LIBRARY Michigan State University This is to certify that the thesis entitled BREEDING FOR RESISTANCE TO BEAN COMMON MOSAIC NECROSIS VIRUS AND MOLECULAR TAGGING OF bc—3 GENE IN COMMON BEAN presented by Gérardine Mukeshimana has been accepted towards fulfillment of the requirements for the MS. degree in Plant breeding and genetics- Crop and Soil Sciences ,/ ' //é’/<2{ James D. Kelly Major Professor’s Signature 7’ 02 _, 0 3 Date MSU is an Affirmative Action/Equal Opportunity Institution PLACE IN RETURN BOX to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 6/01 c:/CIRC/DateDue.p65-p.15 BREEDING FOR RESISTANCE TO BEAN COMMON MOSAIC NECROSIS VIRUS AND MOLECULAR TAGGING OF bc-3 GENE IN COMMON BEAN By Gérardine Mukeshimana A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Plant Breeding and Genetics Program-Department of Crop and Soil Sciences 2003 ABSTRACT BREEDING FOR RESISTANCE TO BEAN COMMON MOSAIC NECROSIS VIRUS AND MOLECULAR TAGGING OF bc-3 GENE IN COMMON BEAN By Gérardine Mukeshimana Studies were conducted to assess the disease resistance present in local Rwandan bean varieties, transfer genetic resistance to bean common mosaic necrosis virus (BCMNV) into susceptible Rwandan bean varieties by backcrossing, and search for molecular markers linked to the bc-3 resistance gene using RAPD and AFLP techniques. Using phenotypic and marker data many Rwandan varieties appeared to possess a number of resistance genes suggesting that these varieties can constitute valuable sources of resistance in breeding for disease resistance. The bc-3 gene for BCMNV resistance was introgressed into eight Rwandan varieties using the backcross breeding procedures. BC1F2 populations were screened for the presence of the bc-3 gene and resistant plants were used as male parents to produce the BCzF1 generation. The RAPD technique failed to generate a useful marker linked to the bc-3 gene. AFLP analysis proved to be more powerful than RAPDs enabling the finding of a codominant EACAMCGG169/172 marker linked at 4.4cM from the bc-3 gene. The EACAMCGG marker was transformed into a sequence tagged site (STS) marker that proved to be present in diverse group of genotypes carrying the bc-3 gene and should provide bean breeders with a ready tool to introgress the bc—3 gene into new bean cultivars. In memory of my parents. iii ACKNOWLEDGEMENTS I am grateful to many individuals and organizations for their generous support and contribution to the completion Of this study. First of all, I would like to express my gratitude and sincere thanks to Dr. James D. Kelly, my major professor for his unlimited support and encouragement throughout my study. Without his insightful advice, patience, and endless support, this dissertation could not have been completed within the set time frame. Special gratitude is extended to Dr. Rebecca Grumet and Dr. Eunice F. Foster for being members of my guidance committee. I would also like to thanks dry bean breeding and genetics lab group for their friendship and help during my time spent at MSU. ’ To Dr. Freed Russell, thank you so much for kindly making available your computer that I used to write my dissertation. I will always remember my friends Gaudence Kayitesirwa, Judith Nyiraneza, Jean Claude Bizimana, Belinda Roman, and Halima Awale for their inestimable friendship. I am deeply grateful for the Partnership for Enhancing Agriculture in Rwanda through Linkages (PEARL) project for the financial support. iv Finally, my deepest love and thanks go to my husband Adelit Nsabimana, our daughter Karen Teta Umubyeyi, and my Sister Beata Mukabera for their constant love, support, encouragement, and sacrifices Showed during the hard time of my absence. TABLE OF CONTENTS LIST OF TABLES ................................................................................. viii LIST OF FIGURES ................................................................................. x Chapter 1. Literature review .................................................................. 1 Beans in Africa ...................................................................................... 1 Bean common mosaic virus (BCMV) and been common mosaic necrosis virus (BCMN V) ............................................................................................. 5 Pathogen variability of BCMV and BCMN V ................................................. 7 BCMN V ............................................................................................ 10 BCM V .............................................................................................. 1 1 Resistance genes ................................................................................. 12 Dominant resistance gene ..................................................................... 12 Recessive resistance genes ................................................................... 14 Genetic markers .................................................................................. 18 Types of genetic markers ...................................................................... 19 Morphological markers ........................................................................ 19 Biochemical markers ............................................................................ 19 RFLP markers .................................................................................... 20 RAPD markers .................................................................................... 