Breeding, Part B. Elsevier, Amsterdam. “I. -, r I. ) 138 Gorman, M. B. 1983. An electrophoretic analysis of the genetic variation in the wild and cultivated soybean germplasm. Ph.D. dissertation. University of New Hampshire, Durham. Hanson, J. B. and Day, D. A. Plant mitochondria. pp 315-358 in The Biochemistry of Plants, Vol. 1. Academic Press, New York. Harris, H. and Hopkinson, D. A. 1976. Handbook of enzyme electrophoresis in human genetics. North-Holland/American Elsevier, Amsterdam/New York. Hart, 0. E. 198“. Triticum aestivum (wheat-biochemical). pp “85-“90 in Genetic Maps 198“. S. J. O'Brien, ed. Cold Spring Harbor Laboratory. Kelly, J. D. and Adams, M. W. 1987. Phenotypic recurrent selection in ideotype breeding of pinto beans. Euphytica 36:69-80. Kiang, Y. T. and Gorman, M. B. 1983. Soybean. pp 295-328 in Isozymes in Plant Genetics and Breeding, Part B. Elsevier, Amsterdam. Ma, Y. and Bliss, F. A. 1978. Seed proteins of common bean. Crop Science 18:“31-“37. Martin, 0. B. 198“. Genetic diversity of bean landraces in northern Malawi. M. S. thesis. Michigan State University, East Lansing. Martin, G. B. and Adams, M. W. 1987. Landraces of Phaseolus vulgaris (Fabaceae) in northern Malawi. 1. Regional variation. Economic Botany “1:190-203. Newton, K. J. 1983. Genetics of mitochondrial isozymes. pp 159-176 in Tanksley, S. D. and Orton, T. J. (eds.) Isozymes in Plant Genetics and Breeding, Part A. Elsevier, Amsterdam. Quiros, C. F. 1983. Alfalfa, luzerne. pp 253-29“ in Tanksley, S. D. and Orton, T. J. (eds.) Isozymes in Plant Genetics and Breeding, Part B. Elsevier, Amsterdam. Rick, 0. M. 1983. Tomato. pp 1“7-166 in Tanksley, S. D. and Orton, T. J. (eds.) Isozymes in Plant Genetics and Breeding, Part B. Elsevier, Amsterdam. Schwartz, D. 196“. A second hybrid enzyme in maize. Proc. Natl. Acad. Shaw, C. R. 196“. The use of genetic variation in the analysis of isozyme structure. pp 117-129 in Subunit Structure of Proteins: Biochemical and Genetic Aspects. Brookhaven National Laboratory, Upton. Shi, C. T., Mok, M. C. and Mok, D. W. 1981. Developmental controls of morphological mutants of Phaseolus vulgaris L.: differential expression of mutant loci in plant organs. Developmental Genetics 2:279-290. 139 Shi, C. T., Temple, S. R. and Mok, D. W. 1980. Expresson of developmental abnormalities in hybrids of Phaseolus vulgaris L.: interaction between temperature and allelic dosage. J. Heredity 71:219- 222. Singh, S. P. and Gutierrez, J. A. 198“. Geographical distribution of the DL1 and DL2 genes causing hybrid dwarfism in Phaseolus bulgaris L., their association with seed size, and their significance to breeding. Euphytica 33:337-3“5. Tanksley, S. D. and Bernatzky, R. 1987. Molecular markers for the nuclear genome of tomato. pp 37-““ in D. J. Nevins and R. A. Jones (eds.) Tomato Biotechnology. Alan R. Liss, New York. Tanksley, S. D. and Orton, T. J. 1983. (eds.) Isozymes in Plant Genetics and Breeding, Part B. Elsevier, Amsterdam. Tanksley, S. D. and Rick, 0. M. 1980. Isozymic gene linkage map of the tomato: applications in genetics and plant breeding. Theor. Appl. Genet. 57:161-170. Tohme, J. M. 1986. Relationships between morphological and physio- logical characterisitcs and yield of dry bean (Phaseolus vulgaris L.) cultivars differing in their plant architecture. Ph.D. dissertation. Michigan State University, East Lansing. Vallejos, C. E. 1983. Enzyme activity staining. pp “69-516 in Tanksley, S. D. and Orton, T. J (eds.) Isozymes in Plant Genetics and Breeding, Part A. Elsevier, Amsterdam. Vieira, C. 1973. Plant introduction and germplasm of Phaseolus vulgaris and other food legumes. pp 239-258 In Potentials of field beans and other fecd legumes in Latin America. Cent. Intern. Adric. Tropical, Cali, Colombia. Wakarchuk, D. A. and Hamilton, R. I. 1985. Cellular double-stranded RNA in Phaseolus vilgaris. Plant Molecular Biology 5:55-63. Weeden, N. F. 1983. Isozyme variation at selected loci in Pisum. Pisum Newsletter 15:58-59. Weeden, N. F. 198“. Distinguishing among white seeded bean cultivars by means of allozyme genotypes. Euphytica 33:199-208. Weeden, N. F. 1985. An isozyme linkage map for Pisum sativum. pp 55- 66 in The Pea Crop: A Basis for Improvement. Hebblethwaite, P. D., Heath, M. C. and Dawkins, T. C. K., eds. Butterworths, London. Weeden, N. F. and Emma, A. C. no date. Horizontal starch gel electrophoresis laboratory procedures. New York State Agricultural Experiment Station, Geneva. Weeden, N. F. and Marx, G. A. 198“. Chromosomal locations of twelve isozyme loci in Pisum sativum. J. of Heredity 75:365-370. mo Worthington Manual. 1968. Enzymes. Worthington Biochemical Corporation. York, D. W. and Dickson, M. H. 1975. Segregation of a semi-lethal or crippled condition from crosses involving P.I. 165“35. Ann. Rept. Bean Improvement Cooperative 18:88-89. P‘s-u.“ .- tum CHAPTER THREE THE MAINTENANCE OF SEED CLASS IDEOTYPES IN VARIETAL MIXTURES IN TRADITIONAL AGRICULTURE IN MALAWI INTRODUCTION The planting of combinations of diverse seed types, termed varietal mixtures, are a feature of common bean (Phaseolus vulgaris L.) production under traditional agriculture in several parts of the world. Mixtures of climbing beans are planted with maize in the highland regions of both Mesoamerica and the Andes, and there is evidence in Mexico that specific mixtures of seed class components have been in use for as long as 1,000 years (Kaplan 1981). The wide range of seed type diversity among the dry bean mixtures grown by small-scale farmers in African countries such as Malawi, Tanzania, Kenya, Rwanda, Zaire, and Uganda has been described in relation to breeding and development programs (Rheenen 1979. Anon., 198“, Lamb and Hardman n. d., Ferguson 1987). The array of seed types present in African landraces suggests that they are a valuable accumulation of locally adapted and variable germplasm. The use of intraspecific varietal mixtures in beans is an example of the use of diversity as an agronomic strategy. Marshall (1977) noted that genotypic diversity within and between traditional crop varieties developed in response to, and was maintained by, diversity of cultural and environmental regimes. Allard (1961) associated the superiority of complex lima bean (S; lunatus) hybrids over simple mixtures not with heterosis but with the ability of variant genotypes to exploit 1“1 1“2 microsites. He suggested that the mixing of pure line components of a self-pollinated crop in traditional agriculture mimics a complex hybrid population by providing diverse genotypes able to respond differentially to climatic variation, and concluded, from tests of two- and three- component populations, that mixtures confer stability of performance irrespective of the number of components involved. A summary of performance tests of multilines and varietal mixtures (Marshall 1977) indicated that they cannot be expected to increase in yield or yield stability over the best component, but the study did not define changes in disease avoidance. In searching for a reason for the use of bean varietal mixtures more precise than the farmers' assertion that "they grow better that way", Kaplan focused on variation in seed germination rates as the adaptive factor (1981). Presumable, however, it is variation for numerous quantitative traits, in addition to germination, which allows a mixed planting to exploit the phenological and ecological environments presented by non-unifbrm plots of land over many seasons. Kaplan (1981) suggested that growers did not originally plant bean mixtures, and did not produce the initial components, but assembled already diversely adapted types into varietal mixtures broadly suited to local conditions. Yet the use of varietal mixtures in traditional agriculture is not entirely explained by agronomic considerations. Small-scale farmers are interested in stability of production, but also in the maintenance of diverse types to fill their nutritional and economic needs. The operation of human selection places the discussion of variation in food crops into a unique realm of plant population genetics. Harlan (1975) points out that human decisions about crop phenotypes superimposes a 1“3 selection pressure which can be "intense and absolute and is often biologically capricious or even whimsical", and that variation patterns of crops are artifacts of human activity. ...most cultivators in what we call "primitive agriculture' are very particular about the seed they sow...The population becomes an array of deliberately chosen components. It may still be rich in variation because cultivators of traditional agriculture have an appreciation for mixtures, but the mixtures will conform to whatever an individual selector chooses. The total potential range of variation will be fragmented into landrace populations or primitive cultivars. Different cultivars will be grown for different purposes or to fit different ecological niches of the agricultural system. The question of how much the bean mixtures in Malawi are affected by natural selection, and how much they are a result of farmer choice, conscious or unconscious, has been discussed (Martin and Adams 1987b, Ferguson 1987). There are several lines of evidence which point to the dominant influence of human intervention on the composition of on-farm bean stocks in Malawi. These focus on the range of seed size and growth habit, and on the maintenance of specific seed types. Intraspecific mixtures of predominantly self pollinating crops, if uninfluenced by human selection, tend to change their composition over successive plantings until they are dominated by the most competitive of the locally adapted genotypes. For seed-propagated crops sown annually on the basis of seed weight or volume, higher yielding smaller seeded lines are able to dominate (Allard 1960). While intermating between lines may broaden the genetic base of the genotypes present, and heterozygosity may persist for long periods (Allard et al. 1972), they do not reverse the selection for competitive traits. These predictions have been tested in bean mixtures under natural selection, with consistent results. Vieira (1975) showed that in bean mixtures planted at a density of 1““ 250,000/ha, small seeded indeterminate cultivars out-competed large seeded determinate ones in a few generations, and all but eliminated them from the mixture. Bean cultivars with the highest seed yield per plant dominated mixtures after three seasons in the field, and after 15 computer simulated cycles (Roos 198“), in a close approximation to the component changes seen by Harlan and Martini (1938) in barley mixtures. Dessert (1987) pointed out the influence of seed size on varietal mixtures of beans planted for six successive seasons in Rwanda without human selection. In experiments replicated at three sites, those lines with indeterminate growth habit and small seed size (21 g/100 seed and 28 g/100), increased in proportion in the mixture, from 10$ each to 28.6$ and “8.8$. Tests of navy bean mixtures, composed of lines with similar seed size (approx. 25 g/100), in Michigan, showed that the long- season components dominated the early-season types, the vine types dominated the bush, and that types combining lateness and vine habit were most competitive (Adams pers. comm.). Ayeh (in preparation) found in similar experiments in Malawi that four traits predominate in bean mixtures over time: small seed size, climbing (vining) growth habit, late maturity, and fecundity. The positive correlation in beans between the yield components of small seed size and number of seed contributes to these changes. Martin and Adams (1987a) were aware of the anomalous situation in Malawi, where diversity for growth habit and seed type is maintained at farm sites across the northern region, in the face of natural selection against them. They suggested that lack of intraspecific competition in the field allowed variation in growth habit to be maintained. In a mixed population where outcrossing takes place, heterogeneity 1“5 for characters not under selection may increase. In a lima bean mixture of two parental types undergoing 5$ outcrossing without human selection, Allard et. al (1966) found that a hybrid swarm predominated by the fourth generation, and that even the competitive parental line was reduced to 10$ of the population by the 12th generation. Natural outcrossing, at a rate of 1$, and seed handling practices, have been identified as significant factors in generating and maintaining variability among Malawian bean landraces (Martin and Adams 1987a). It can be hypothesized that there is another element of selection functioning in the system, grower choice, which is capable of overcoming or enhancing natural selection for seed size, growth habit, and seed type. The maintenance by farmers of multiple types in local and regional populations is able to provide a genetically variable substrate on which outcrossing and further selection may continue to act, generating additional variability. Detailed information has been collected about the characters for which farmers maintain a selection pressure (Martin and Adams 1987a, 1987b, Ferguson 1987), and the varied decisions which influence choice of planting stock in each season (Ferguson 1987, Ferguson and Sprecher 1987). The variation found in Malawian bean mixtures can be attributed in large measure to selection by growers. This intervention may be defined as the maintenance of multiple ideotypes via mass selection. Ideotype selection is suggested as a model for several reasons. Bean seeds are a comparatively large planting unit (12 - 65 g/100 seed) and amenable to individualized handling and decision-making. Bean seeds are both the plant part used and planted, and the variability present in Malawi allows memorable color/pattern combinations to be generated. The 1“6 distinct seed types (comparable to commercial classes in modern agriculture) in this predominantly self-pollinated crop allow information about the performance of a genotype to be associated with a phenotype which is selectable at planting time. Common beans fit Wright's case of ”unrestricted component selection" (1983) in compositing.mixtures, because, although they may not be separately harvested or stored, the seed type identification allows separate assessment and manipulation of the components of the mixture. The close contact among bean growers/preparers/consumers in Malawi (often the same individual), allows knowledge about the agronomic and culinary properties of seed types to accumulate and to influence selection. The situation is comparable to the mass selection in early corn breeding, where each ear of corn was handled by the farmer who saw the plants and was able to select on the basis of close knowledge (Poehlman 1979). This situation allows the genetic association between seed appearance and plant/seed traits to become closer as such selection is applied over time. From observations and information collected in Malawi, two factors considered important by growers in identifying various bean ideotypes are seed appearance (size and pattern), and plant growth habit. Expectations of the perfbrmance of ideotypes are both agronomic (vigor, yield, drought resistance), and socio-economic (palatibility, cooking behavior, marketability) (Martin and Adams 1987b, Ferguson 1987). The concept of crop ideotypes as an influence in modern breeding and improvement was recognized by Donald (1967), who defined growth habit and phenological characters which could be expected to contribute to increased production levels of wheat under conditions of mechanized 1“7 farming. Adams (1973) developed an ideotype for dry bean production in humid temperate regions, combining a specific arrangement of branches and reproductive parts with indeterminate growth habit, and produced cultivars incorporating the desirable plant architecture. These were successfully used in the intensive production systems for which they had been planned. In traditional agriculture, several ideotypes may be needed to suit different cropping regimes and seasonal climates, and to fulfill diverse culinary requirements. Types will be planted on the basis of how well the farmer expects them, from past experience, to fill his or her needs: and to a certain extent they will be re-selected in proportion to how successfully they came up to expectation, and approached the conceptualized ideal. Named bean types in Malawi are most readily defined by their seed appearance and growth habit, but additional expectations of their performance are associated with many of them. Thus, an example of one common ideotype is the large red kidney bush bean which matures in about 90 days, cooks quickly and is so flavorful that it tastes good even without salt (Ferguson 1987). If seed planted as such should fail to live up to its ideotypic description, the fitness value of the genotype would change. The maintenance of ideotypes in Malawi will be explored in this chapter by examining variation found within seed classes and across farms. Isozyme analysis of 373 bean lines in a sample from northern Malawi produced information about the major gene pools of the crop (Chapter One), but left within-pool relationships mostly unexplored. Some insight into the crop germplasm is possible by combining isozyme data with morphological and phenological measurements. What emerges is 1“8 a system of sub-populations formed within the large and small seeded gene pools, as a result of grower decision-making. Martin and Adams (1987b) point out the importance of "first understanding the farmer's requirements and perceptions of her own indigenous varieties." It is these requirements and perceptions which influence the farmers' planting decisions, and consequently produce the genetic pool found in their fields. MATERIALS AND METHODS The Malawian sample of 375 lines previously analyzed for isozyme variation (Chapter One) had been collected as a random sample from farmers' fields and stocks (Martin and Adams 1987a), and represented the frequency of genotypes found on each farm, rather than a full inventory of all types being grown. For the current study the lines were grouped according to the appearance of the seed, regardless of the site where they were collected. The groupings were based on a combination of color, pattern, shape and size, and identified by a consecutive number. These testa types are analogous to the seed classes in commercial dry bean agronomy and marketing, ie., visually similar lines marketed and processed as a unit. The sample had been grown in Malawi in 1983 under local standard experimental conditions in three replications, and line means fer morphological and phenological traits were calculated (Martin 198“). These line means, together with the isozyme infermation, were analyzed by seed class groupings in order to identify similarities of seed classes across farms. The isozyme data were summarized by assigning each line to either the Andean or Mesoamerican gene pool on the basis of the closeness of its isozyme complement to the most typical allozyme combinations, #1 or #7 (Chapter One). The data describing growth habit, and the presence of any segregation for qualitative traits found during planting in Malawi (Martin 198“), are listed in Appendix A. 1“9 150 In order to identify local names for the seed classes in the northern sample, an interview was conducted with Sheila Mkandawire (nee Konje), while she was living with her husband, Alex, in East Lansing, Michigan. Two to three seeds of each of the seed classes, taken from one accession, were examined by Mrs. Mkandawire. She commented on these types, in response to such questions as: 1. Have you seen this type in Malawi? 2. What are the local names, in which language(s)? Do the name(s) mean anything? 3. Is this bean tasty? Does it cook easily? A summary of her comments are listed in Appendix B. RESULTS Variation among Seed Classes In spite of the variability seen among Malawian bean landraces, certain phenotypic and genotypic variants known elsewhere have not yet been identified in the country. Gepts (198“) analyzed phaseolin, the major bean seed storage protein, in 81 northern Malawi germplasm accessions, not part of the Martin sample. Eighteen of the accessions were heterogeneous for phaseolin, and of the 101 total genotypes found, 62$ were the 'T' variant (associated with large seed), 20$ were '0' (snap bean) and 20$ were '8' (small seed) types; only one 'H' type was found. No 'B' or 'A' phaseolin types were feund. The distinctive "pop bean" type, eg. 