21 AFLP markers ..................................................................................... 22 Simple sequence repeat (SSRs) and single nucleotide polymorphism (SNPs)...26 Selection efficiency of indirect selection .................................................... 26 Resistance genes pyramiding ................................................................ 28 Chapter 2. Assessing the disease resistance in local Rwandan bean varieties using a combination of direct and indirect screening methods and transfer of BCMNV resistance into susceptible Rwandan bean cultivars by backcrossing .................................................................................... 30 Introduction ........................................................................................ 30 Materials and methods .......................................................................... 35 Phenotypic and molecular evaluation of Rwandan cultivars for disease resistance .......................................................................................... 35 BCMN V and Anthracnose evaluation ....................................................... 37 Marker evaluation ................................................................................. 38 Transferring resistance to BCMNV by introgressing bc-3 gene into Rwandan cultivars using the backcross method ...................................................... 41 Results and discussion ........................................................................ 46 vi Phenotypic evaluation .......................................................................... 46 Marker evaluation ................................................................................ 48 The transfer of the bc—3 gene ................................................................. 54 Conclusion ........................................................................................ 60 Chapter 3. Search for a new marker linked to bc-3 gene .......................... 62 Introduction ......................................................................................... 62 Materials and methods .......................................................................... 64 Mapping population .............................................................................. 64 Identification of a marker linked to bc-3 gene ............................................. 66 RAPD analysis .................................................................................... 66 AFLP analysis ..................................................................................... 67 Sequence tagged sites (8 TS) marker development ..................................... 69 Segregation analysis ............................................................................. 71 Linkage analysis .................................................................................. 71 Results .............................................................................................. 72 RAPD analysis .................................................................................... 72 AFLP analysrs ................ 73 Discussion .......................................................................................... 78 Conclusion .......................................................................................... 82 Appendix ........................................................................................... 84 References ....................................................................................... 85 vii LIST OF TABLES Table 1. Disease and Abiotic constraints affecting beans in East Africa ............ 4 Table 2. Interaction of BCMV and BCMNV pathogenicity groups, strains, and pathogenicity genes with beans genotypes .................................................. 9 Table 3. Characteristics often Rwandan bean cultivars ............................... 36 Table 4. Description of molecular markers linked to different disease resistance genes in common bean ......................................................................... 40 Table 5. Characteristic of the donor parents used in the backcross programs...43 Table 6. Crossing block between Rwandan genotypes and US. lines ............ 43 Table 7. Phenotypic reaction of Rwandan cultivars to races 7, 73, and 256 of C. lindemuthianum and NL3 strain of BCMNV ................................................ 47 Table 8. Molecular markers linked to genes conditioning resistance to anthracnose in Rwandan cultivars ........................................................... 49 Table 9. Molecular markers linked to the genes and quantitative trait loci (QTL) conditioning resistance to angular leaf spot (Phg 2), bean common mosaic necrosis virus (I), common bacterial blight (QTL), and rust (Ur-3 and Ur-5) in ten Rwandan bean cultivars ....................................................... . ................. 52 Table 10. Observed reactions in BC1F2 populations after inoculation with NL-3 strain of BCMNV .................................................................................. 55 Table 11. Resistant vs susceptible, segregation ratios of BC1F2 populations, goodness of fit to expected ratios, and p value ....................................... 56 Table 12. Observed reactions in BC1F2 populations after inoculation with NL-3 strain of BCMNV ................................................................................. 56 Table 13. Resistant vs susceptible, segregation ratios of BC1F2 populations, goodness of fit to expected ratios, and p value ....................................... 