'Nuna Huevo de Huanchaco' which carries 'H' phaseolin, has not been reported from observations of phenotype. Certain alleles of two NADH diaphorase loci (2}32:1 I. 2133;; g, Sggngig and 232222.!) were not found in the Malawian sample (Chapter One). Both null alleles are infrequent in New Werld material, and may not have been introduced into Malawi, or may not have been maintained after arising through mutation. The absence of the alleles found in tropical black and tropical red germplasm, Signg‘fi and 233221 I, may be associated with the fact that black beans are not popular in Malawi (Edje et al.), and that in some areas of eastern Africa shiny red beans are reported to lend a color to soup which is undesirable (Leakey 1970). Conversely, all allozymes found in Malawi were also present in the range of cultivars and accessions from Europe and North and South 151 152 America (Chapter One). Although in Malawi the use of varietal mixtures allows lines from the Andean and Mesoamerican gene pools to be interplanted, variability produced by intermating between gene pools was found to be at a low level (Chapter One). When the lines in the northern sample were grouped on the basis of testa type, 32 seed classes were identified among the 15 farms (Table 1). These classes contained from 1 to 57 lines each, with an average of 11.7 :_13.8. The distribution of lines per class was non-normal, with the five most frequent classes accounting for half of the lines present (Table “). Each class occurred in an average of 2.“ 1 2.1 farms, ranging from 1 to 10 (Table 1). The average number of classes present in the sample collected from each farm was 5.2 1 2, with a range from 2 to 9 (Table 2). No classes were present in all farms. When the seed classes were analyzed by color and pattern, the ratio of solid colored seed to patterned was found to be 20:12, or 62.5$:37.5$, comparable to the 7“.8$:2“.2$ feund by Edje et al. (1981) in 11,255 accessions collected from the entire country. Excluding heterozygous lines, eight seed classes were found to include more than one allozyme combination (Table 3). The limited variability for allozymes within class reflects the limited number of combinations found in the population as a whole. When seed classes were sorted into gene pools on the basis of isozyme pattern (Table “), the distribution of classes was predominantly into one pool, or one pool plus between-pool heterozygotes (eg., seed classes 8, 12, 23 and 28). This distribution is confounded by the fact that seed size differs across gene pools, but it may reflect the evolution of distinct testa color, shape and pattern genes within each pool. However, the major 153 1 Table 1. Seed classes among 375 lines collected in northern Malawi. Seed wt. Total no. No. Seed Description g/100 of farms Total # Freq. seed occurs in 1 brown elongate 52.5 2 1“ .037 2 purple speck on cream “7.8 1 7 .019 3 yellow elongate ““.2 5 “7 .125 “ cream elongate 30.5 2 3 .008 5 red mottle on cream ““.8 10 57 .152 6 brown rounded 29.6 1 1 .003 7 pink kidney “6.5 1 1 .003 8 white kidney ““.1 3 12 .032 9 pink rounded 28.2 1 1 .003 10 red mottle on yellow “5.3 2 2 .005 11 white small round 23.3 5 35 .093 12 red elongate/kidney “6.2 7 28 .075 13 brown mottle on cream “8.5 2 7 .019 15 black eye on gray 25.5 1 8 .021 16 brown stripe on cream ““.5 1 8 .021 17 black stripe on grey 25.5 1 1 .003 18 orange stripe on yellow 50.3 1 1 .003 20 red stripe on cream 52.7 3 21 .077 21 light purple elongate 21.1 1 1 .003 22 shiny red round 27.5 1 1 .003 23 slender red kidney 36.1 3 17 .0“5 2“ opaque green round 37.3 2 19 .051 25 purple on tan elongate “0.6 5 9 .02“ 26 red spot on white 52.6 1 2 .005 27 purple mottle on cream “9.9 2 19 .051 28 opaque red round/oval 52.1 5 20 .053 29 khaki oblong 58.6 1 3 .008 30 yellow round 38.2 1 2 .005 32 bright purple kidney 37.5 1 2 .005 3“ light purple round 56.6 2 15 .0“0 “6 dark shiny rectangular “1.0 3 8 .021 “7 brown rectangular 3“.5 1 3 .008 Mean 2.11:2.1 11.71133 1 Collected in Malawi (Martin 198“). 15“ Table 2. No. and type of seed classes in each 25 seed farm sample. Total Seed Type Farm Site lines/ ID # 1 2 3 “ 5 6 7 8 9 10 11 12 13 1“ 15 class —l O I _II —J 112---2--- - -- 111 2 7----- ------- 7 3 51558-1“---------“7 111--------2----- 3 5 -8153111“-9--1-1“57 6 -1—------- - --1 7 1--------—----1 8 --3---- -u-- 512 9 --1-- - -— ----1 1o --1-- --1------ 2 11 - - - 7 u — 22 - 1 - - - - 1 35 12 ---1---311-11“7-28 13 ---s----1-----— 6 111 ---1-------—---1 15 ----8--- ------ 8 16 ----8---------- 8 17 ----1--- -- -1 18 ----1--- -----1 20 -----19--11----21 21 --— - 1-..-- - -1 22 ------1---- --1 23 ----------- 7 9 9 2 I IINIU'II _n we N .4 I I I I I I I I I IIO‘IIRI can-I x:- II-eIN-‘NNI I INLA’tUlI-‘I N N 0 w 0 IIII IIIO‘II N w .1: I I I I 15 8 3 8:375 I I I IIIIII I I I II INII I _e g .1: 0" I I I I 13.“)! (D I :5 1 I 1 1 1 I 1 1 1 uiui 1 no. of classes “ “ 5 6 7 2 “ 2 6 9 “ 6 7 5 8 x=5.2:2.0 present 155 Table 3. Allozyme combinations occurring in each seed class. 1 Seed Isozyme Allele Combination TYPe # Total 1 2 3 “ 5 6 7 8 9 1O 11 12 Hets. 1 111 12 - - - 1 - - 1 - - - 2 7 5 - 2 - - - - - - - 3 117 33 7 - 6 - - - - - - - - - t1 3 1 - - - - — - - - - 2 5 57 52 - - 2 - - - - - - - - 2 6 1 - - - - - 1 - - - - - - - 7 1 1 - - - - - 8 12 8 - - — - - - - - - 11 9 1 - - - - 1 - - - - - - 1o 2 1 - — — — - - — - — - 1 11 35 - - - 33 - - - - 2 12 28 15 - - 2 - - - - - - — 11 13 7 - - - - 5 1 - - - - 1 15 8 - - - 7 - - - - 1 16 8 t1 - - — - - - - - 1 2 - 1 17 1 - - - - - - 1 - - - - - - 18 1 1 - - - - - - - - - - - .- 2o 21 21 - - - - - — - - - - - .- 21 1 — - - - - - - - - - - 1 - 22 1 - - - - — - 1 - - - - - .- 23 17 11 - - - - - - - - - - - 6 211 19 18 - - 1 - - - - - - - - - 25 9 1 - - - - - - - - - - - 8 26 2 2 - - - - - - - - - - - - 27 19 19 - - - - - - - - - 28 20 16 - - - - - - - I1 29 3 3 - - - - - - - - - - - - 3o 2 2 - - - - - - - - 32 2 1 - - - - - - - - - - 1 311 15 111 - - - - - - - - - - - 1 “6 8 - - - - — - 8 I17 3 - - - - - - - 3 Total 2u1 7 2 11 1 1 “8 1 1 1 2 1 56 0:373 See Chapter One, Table 6 for allozymes found in each combination. 156 Table “. 32 Malawi seed classes sorted by frequency and gene pool. Gene Pool Seed class1 Andean Mesoamer. Heteroz. Total no. of lines 5 5“ 0 2 56 3 “6 O 0 “6 11 0 33 2 35 12 17 0 11 28 20 21 0 0 21 28 16 0 “ 20 2“ 19 0 0 19 27 19 0 O 19 23 11 O 6 17 3“ 1“ O 1 15 1 1“ 0 O 1“ 8 8 0 “ 12 25 1 0 8 9 15 0 7 1 8 16 7 O 1 8 “6 0 0 8 8 2 7 O 0 7 13 0 6 1 7 11 o 1 2 3 29 3 0 0 3 “7 0 0 3 3 10 1 O 1 2 26 2 0 0 2 30 2 0 0 2 32 1 O 1 2 6 0 1 0 1 7 1 O O 1 9 0 1 O 1 17 0 1 O 1 18 1 O O 1 21 1 0 0 1 22 ' 0 1 O 1 Total 266 51 56 373 1 See Table 1 for descriptions of seed classes. 157 mechanism preventing isozyme variability within seed class is the complex of genetic barriers which eliminate gene pool-recombinant types in general, limiting them to 7.5$ (28 lines) of this sample. Grouping the lines found on each farm by gene pool (Table 5) shows that certain farm samples differ from the average distribution of such groups, 71.3$ Andean, 13.7$ Mesoamerican, and 15.0$ heterozygous or segregating. Significant departure from the average of 17.8, 3.“, and 3.8 gene pool lines per farm is evident in ten of the farms. Direct selection by the grower for seed class is probably the indirect cause of this variation. Seventy-five per cent of the isozyme heterozygosity is present in two farms, #10 and 12. The level of heterozygosity, H, on these two farms is .577, while it is .022 on the remaining ten farms, and .097 in the sample as a whole. The localization of this variability suggests selection for heterozygosity, or some trait associated with it. In the seed, distinctive testa patterns are found in segregating generations. The traits found to express heterosis in gene pool crosses were days to emergence (increased), days to end of flowering (increased), days to physiological maturity (increased), and yield per plant (increased). Increased yield, and the associated increase in lateness, seem most amenable to grower selection. It is interesting to observe that on neither Farm #10 or #12 are Mesoamerican lines present, and the source of the alternate gene pool alleles is not apparent. It may be that unpopular segregant testa types, outcasts from another farm, were being planted. Although data collection in Malawi did not fbcus on quantifying heterozygosity, in the course of field evaluations of this sample 55 158 Table 5. Number of lines belonging to each gene pool on 15 farms. Gene Pool Farm ............................... Andean Mesoamer. Hets. Total 1 2“ 1 0 25 * 2 2“ 1 O 25 * 3 2“ 1 0 25 * “ 12 11 2 25 ee 5 11 12 2 25 '* 6 23 O 0 23 NS 7 2“ 1 O 25 * 8 3 22 0 25 *' 9 22 1 2 25 NS 10 “ 0 21 25 " 11 25 0 0 25 ' 12 “ 0 21 25 " 13 23 0 2 25 NS 1“ 20 0 5 25 NS 15 23 1 1 25 us Total 266 51 56 373 Mean 17.8 3.“ 3.8 $ 71.3 13.7 15.0 NS = Not signficantly different from the mean distribution of types per farm at (.05 level. Other farms differ from mean at P<.05 level (') or P<.O1 (*') level. 159 lines were identified as segregating for qualitative traits, such as hypocotyl and flower color, growth habit and testa appearance (Martin 198“). This heterozygosity would be expected to be found within and between gene pools in ratios resulting from the gene frequencies in each pool, and would be expected to be most common in the more frequent large seeded lines homozygous for the Andean isozyme pattern. However, “2 of of these 55 morphological heterozygotes were found to be segregating for isozymes, meaning that they resulted from between-pool crosses (Table 6). Additional lines were identified as heterozygous when seed increases were done-at MSU, but the proportion of inter-pool crosses among the morphologically segregating lines remained high, 62.0$ (Table 6). This suggests that within-pool variability may be low, even for non-isozyme characters. In beans the trait most amenable to grower selection, after seed appearance, is probably growth habit. The single allele difference (Bliss 1971) between determinate and indeterminate types produces several changes in plant structure which are easily seen, and which cause important consequences at many points in a cropping system. Among Malawian farmers there is a general expectation that certain seed classes are bush, semi-climbers, or climbers, and a grower has precise knowledge about the growth habit differences within classes within his or her own bean stocks (Ferguson 1987). The maintenance of lower- yielding bush types in varietal mixtures and within seed classes, where natural selection pressure is towards the more fecund climbing type, suggests human intervention. In a nationwide sample of bean germplasm, over 75$ of the lines were found to be climbers, explained as a result of the need for 160 Table 6. Total heterozygosity associated with isozyme heterozygosity. Lines identified as heterozygous via Allozyme ............................................... combinations Isozymes Malawi/field Total morphological :ndean — —————— ---------- ...... ------ ........ - ........ - .............. 1 2“1 6 16 2 7 0 1 3 2 0 1 “ 11 O 1 9 1 O O 11 2 1 1 12 1 O 0 Mesoamerican 5 1 0 0 O 6 1 0 0 7 “8 5 6 8 1 1 1 Segregating 56 “2 an 161 indeterminate types to interplant with maize (Edje et al. 1981). Bush lines make up 21.6$ of the northern sample, comparable to 2“.6$ of the countrywide sample. Determinate habit is present only in the Andean types of the northern sample (Table 7 and 8), which suggests that of the six different races defined by Evans (1975) on the basis of seed size and growth habit, Race “, the small-seeded determinate, is not present in Malawi. The determinate small seeded type is unknown in the wild (Gentry 1969) and uncommon in improved agriculture. Crossing programs in the U.S. were unsuccessful in generating a bush navy bean type (Adams pers. comm.) until one was found in irradiated material in a mutation breeding program (Andersen et al. 1960). In Malawi, in spite of interplanting of both gene pools, and the use of types of both growth habits, it appears that either a viable recombination of these traits has not occurred or has not been maintained. Among stocks collected at Dedza (Ferguson 1987) two farms had small white and small red seeded lines which growers identified as bush types. These may be indeterminate bush, but it would be be of interest to know whether they represented successful inter-pool recombination. The gene pool heterozygostes are predominantly indeterminate (Table 8). When data on growth habit within seed classes are examined (Table 9), it appears that the more popular seed types are maintained in both growth habits. In seed classes #3, 5, and 12, the three most frequent Andean classes in the sample, both determinate and indeterminate types are present in about equal numbers. The remaining classes consist almost entirely of only one growth habit or the other. The presence of both growth habits in the most popular seed classes suggests that 162 Table 7. Occurrence of growth habit types among isozyme combinations. Growth Habit Allozyme .............................. Combination Determinate Indeterminate Total 0 3 53 56 1 72 169 2“1 2 0 7 7 3 0 2 2 “ 3 8 11 5 0 1 1 6 0 1 1 7 0 “8 “8 8 0 1 1 9 O 1 1 1O 0 1 1 11 O 2 2 12 1 0 1 Total 79 29“ 373 Table 8. Growth habit among gene pools in the Malawian sample. Growth Habit Gene Pool Determinate Indeterminate Total Andean 76 190 266 Mesoamerican 0 51 51 Segregating 3 53 56 163 Table 9. Growth habit in 32 seed classes in the Malawi sample. Growth Habit Seed .................... Class Determ. Indeterm. Total 1 0 1“ 1“ 2 0 7 7 3 18 29 “7 “ 0 3 3 5 26 31 57 6 0 1 1 7 0 1 1 8 1 11 12 9 O 1 1 1O 1 1 2 11 O 35 35 12 12 16 28 13 0 7 7 15 0 8 8 16 O 8 8 17 O 1 1 18 0 1 1 20 21 O 21 21 1 0 1 22 0 1 1 23 0 17 17 2“ O 19 19 25 1 8 9 26 0 2 2 27 0 19 19 28 0 20 20 29 0 3 3 30 0 2 2 32 0 2 2 3“ O 15 15 “6 0 8 8 “7 0 3 3 Total 81 29“ 375 Freq. .216 .78“ Growth habit data from Martin (198“). 16“ farmers want the planting options provided by both growth habits, and have maintained bush types. This is consistent with the findings that farmers can usually identify bush or climbing types in their own seed stocks. The other of the four most frequent seed classes, #11, is only present in the sample as an indeterminate type. This small white seed type is Mesoamerican, and the indeterminate semi-climber type may be the only bush-like option available, because of the lack of genetic exchange between the pools. Most farms did not differ significantly from the average distribution of growth habit in the sample as a whole (Table 10). However, departure from the population norm of 5.“ determinate and 19.6 indeterminate lines per site at Farms 6, 7 and 9, suggests positive selection for bush types by these growers. The genetic similarity of seed stocks grown on the same farms was one of the questions addressed by Martin (198“). Using principal components analysis scores, he indicated pairs of minimum and maximum distance between lines on each farm. These are shown in Table 11, with the isozyme and seed class designations produced in the current study. The majority of minimum distance pairs belong to the same seed class, and contain the same combination of isozyme alleles. Variation within Seed Class The amount of genetic variation present within Malawian seed classes will affect future improvement programs. Martin concluded from analyses of variance of 18 traits from three seed types, found in Farm #1 (seed classes #1, 2, and 3), that homogeneous seed type groups were present, but that they were not necessarily similar for metric traits. 165 1 Table 10. Distribution of different growth habit lines on 15 farms. Growth Habit Farm Determ. Indeterm. Total 1 0 25 25 9' 2 1 2“ 25 * 3 2 23 25 NS “ 8 17 25 NS 5 1 2“ 25 ' 6 23 2 25 *9 7 2“ 1 25 '9 8 2 23 25 NS 9 1O 15 25 9 1O 1 2“ 25 9 11 1 2“ 25 F 12 O 25 25 " 13 1 2“ 25 9 1“ 7 18 25 NS 15 0 25 25 '* Total 81 29“ 375 Farm av. 5.“ 19.6 1 Data collected by Martin (198“). 5,5! = Signficantly different from the average farm distribution of types at the .05, .01 level. NS = Not significantly different at .05 level 166 1 Table 11. Lines with minimum and maximum distance in 15 farm samples. Minimum Distance Maximum Distance Farm Line # All6zyme—_-CI;;; Line;—# Allozyme -0155; 1 9.10 3.3 2,2 2,17 1.“ 2.3 2 9,10 1,1 3.3 20,23 1.1 5.3 3 12,22 1,1 5.5 16,25 7.1 9,8 “ 10,11 0,7 11,11 20,23 “,1 5.3 5 13.2“ 7.7 15,11 6,25 1,0 5,16 6 1,23 1,1 3,3 10,22 1,1 3.5 7 15,2“ 1,1 5,20 13,23 1,7 20,22 8 “,12 7.7 11,11 20,21 7.7 11,11 9 3,13 1,1 2“,2“ 10,22 1,1 5.25 10 23,25 1,1 28,28 5,8 1,0 20,25 11 1,2“ 1,1 27,27 6,13 1,1 20,28 12 11,19 0,0 “6,12 10,17 0,0 23.12 13 2,1“ 1,1 2“,2“ 10,25 1,1 12,12 1“ 19.21 1,1 28,28 2,6 0,1 23.12 15 7,21 1,1 27,27 “,25 7,“ 11,5 1 Minimum and maximum distance pairs from Martin (198“). I'l-u‘ I? 7'71"“ 167 In the current study, analyses of variance of the 21 quantitative traits measured in Malawi (Chapter One, Table 8) were done on seed classes across those farms where more than one line of the class occurred. In certain seed classes there was significant variability both within and among farms, suggesting that mass selection would produce improvement of seed class groups. . Where bush and climbing types of the same seed class are known, the two ideotypes would be expected to be genetically distinct for more than growth habit loci beacuse they are being handled differently by farmers and being exposed to different possible mates. For example, farmers may choose to grow only one growth habit type, or both, during the main or secondary seasons, in various parts of the field or row, with different neighboring lines, etc. For seed class #5, the popular red mottle on cream "sugar bean", the sample suggests that farmers select and maintain either determinate or indeterminate types in their stock, but not both (Table 12a), although it is possible that this array may reflect sampling from only one growing season, or from only part of the seed stock belonging to the farmer. The large red kidney class, #12, also appears to be maintained as growth habit ideotypes on those farms with the largest number of these lines, Farm 12 and 1“ (Table 12b). The lack of isozyme variability found within gene pool, and consequently within seed class, allows little analysis of the genetic differences between determinate and indeterminate types of the same class. The only seed class containing appreciable diversity for isozyme alleles and growth habit is #3, with yellow elongate seeds (Table 12c). The “7 lines present across five farms include allozyme combinations #1, 2, and “. The last two differ from #1 by one allele each. Allozyme 168 Table 12. Growth habit differences among three seed classes, by farm. Farm Growth Habit Determ. Indeterm. Total a) Seed class #5, red mottle on cream (040(11sz 12 OOOQtd—IWOO z—s—smoooomoo —. z—a—axoc—a—awmoo 15 Total 26 u. c—D U1 -4 Freq. .“6 .5“ b) Seed class #12, red elongate/kidney “ 1 0 1 8 2 1 3 9 1 O 1 10 0 1 1 12 O 11 11 13 1 3 “ 1“ 7 0 7 Total 12 16 28 Freq. .“3 .57 c) seed class #3. Yellow elonagate 1 0 5 5 2 1 1“ 15 3 1 “ 5 “ “ “ 8 6 12 2 1“ Total 18 29 “7 169 combination #2 was found only in this seed class, while #“ was found in two other large seeded classes. The distribution of isozyme types found in this seed class across farms (Table 13) showed that combinations #2 and “ occurred only in the Chitipa North area; farms in the Misuku Hills area had only allozyme combination #1. (These areas are separated by transportation barriers which may reduce germplasm exchange.) Isozyme combinations #2 and “ also did not occur in determinate lines (Table 1“). Thus, four different populations, determinate isozyme #1, indeterminate isozyme #1, indeterminate isozyme #2, and indeterminate isozyme #“ (Table 1“). occur in the sample. These genotypes were identified on the plots produced from principal component analysis of individual farm samples (Martin 198“). In Farm 1, the indeterminate #2 lines (5. 6 and 15) clustered on principal component axis (PC) 1 separately from indeterminate #“ lines (1“ and 17). In Farm 2, the 15 yellow seeded lines grouped centrally on both axes; all four growth habit/isozyme combinations were present on this farm, but the single determinate #1, line 23, was most separate on PC 1. Farm 3 also had an out-lying determinate #1, line 8, but it was distant on both axes. In Farm “, the determinate #1 (15, 17. 21, 23) and indeterminate #1 lines (1, 6, 7. 1“) clustered separately on PC I. On this farm there were significant differences in seed size between the determinate and indeterminate lines, 52.3 g/100 seed and 36.2 g/100 respectively, compared to “9.0 and “2.“ g/1OO between the two growth habits in this seed class overall. Size differences of this kind could aid in the identification of bush and climbing ideotypes by seed alone, and indicates the amount of genetic differentiation the two growth habit types have undergone. In Farm 6 the determinate #1's (1, 3, “, 5. 6, 170 Table 13. Allozyme combinations found in seed class #3. Allozyme Combination Area Farm #1 #2 #“ Total Chitipa North 1 0 3 2 5 2 8 “ 3 15 3 “ O 1 5 Misuku Hills “ 8 O 0 8 6 13 0 0 13 ............................... - --;- Total 33 7 6 “6 *One line was lost during seed increase at MSU. Table 1“. Isozyme-by-growth habit groups in seed class #3. Growth Habit Allozyme -------------------- Combination Determ. Indeterm. Total 1 17 16 33 2 0 7 7 “ 0 6 6 - ........................................... ' -- Total 17 29 “6 ' One line was lost during seed increase at MSU. 171 12, 15, 16, 18, 19, 23) clustered centrally on both axes, with the indeterminate #1's (8, 10) together as outliers. Although the the principal components were influenced by traits involved in growth habit differences (number of nodes per main stem, growing season, etc.), there appeared to be genetic differences within this yellow seed class. In order to examine the genetic relationships among all lines of seed class #3, a principal component analysis was made of them as a group, based on the data used by Martin (198“), and leaving out the variable for number of nodes on the main stem (Chapter One, Table 8). The first two principal components, which accounted for 9“.81$ of the variance, were plotted (Figure 1). The determinate types carrying isozyme combination #1 clustered on PC I (Figure 1a), and did not occupy the range shown by the indeterminate #1 types (Figure 1b). The indeterminate lines with isozyme combinations #2 and “ (Figure 2c,d) overlapped only in the center of their distributions on PC I, and together approached the range covered by indeterminate isozyme type #1. The traits contributing to the first principal component, which appears to distinguish between growth habit differences within isozyme combination #1 (Figure 1a and b), and between isozymes #2 and #“ (Figure 1c and d), are those directly and indirectly involved with seed size: seed weight, length, width, seed yield, and yield components. This indicates that seed size differences, which have the potential to be selected by growers, are contributing to the process of genetic differentiation within the seed class. Figure Figure Figure Figure Figure 1. 1a. 1b. 1c. 1d. 172 Four growth habit and isozyme groupings of seed class #3 plotted on the first two principal component axes following analysis of quantitative data for 20 morphological traits. Determinate growth habit, Isozyme combination #1 Indeterminate growth habit, Isozyme combination #1 Indeterminate growth habit, Isozyme combination #2 Indetermiante growth habit, Isozyme combination #“ 173 open .oH shaman F ...ZMZOQZOQ .._a essay >o.a..a- acmu=-_m names. _omuoam >aeoaoe. -=_.eao >.__L.w.._.m\-u..p ea...m\>u..a..= so ow en an on on on su «a mm vs nu «a «a on a. so o. n. no N_ __ o. o a s o n ¢ m u =5 .8... =e.~au.o.=_ ea an»— ...zsaueuuz somogm x. uaasaaeaceu a...._u a... Lace. ans-_soe. _ueucum .n. a~gap Table 16. Planting 179 and seed selection practices from S. Mkandawire. # Seed Class Information 1 brown elongate 5 red mottle on cream 12 red elongate/ kidney 16 brown stripe on cream 29 khaki oblong Planted in February because early maturing; in high demand at planting time and therefore seed saved directly after harvest in ashes Planted with maize, not separately in own part of field, seed saved for planting Planted with maize because it took longer to mature Planted separately in its own part of the field, seed saved for planting Planted separately in part of the field because it is particularly tasty, seed is saved separately for next season, it does not grow well with maize DISCUSSION Comparisons between Northern and Central Malawi Samples Evidence that grower selection is the major ferce producing the genetic structure found among farm stocks in Malawi comes from the predominance of large seeded beans, and the presence of determinate growth habit, counter to expectations for bean populations under natural selection. The situation in Malawi is in accord with Harlan's prediction that under grower selection in traditional agriculture "the population becomes an array of deliberately chosen components” (1975). Harlan's assumption that growers will know such traits of the landraces as maturity, soil type adaptation, and preferred culinary use, and that landraces are adapted to local climate, cultural practices, diseases and pests, remains to be verified in the northern region, although anecdotal evidence has been feund (Bortei-Doku et al., Martin and Adams 1987a). Detailed interviews with 17 growers in the central part of the country have recorded examples of the knowledge associated with bean ideotypes (Ferguson 1987). Analysis of the bean types grown in the area shows several parallels with the northern Malawi seed population, although the collection strategy differed. The northern sample reflects the frequency of local bean types because it was a random collection of 25 of the lines growing or harvested on each of 15 farms. In the Dedza area of the Central Region a comprehensive collection was made of each seed type grown by 17 farmers in the 1986-87 rainy season, and information on frequency was acquired through 180 181 interviews with growers on each farm. The distribution of seed classes is comparable between the two areas studied. The 220 lines collected in Dedza were sorted into 36 seed type classes, using the same visual clues which identified 32 classes from the farms in the north. Seventeen of the seed classes are common to both areas. The 11 most frequent classes in Dedza accounted for 65$ of the collection, while eight of the classes accounted for 65$ of the lines in the northern sample. Across farms in both areas many classes are found to be localized, present at one to four farm sites, while a few classes are widely distributed (Table 7). The most widely grown classes among the Dedza sample were the bush white, the khaki, the red mottle on cream, the small white, and the large red kidney. Only the bush white does not appear at all in the northern sample, and the red mottle on cream, the small white, and the large red kidney, seed classes #5, 11 and 12, ranked as the first, third, and fourth most common in the north (Table 5). The khaki, #29, is present only in the southernmost farm of the northern sample, within the boundaries of the Central Region. In Kenya a similar distribution of local and cosmopolitan types occurs, where a sample of 997 lines from across the country were found to fall into 78 seed types; 62.2$ of the lines belonged to 10 seed classes, and the remaining 377 lines were distributed among 68 classes (Rheenen 1979). This diversity of acceptable seed types has been noted in several African countries, Ethiopia (Westphal 197“), Uganda (Leakey 1970), Kenya (Rheenan 1979), and Rwanda (Lamb and Hardman n.d.). and suggests a consumer flexibility that will allow breeders to focus on requirements other than seed class (Rheenan 1979). Nevertheless, 182 regional preferences are in evidence, and preferred classes such as the cosmopolitan types in Malawi are a logical starting point for improvement. Plant growth habit distribution differed between Dedza and the north. Bush and climbers were equally represented in lines from the Dedza sample, “6.8$ to “5.5$, with 6.3$ semiclimbers and a few unknowns. In the north, bush types were less than a quarter of the sample (21.2$) (Table 9). This may reflect sample taking during different planting seasons of the year, when different types are in the field or in supply. The ability of the Dedza farmers to identify 98.6$ of their varieties fer growth habit suggests the importance of this information when making planting decisions. Habit affects time to maturity and determines the optimum planting season, the need fbr support from maize, and the need to delay planting of vigorous climbers which may otherwise outcompete maize. In Dedza, bush and climbing ideotypes are known in some seed classes, such as the common red mottle on cream, but in most cases only one such type is grown on a farm (Ferguson 1987). The northern sample had six classes which contained bush and climbing varieties, including three of the four most frequent types (Table 9). In these three classes, seven out of 22 farms had varieties of both habit. The recognition of hybrid or off-types by farmers, and their fate during seed selection, are important factors in the genetic structure of the farm and regional populations. Lines resulting from intermating between two different seed types are identifiable when, following the first generation after the cross, their testa pattern differs from the parent plant seed type. The occurrence of three layers of color in the testa pattern, or combinations of two pure-line patterns in the same seed, can identify non-true breeding seed by its appearance alone. 183 These off-type patterns are sometimes named, sometimes given the epithet ”stranger” (Martin and Adams 1987), or left unnamed. In the Dedza sample 30 lines (13.6$) were unnamed, and farmers reported that most of these types (56.3$) "came from" known varieties, or appeared "uninvited" in the field; such types are sometimes not regarded as real varieties. Preliminary results suggest that unnamed types are rejected from planting in much higher frequency (38.7$) than seeds belonging to named classes (7.“$) (Ferguson 1987). It is possible that farmers reject heterozygous patterns not only because the seed type is unfamiliar and undesirable, but also because their agronomic characters are not predictable; a bean planted with the expectation that it is a bush may turn out to be indeterminate. The Maintenance 25 Ideotypes A progression from simple mixtures of classes, through hybrid swarms, to predominantly recombinant types (Martin and Adams 1987a and b), does not appear to be the norm in individual farm stocks in Malawi. The combined results from Malawi support the idea that some bean classes have the status of ideotypes, maintained or rejected by farmers depending on how well they fit the agronomic and utility niches available, and that many non-varietal seed type segregants are eliminated through positive selection for recognized ideotypes. Not all genetic recombinants are eliminated under this pressure, and new classes are able to arise, but variation within classes may be progressively reduced, with the result that most diversity would be expected to occur between classes. The linkage between cosmopolitan seed types and the ideotypic traits associated with them would also be expected to become 18“ closer over time. Thus, when sampling genetic diversity, the localized lines found on one to a few farms may provide more genetic variation than samples of the most frequent types. To collect diversity within popular seed types may require sampling from as wide an area as possible. The maintenance of seed class ideotypes reduces variability on a single farm but allows decision-making and planning. The occurrence of a few widespread types and many diverse types found in low frequency across farms appears to be the structure within which variation is retained in the population as a whole. Gentry (1969) points out the ease with which ideotypes can be maintained, describing the case in Mexico, Phaseolus...is self fertile and selections are easily made with seeds from individual plants. Because of the limited segregates carried in the succeeding selfing generations, pure lines are soon established by avoiding seed of the undesirable variants. The occasional cross-pollinations...result in new genetic combinations, or new choices for the planter... The traits selected by farmers in Mexico--large seeds, flavor, quick- cooking properties, fancy seed colors and growth habit--have been incorporated into ideotypes in Malawi, and the strategies of using bush types for short seasons and later maturing climbing types in the long growing seasons, have also been repeated in Malawi (Gentry 1969). Kaplan's (1981) assertion, on the basis of market samples, that specific mixtures of g; vulgaris components have been stable in Mexico for long periods of time, also suggests human-directed maintenance of seed types. A parallel occurs in cowpeas in Botswana, which are often maintained as varietal mixtures (deMooy 1985). Cowpeas sampled from farm stocks were found to have the highest diversity indices among 185 qualitative characters for the traits of growth habit and seed color. Both have been shown in Malawian beans to be important selection criteria, involved in decision-making by the grower. In Botswana it was seen that even limited introductions of new types were able to add to the richness of the stocks of the farmers. A Californian cultivar of cowpea was introduced over a decade ago, and has become one of the lines maintained on many farms, although it occurs at low frequency in the farmers' inventories. Gene exchange between this cultivar and local material has taken place (deMooy 1985). Perceptual Distinctiveness A theory proposing selection for perceptual salience in crops may be applied to the bean diversity of Malawi to suggest that one of the major pressures maintaining the array of visual variability is the requirement for identification by growers and consumers. Easter (198“, 1985) examined the question of crop diversity in traits unrelated to use, and proposed that it results from the need to distinguish and maintain cultural inventory. Genotypes must be distinguishable before they can be selected on the basis of utility, or else inventories are reduced to types producing more planting material (Boster 1985). Growers distinguish types via those plant parts which show the greatest range of variation and "perceptual salience", and consequently variants are accumulated. Boster's theory of selection for perceptual distinctiveness (SPD) developed from observations of the range of visible phenotypic differences among landraces of Manihot esculenta. He suggested that manioc landraces have become extremely variable for stem color, petiole 186 color, leaflet shape, etc. via human selection, to allow tuber types to be differentiated by farmers. He expected that SPD occurs most readily in clonal tuber crops where newly arisen genotypes are replicated in precise correspondence with the distinctive plant characters, and where the primary edible part is usually hidden during growth and harvest. Boster makes several predictions based on his observations of manioc. He proposes that for a morphological character to be used in perceptual selection it cannot interfere with use, and consequently variation within a crop will increase the most in non-adaptive characters. SPD will not allow the evolution of geographic races differing for the visually distinct traits being used as taxonomic markers, because they will be equally useful in all regions. The characters used most frequently for dry bean variety identification in all types of agriculture are those involved in the appearance of the seed coat, and it is the most visually differentiated part of the plant under domestication. Wild beans tend to be mottled in shades of tan suitable fer camouflage on soil (Gentry 1969). In contrast, cultivated types have accumulated immense diversity for color and pattern. While release from natural selection pressures does allow an increase in variation in certain plant characters under domstication, Boster points out that this leaves unexplained why so many of these characters are perceptually salient to humans (1985). This idea can be used to address the paradox of the low isozyme diversity feund in a crop which is so extremely variable for the morphological appearance of its seed. The isozyme variation seen to date in g; vulgaris is not unusually low when compared to certain other self-pollinating crops, such as tomato (Chapter One). These enzyme loci 187 are expected to be selection-neutral with regard to humans, and there is. no evidence that they are not. However, all the genes influencing the perceptible seed coat traits of color and pattern in bean are under strong human selection which can operate to accumulate as many allelic variants or combinations as arise in the population. As Boster's hypotheses suggest, the trend is toward positive selection for variation in these traits. In addition, each isozyme can represent variation at only one locus, while multiple genes produce each color or pattern. The accumulation of alleles at many loci geometrically increase the possible combinations of alleles affecting testa variation. However, bean departs from Boster's prediction that visual variation will increase in parts unrelated to utility. The color variation of the seed is under selection because the seed is the part eaten, and its pigments affect the appearance of the cooked dish. Some seed colors are preferred in Malawi for their resemblence to meat in soup, while black beans, although distinctive and often high yielding, are often rejected from the inventory because of the undesirable color they produce when cooked (Edje et al. 1982). Thus, while Boster suggests that pigment systems are most useful as SPD characters because they are of low cost to the plant and usually neutral under other selection pressures (198“, 1985). the bean model shows that great variation can occur in characters which are under selection for both utility and distinctiveness. Numerous seed types are acceptable to at least some portion of the population in Malawi, and are maintained, if at a low frequency. In certain Central American countries the stringency of preference for seed type is high, to the point that only a single class of beans may be grown. 188 Regional diversity is known for the perceptually distinctive trait, seed appearance, in traditional Mesoamerican bean culture (Kaplan 1981). and is obvious in Malawi where the predominant seed colors change in frequency as one moves southward in the region sampled (Martin and Adams 19878). Local preferences are common in most fbods, and this departure from Boster's hypothesis that geographic races differing for the distinctive traits will not occur is logical, because the seed is the edible part. In beans, where recognition of a type is primarily by the appearance of the dry, mature propagule, additional identifying plant parts are not needed at harvest and sowing time, as they are in manioc. Seed propagation means that seed patterns are highly selectable, and the maintenance of types is aided by the predominantly self-pollinating nature of beans, and the distinct patterns produced by some forms of heterozygozity. Boster found that manioc growers each maintained at least some of all manioc types known to them, but Malawian farmers recognize more seed types than they themselves plant (Appendix B, Ferguson 1987). Boster thought that the former practice resulted from the ability of SPD to provide a way of maintaining all stocks, but it may reflect inaccessibility of markets or other exchanges to which Malawians have access, or the difficulty of storing propagating material of manioc. Boster assumes that taxonomic characters will vary independently from each other, and that combinations of all characters will be produced as they are selected and maintained over time (1985), but in bean the barriers between the gene pools appear to prevent certain combinations of the selectable characters present. 189 Perceptual distinctiveness relates to maintenance of diversity and to improvement programs. Boster defines the condition which promotes selection fer perceptual distinctiveness as being the maintenance of many cultivars by people who appreciate the value of having crop diversity. Distinctiveness becomes more important as the number of varieties in the inventory increases and recognition of many types is required. This is the case in Malawi, where bean growers recognize and maintain many types, and often have precise reasons for using the types feund in their stocks. Loss of knowledge of cultivars will result in loss of diversity. In Malawi there is some evidence that older, experienced growers maintain more classes in their bean inventory than younger ones, and that knowledge is related to diversity (Ferguson 1987). In manioc culture the process of introduction, selective maintenance and random loss resulted in a few common, widely known cultivars, and in a much larger number of lines grown by few people (Boster 198“). This is the situation among landrace beans in Malawi. Boster concludes that the maintenance of crop genetic diversity depends on the cultural practices of the growers, and this is compatible with what the evidence shows in Malawi. In modern agriculture bean cultivars can be extremely alike as to seed phenotype, yet easily distinguished on the basis of administrative records or chemical/genetic analysis (Chapter One), but traditional farmers in who seldom buy seed must remember the qualities of their lines by associating them with testa differences. Thus, perceptual distinctiveness must be considered in bean breeding and improvement programs for subsistence agriculture where numerous varieties are maintained. 190 New varieties which arise in traditional agriculture need to be distinguishable on the basis of their combination of folk taxonomy characters from those already in cultivation, and problems may arise if new types are too similar to old ones (Boster 1985). These conditions will also apply when improved cultivars are introduced to growers who are going to incorporate them into their stock mixtures. "Cultivars that demand careful attention to be discriminated will not diffuse as rapidly", and low perceptual distinctiveness lowers the probability of diffusion (Boster 198“). It is not expected that growers will automatically eliminate their old stock of a seed class when an improved line is available, and, where seed type is identical, there may be a problem of maintaining improved types in the farm population without the benefit of a selection criterion. Breeding lines resembling desirable grain types fer incorporation into variable farm stocks is appropriate under current agricultural conditions in Malawi, since the use of mixtures is probably the optimum strategy fbr stabilising production and maintaining diversity. It is being carried out in eastern African bean improvement (Anonymous, 198“). However, some seed pattern element may need to be distinctive in order for new lines to be discriminated and maintained by growers. It is possible that another morphological character can be used to distinguish bean genotypes. Multilines have advantages of homeostasis and give a uniform seed type (Marshall 1977). However, the fact that they differ by one allele, or a small linkage group, means that they cannot supply the same genetic buffering as the broad-based stocks already in use. They do not supply the diversity of classes desired by the grower/consumer. Certainly, 191 because of the unifbrm appearance of the component lines and lack of perceptual distinctiveness, there would be little chance that any specific frequency, whether the original one or some other suited to changing epidemiology, could be reconstituted from season to season by the farmers. There are diverse genetic elements present in bean landraces in Malawi, but the genetic dynamics of the crop are dominated by the growers. LIST OF REFERENCES Adams, M. W. 1973. Plant architecture and physiological efficiency in the field bean. In "Potentials of Field Bean and Other Food Legumes in Latin America". Series Seminars No. 2E. Centro International de Agricultura Tropical, Cali, Colombia. Allard, R. W. 1960. Principles of Plant Breeding. John Wiley & Sons, New York. Allard, R. W. 1961. Relationship between genetic diversity and consistency of performance in different environments. Crop Science 1:127-133. Allard, R. W., Jain, S. K. and Workman, P. L. 1966. The genetics of inbreeding papulations. Advances in Genetics 1“:55-131. Allard, R. W., Kahler, A. L. and Weir, B. S. 1972. The effect of selection on esterase allozymes in a barley population. Genetics Andersen, A. L., Down, E. E., and Whithrd, G. 1960. The Sanilac pea bean--its history, deveopment and characterisitics. Quarterly Bulletin, Michigan Agricultural Experiment Station “3:21“-236. Anonymous. 198“. Annual Report 198“: Bean Program. Centro International de Agricultura Tropical. Bortei-Doku, E., Barnes-McConnell, P. W. and Edje, 0. T. 198“. Malawi: focus on a bean culture. Bean/Cowpea CRSP, Michigan State University, East Lansing. Bliss, F. A. 1971. Inhericance of growth habit and time of flowering in beans, Phaseolus vulgaris L. J. Amer. Soc. Hort. Sci. 96:715-717. Boster, J. S. 198“. Classification, cultivation, and selection of Aguaruna cultivars of Manihot esculenta (Euphorbiaceae). Advances in Economic Botany 1:3“-“7. Boster, J. S. 1985. Selection for perceptual distinctiveness: evidence from Aguaruna cultivars of Manihot esculenta. Economic Botany 39:310- 325. Briggs, F. N. and Knowles, P. F. 1967. Introduction to Plant Breeding. Reinhold. deMooy, B. E. 1985. Germplasm evaluation of Botswana cowpea (Vigna 192 193 unguiculata (L.) Walp.) landraces. M.S. Thesis, Michigan State University, East Lansing. Dessert, J. M. 1987. Changes over time in the proportion of beans in a varietal mixture. Bean Improvement Cooperative 30:79-80. Donald, C. M. 1968. The breeding of crop ideotypes. Euphytica 17:385- “05. Edje, O. T., Mughogho, L, K., Rao, Y. P., and Msuku, W. A. B. 1982. Bean production in Malawi. In Potential fer Field Beans in Eastern Africa, pp 5“-97. Centro Internacional de Agricultura Tropical. Evans, A. M. 1975. Genetic improvement of Phaseolus vulgaris. In Nutritional Improvement of Food Legumes by Breeding, pp 107-115. John Wiley & Sons, New York. Ferguson, A. E. 1987. Socio-economic factors influencing genetic diversity in Malawian bean landraces: Central Region. Bean/Cowpea CRSP Malawi/Michigan State University Project, East Lansing. Ferguson, A. E. and Sprecher, S. 1987. Women and plant genetic diversity: the case of beans in the Central Region of Malawi. Rural Africana, in press. Gentry, H. S. 1969. Origin of the common bean, Phaseolus vulgaris. Economic Botany 23:55-69. Gepts, P. L. 198“. Nutritional and evolutionary implications of phaseolin seed protein variability in common bean (Phaseolus vulgaris L.). Ph.D. dissertation. University of Wisconsin, Madison. Harlan, H. V. and Martini, M. L. 1938. The effect of natural selection in a mixture of barley varieties. Journal of Agricultural Research Harlan, J. R. 1975. Crops and Man. American Society of Agronomy/Crops Science Society of America, Madison. Hawkes, J. G. 1981. Germplasm collection, preservation and use. pp. 57-83 in Frey, K. J. (ed.). Plant Breeding II. Iowa State University Press, Ames. Jain, S. K. 1983. Domestication and breeding of new crop plants. pp. 177-197 in Crop Breeding, Wbod. D. R. editor. American Society of Agronomy, Inc. and the Crop Science Society of America, Inc. Madison. Kaplan, L. 1981. What is the origin of the common bean? Economic Botany 35:2“0-25“. Lamb, E. M. and Hardman, L. L. n.d. Final report of survey of been varieties grown in Rwanda, January, 198“ - June, 1985. US-AID/ University of Minnesota and the Government of Rwanda. 19“ Leakey, C. L. A. 1970. The improvement of beans Phaseolus vulgaris (Fabaceae) in East Africa. In C. L. A. Leakey, ed., Crop Improvement in East Africa, Farnham Royal, England. Marshall, D. R. 1977. The advantages and hazards of genetic homogeneity. Annals New York Academy of Sciences 287:1-20. Martin, G. 198“. Genetic Diversity of Bean Landraces in Northern Malawi. M.S. Thesis. Michigan State University. Martin, G. B. and Adams, M. W. 1987. Landraces of Phaseolus vulgaris (Fabaceae) in northern Malawi. 1. Regional variation. Economic Botany Martin, G. B. and Adams, M. W. 1987. Landraces of Phaseolus vulgaris (Fabaceae) in northern Malawi. 11. Generation and maintenance of variability. Economic Botany “1:20“-215. Mehra, K. L., C. B. Singh and K. S. Kohli. 1970. Phenotypic diversity and breeding of forage cowpea. XI International Grassland Congress. (ed.) M. J. T. Norman. Surfers Paradise, Queensland, Australia. pp 293-299 0 Poehlman, J. M. 1979. Breeding Field Crops. 2nd ed. AVI publishing Co., Westport. Rheenen, H. A. van. 1979. Diversity of food beans in Kenya. 33:““8- “5“ Roos, E. E. 198“. Genetic shifts in mixed bean populations. 11. Effects of regeneration. Crop Science 2“:711-715. Simmonds, N. W. 1979. Principles of Crop Improvement. Longman, London. Vieira, C. 1975. Change in varietal composition of bean mixtures after successive plantings. Bean Improvement C00perative 18:85-87. Wright, A. J. 1983. The expected efficiencies of some mthods of selection of components for inter-genotypic mixtures. Theoretical and Applied Genetics 67:“5-52. APPENDICES APPENDIX A Data for 375 Bean Lines Sampled from 1S Farms £2 Northern Malawi. # = Case number FARM = Farm number, 1 to 15 LINE = Line number within farm, 1 to 25 SC = Seed class (see Chapter One, Table 1), 1 to “7 ALZ = Allozyme combination, 0 to 12 (Chapter One, Table “) In the following isozyme data, alleles are coded as 1 = Fast, 2 : Slow, 3 = Heterozygous or segregating DIA = Diaphorase ME = Malic enzyme PRX = Peroxidase PRO = Seed protein RBO = Rubisco SKD = Shikimic acid dehydrogenase GPL = Gene Pool 1 = Andean, 2 = Mesoamerican, 3 = Heterozygous GRO : Growth habit 1 : Determinate, 2 = Indeterminate. HET = Segregation fer morphological characters 0 = No, 1 = Yes. ' This data was collected by Martin (198“). Allozyme Data # FARM LIN SC ALZ DIA ME PRX PRO R80 SKD GPL 0R0 HET 1 1 1 1 1 1 2 2 1 2 2 1 2 0 2 1 1 2 1 1 2 2 1 2 2 1 2 0 3 1 3 1 1 1 2 2 1 2 2 1 2 0 “ 1 “ 2 1 1 2 2 1 2 2 1 2 0 5 1 5 3 2 1 2 2 1 2 1 1 2 0 6 1 6 3 2 1 2 2 1 2 1 1 2 0 7 1 7 1 1 1 2 2 1 2 2 1 2 0 8 1 8 2 1 1 2 2 1 2 2 1 2 0 9 1 9 2 3 1 1 1 1 2 2 1 2 0 1O 1 10 2 3 1 1 1 1 2 2 1 2 0 11 1 11 1 1 1 2 2 1 2 2 1 2 O 12 1 12 1 1 1 2 2 1 2 2 1 2 0 13 1 13 1 1 1 2 2 1 2 2 1 2 0 1“ 1 1“ 3 “ 1 1 2 1 2 2 1 2 0 15 1 15 3 2 1 2 2 1 2 1 1 2 0 195 196 Appendix A (cont'd.) # FARM LIN SC ALZ DIA ME PRX PRO RBO SKD GPL 0R0 HET 0000000000 2222222222 411111111121 2222222222 2222222222 1111111121 2222222212 2122222212 1111111111 1“11111151 13121211u1 0000000000000000000000000 2222222222222222222222122 2211122222222122221222222 2222122222222222222222222 2222122222222222222222222 24122222122222222222212222 1111211111111111111111111 1422611u111112111121u1111| 5333635333353355373533355 123u567890123u567890123u5 1111111111222222 2222222222222222222222222 67890123u567890123u567890 22223333333333uuuuuuuuuu5 0000000000000 2122222122222 2222222222222 2222222222222 1111111111111 2222222222222 2222222222222 1111111111111 1111111111111 5855333355855 4123.“.56789 10 11 12 13 3 3 33333333333 78 55 59 60 62 63 123.456 1 555555 6 197 Appendix A (cont'd.) # FARM LIN SC ALZ DIA ME PRX PRO RBO SKD GPL GRO HET 000000000010 222222222222 112111111111 221222222222 221222222222 112111111111 221222222222 121-222222222 112111111111 ”17111111111 359555555508 1 u.nasv7.nvo.n.1.9.24u.fia 1.1.1.1.1j1.¢.4~4.4aca¢ 24242424242424242424232. u.RJAv7.n.o,nv1.9.24u.RJ ,O,O,O,b,o,o.lqlq17a7.7. 0000100110000000011000010 2212222222222211122111122 1.9.1.9.9.1.1.9.9.2.9.9_9.1.1.1.1.9.241.1.1.1.9_9. 21.21122113111222213222211 12122112222221.11123111122 2121122113111222213222221 2111122111111222213122211 1212211222222111123111112 17u771177077711117.0”11187 3121333331111353333535331 1111 1111-11 11 11! 123u567890123hfi567890123u5 11.11111111222227— uuuuunuuunuuuuuuuuuuuuuu“. 67890123u56789o123u567890 7777888888888899999999990 1 00000000100 22222122222 22122111211 11211222122 22122122211 11211221122 11211222122 9.9.1.9.9_1.1.1.9.1f1 198 Appendix A (cont'd.) I FARM LIN SC ALZ DIA ME PRX PRO RBO SKD GPL GRO HET 12 13 111 15 16 112 113 1111 00100000100000 22222222222222 32121 11212 11212 22121. 11212 31212 150 5 5 1S 7 18 1 S 5 15 7 16 115 116 1 5 122121223 211212112 211212112 122121221 111112113 211212112 122121221 177171770 1 1| 611656516 111111111 7890123u5 111222222 555555555 7890123hfl5 111222222 111111111 1000000000000100000000000 2022222222222022222222222 2022222222222022222222222 2022222222222022222222222 2022222222222022222222222 1 11111111111 11111111111 3535335353533333533555355 123.“.567890123“.567890123n5 1.111111111222222 6666666666666666666666666 67890123u567890123u567890 22223333333333uuuuuuuuuu5 1111111111111111111111111 000000000 222222222 222221222 111111111 222222222 222222222 111111111 111112111 1 000051500 2222 2 22 123.456789 777777777 151 152 153 1511 155 156 157 158 159 199 Appendix A (cont'd.) # FARM LIN SC ALZ DIA ME PRX PRO RBO SKD GPL GRO HET 0000000000000000 1111111111111211 1111111111111211 2222222222222122 2222222222222122 1111111111111211 2222222222222122 2222222222222122 1111111111111711 0000050000050200 22222 22222 2222 0123u567890123u5 111-1111111222222 7.7.7.7.7.7.7.7.7.7.7.7.7.7.7.7. 00000000 22222222 22222122 11111211 11111211 22222122 11111211 11111211 ooooooooooooooooo 29.222221222222212 29.222221222222212 11111112111111121 9.2222221222222212 9.2222221222222212 7.7.7.7.7.7.7.1.7.7.7.7.7.7.7.1.7. 11111112111111121 1.1111111111111111! 90123u567890123u5 1111111111222222 88888888888888888 ”567890123u567890 88888899999999990 11111111111111112 000000 212211: 111131 222222 222232 222232 222232 111131 111101 ”sun—.35 222 123.".56 999999 201 202 203 20” 205 206 200 Appendix A (cont'd.) # FARM LIN SC ALZ DIA ME PRX PRO RBO SKD GPL GRO HET 0000000000010000000 2222212112212211212 1111111111132111111 2222222222221222222 2222222222221222222 111-11111111112111.1111! 2222222222221222222 2222222221231222222 1.1.1.1.1.1.1.1.1.1.1.1.9_1.1.1.1.1Z1 nuu5u5n55n503u55n2n 222 2 2 2 1.12 22412 7890123u567890123n5 1111111111222222 9999999999999999999 7890123u567890123u5 0001111111111222222 2222222222222222222 1111011010111001111111010 2222122222222222222222222 33334133333333133333333131 2233232333333222332333222 33311233322333223333323222 23314133131133.14112332331131 1312222332333223333333232 3222233132333233323133232 24133133241331.311-3333333131. 0000100000000100000000101 n96760785182688885755888n8 “-4.42“ 2.121“. 22 2.3.22 22 2 1233.567890123-4567890123“5 1111111111222222 0000000000000000000000000 11111111111111111111111111 67890.123u567890123u567890 22223333333333uuunnuuunu5 2222222222222222222222222 0000 2222 4141111 2222 2222 2222 2222 1111 1111 7uun 2333 123.4 11 11 11 11 251 252 253 25” 201 Appendix A (cont'd.) LIN SC ALZ DIA ME PRX PRO RBO SKD GPL GRO HET # FARM 000000000000000000000 212222222222222222222 222222222222222222222 222222222222222222222 222222222222222222222 222222222222222222222 111111111111111111111 7.nvnu7.n.7.7.7unuvu7.7.7.u.7.u.u.vuvuvunu 222232222222232332222 567890123n567890123u5 11.114111111222222 111111111111111111111 111111111111111111111 1011100111110111411111101000 2222222222222222222222222 33333133333333333333131.13 3323222233333333333322222 3233222233323332333323222 1n113313331333131333111111 3233223233323332233323222 3332223233323333333323223 31331113333333333333131141 0000010000000000000010110 2566333253622362222225222 1. nwu.9.9.9.1.9.9.n.1.1.9.u.1.1.1.1.1.1.9.1.1.1. 123.“567890123u567890123u5 1111111111222222 2222222222222222222222222 11111111‘1111111111111111 67890123u567890123u567890 7777888888888899999999990 2222222222222222222222223 26 2“ 13 13 301 302 202 Appendix A (cont'd.) # FARM LIN SC ALZ DIA HE PRX PRO RBO SKD GPL GRO HET 01000000000000000000000 22222222222222222222212 13111111111111111111311 23222222222222222222222 23222222222222222222322 13111111111111111111111 23222222222222222222222 23222222222222222222222 11111111111111111111111 10111111111111111111011 ususuzuzuuaunuéajuunuuzz 32323121322233223222311 3u567890123u567890123u5 1111111111222222 33333333333333333333333 11111111111111111111111 3&567890123n567890123u5 00000001111111111222222 33333333333333333333333 0000000000001001100000000 2222212122212221222122121 2222222232223222322222222 2222222222223222322222222 1111111111113111111111111 2222222232223222222222222 2322222232223223322122222 1311111131113111111111111 1011.111101110110011u11111 3333323233828582638288232 2222212122212 21u22122121 123”567890123n567890123u5 1111111111222222 ”an”nunuuuuuuuuuuunuuuunn 1111111111111111111111111 67890123u567890123u567890 22223333333333uuuuuuuuuu5 3333333333333333333333333 203 Appendix A (cont'd.) # FARM LIN SC ALZ DIA ME PRX PRO RBO SKD GPL GRO HET 0000001001000001000100000 2222222222222222222222222 2221222223222222222222222 2221222223222222222222222 2221222223222222222222222 2221222222222222222222221 111711111011111111111111” 5891877805895889758778085 21 22232 2 2 22 222 32 1.23“567890123u567890123u5 111-1111111222222 5555555555555555555555555 1.1111!11111111111111111111 123M“567890123hfl567890123u5 5555555556666666666777777- 3333333333333333333333333 APPENDIX B Infbrmation on Malawian Beans from an Interview with Sheila Mkandawire Mrs. Mkandawire was born in Bulawayo, Zimbabwe on March 31, 1956. Her parents are Malawians, and belong to a Tumbuka-speaking group of the northern region of Malawi. In her family when the daughters reached seven years of age they were sent back to Malawi to live with the a grandmother, at Ekwendeni in the Mzimba district. Sheila used to accompany her grandmother to the gardens and fields, and had some experience of the agricultural year. She reported that her grandmother would plant in February. One method of planting would be for the grandmother to hold maize and bean seed in her apron, make a planting hole, and then place three kernels of maize and two beans in each hole. The following information was supplied in response to questions about individual seed classes. Type 1 eg. 1203 Brown elongate Rumpi district "Nyauzembe", a Tumbuka name, where "Zembe" is a surname or family name, and "Nya is a form of address for women, like "Mrs." For example, Sheila said that in her husband's family she is called "NyaKonJe". This is a kind which is planted in February because it is early maturing, matures in 2 months. Very good, does not take long to cook. This is the kind grandmother planted in February. Sheila brought some of these with her from Malawi when she came to MSU, for cooking. It is in demand. It is hard to buy or to find in the planting season. Therefore it is kept right after harvest to save for seed. It is kept in ashes. Note: Type 2“, which looks quite a bit different, is also called "nyauzembe', and is the nyauzembe type which Greg Martin and Dr. Adams are familiar with. It was interesting to find that this Type 1 had the same name. Type 2 eg. 1202 Purple speck on cream Looked familiar, but Sheila did not know name. Grandmother did not have it in her mixture. Sheila saw it in the market. 20” 205 Type 3 eg. 1u05 Yellow elongate "Katolika", which means Catholic. Good tasting, but not as good as Type 1. Took long time to cook. Type 3b eg. 3506 Yellow elongate Described by the word for shiny, "biribwira". Type R eg. 122“ Cream elongate Sheila had seen a larger size of this type (see Type 29). Type 5 e8. 1307 Red mottle on cream 'Zamabanga", meaning spots. Planted with maize, not separately. Seeds taken from group (picked out from mixture for planting?). Had not heard it called sugar bean. Good tasting. Groundnuts can be added Just befbre the beans have finished cooking, and this makes a dish called "mandavya". Type 6 eg. 1305 Brown rounded Not familiar with it. Type 7 eg. 1318 Pink kidney Looked familiar, but did not know a name. Type 8 eg. 1u02 Large white kidney ”Kayera", meaning white. Not common in the north, more common in the Central region. See Type 11. Type 9 eg. 1u16 Pink rounded Not familiar with this. Type 10 eg. 1u2u Red mottle on yellow Anything with spots is termed ”Zamabanga'. Type 11 eg. 3102 White small round "Kayera", meaning white. This is a ChiChewa name, not a Tumbuka name. A common bean but it does not taste good, not as good as other types. Her grandmother would not plant it. It takes a long time to cook. People like it in the Central Region. Maybe it would be planted for sale. Type 12 e8. 3103 Red elongate/kidney, large (compare to Type 23) "Kamcheche", a Tumbuka name meaning red. It is good tasting. Famous. It makes a thick gravy. It is planted with maize. It is 206 cooked with maize, in a dish called ”nkhove", where the maize is pounded, cooked, and then the beans are added to it. Because the seeds are big they were called ”za mungoma musi" by her grandmother, where "mungoma” means maize, and "musi" means under. These take longer to mature and were therefore planted with maize. They are very common. Type 13 eg. h519 Brown mottle on cream This is a familiar type. Called by the generic term "zamabanga", meaning spots. Type 15 eg. 3u02 Black eye on gray Does not know this one. Type 16 eg. 3&11 Brown stripe on cream "Mabanga ya chimbwe", meaning hyena. Good tasting. Planted separately in its own part of the field. Kept for seed. Her grandmother did plant this type. [Notez "Mabanga" may be term far markings, so that "mabanga ya chimbwe' and ”zamabanga” both use it as a root.] Type 17 - eg. 3&09 Black stripe on grey Did not know this one. Type 18 eg. 3H1” Orange stripe on yellow Had seen this but did not know of a particular name for it. Type 20 eg. “109 Red stripe on cream "Mabanga ya chimbwe", meaning hyena (see Type 16). "Mabanga" for short. Type 21 _ Light purple elongate eg. “106 Not seen befbre. Type 22 eg. #123 Shiny red round "Kamcheche" because it is red (see Type 12). Good tasting, but it takes a long time to cook. Type 23 eg. 5505 Slender red kidney Type 2” eg. h507 Opaque green round Looks like “nyauzembe”, same group as seen before (See Type 1). Many of these where Alex comes from. His mother likes them. 207 Type 25 eg. ”522 Purple on tan elongate Type 26 eg. 6117 Red pot on white ('Jacob's cattle') "Zamabanga", meaning spots. Has never tasted this, and her grandmother never planted it, but she has seen it in the market. Type 27 eg. 5323 Purple mottle on cream "Zamabanga", meaning spots. Type 28 eg. 511” Opaque red round/oval "Kamcheche", meaning red. Type 29 eg. 6503 Khaki oblong, bullet-shaped Says that this is the large kind of Type u. "Kapenta" was the name used by her grandmother, who had it in her mixture. It is tasty, and is planted separately in a part of the field devoted to it alone. It is planted separately because the seed is saved fOr the next year. This method of planting is used with some beans because if planted with maize they do not grow well, and they are ”less happy". Type 30 eg. 6509 Yellow round Not familiar with it. Type 32 e8. 5513 Bright purple kidney Does not know a name. Has seen in in the markets, but her grandmother did not plant it. Type 3a eg. 6115 Round pale purple Does not know it. "‘111111911111111111111111“