57 Table 14. Pattern of the primers that were polymorphic between both parents..71 Table 15. Chi-square analysis of observed ratios for bc-3 gene and the AFLP marker EACAMcsg169/172 segregating for resistant (R) and susceptible (S) in F2 viii generation and for homozygous resistant (RR), heterozygous susceptible (Rr), and homozygous susceptible (rr) for F23 families in RWK10/USCR-7 population .......................................................................................... 74 Table 16. Survey of genotypes carrying bc-3 gene for presence of SEACAMCGG marker .............................................................................................. 77 ix LIST OF FIGURES Figure 1. General organization of a potyvirus genome ................................... 5 Figure 2. Backcrossing Scheme ................ 45 Figure 3. Development of RWK10/USCR-7 mapping population ..................... 65 Figure 4. AFLP gel showing the polymorphism generated by the primer combination EACAMCGG159/172 .................................................................. 75 Figure 5. PCR product obtained using the SEACAMCGG134/137 primer. .............. 76 Figure 6. Sequence of the resistant AFLP fragment ..................................... 84 Chapter 1. Literature review. Beans in Africa Common bean (Phaseolus vulgaris L.) is one of the few crops with a broad range of adaptation to the most varied geographic regions. Its distribution extends to all continents except Antarctica with excellent adaptation to diverse conditions generated by differences in elevation, latitude and diverse farming systems. Common bean is the most important grain legume for direct human consumption in Africa and Latin America where it provides a cheap source of protein, calories, vitamins, fiber, and minerals for low-income populations (Jacobsen, 1999; Broughton et al., 2003). In Africa, beans represent the most important pulse crop consumed (Allen et al., 1989) and they constitute a source of dietary protein for more than 70 million people (Graham and Ranalli, 1997). Bean production is concentrated on small scale farms in many African countries, where they are grown in complex associations of bean mixtures with other crops such as plantains (Musa sp.), maize (Zea mays), and other cereals (Pachico, 1989). In the case of maize-bean associations, favorable relationships have been reported where the bean uses the maize as stakes and in return, maize benefits from the N2 fixed by the bean (Graham and Ranalli, 1997). In addition, bean-maize associations have been shown to lessen disease severity in both crops by impeding spores dispersal, and altered microclimates favorable for disease development (Pachico, 1989). The highest bean production is concentrated in the great lakes region of East Africa and southern part of Africa (Pachico, 1989). Despite the high bean production in this part of the African continent, bean yields are low so the production lags far behind population growth (FAOSTAT, 2003) as bean growers are unable to satisfy the demand. Bean cultivars with higher yield, multiple disease resistance, and acceptable level of tolerance to drought and low soil fertility provide the most effective way to enhance productivity and insure yield stability in the region. In Rwanda, beans are grown by 95% of the farmers and they are central to the farm economy and people’s diet (lSAR, 2001). Both bush and climbing beans are grown in Rwanda depending on altitude. At high altitudes between 2000 and 2300 m with high rainfall and fertile soils, climbing beans are more prevalent. The main reason for such intensive cultivation of climbing beans is the need to intensify the production in parallel to the growing population density of the country (Allen et al., 1989). At lower and medium altitudes between 900 and 2000 m, the complex associations of several bush bean varieties are more common (lSAR, 2001). The variety mixtures provide small farmers with assurance of yield stability under low input conditions by buffering against unfavorable environmental conditions caused by variable microclimates and diseases (Pachico, 1989; Berti, 1985). Per capita consumption of beans is over 60 kg per year in Rwanda, one of the highest rates in the world, and beans provide 32% and 65% of all calorie and protein intake (lSAR, 2001). Beans are consumed in five forms in Rwanda: tender green pods, tender seeds before they dry, dry seeds, beans without seed coats, and tender leaves as vegetables (Nyabyenda et al., 1980). The annual production of around 300,000 tons (lSAR, 2001) does not satisfy the national demand and Rwanda is obligated to import beans from neighboring countries. The main factors that limit Rwandan bean production include the limited land for production, high acidity soils that lack important nutrients such as nitrogen, phosphorus, and potassium, drought, and bean diseases. Acidity, and low soil N and P availability can potentially be alleviated with the use of inorganic and organic fertilizers and planting of specific varieties efficient in nutrient use (lSAR, 2001). However, bean diseases constitute the most important constraint to bean production in Rwanda. The warm and humid environmental conditions that prevail in the tropical and subtropical regions favor pathogen development. Moreover, the planting of 2-3 crop cycles per year on the same land in the absence of crop rotation provides a continuity of disease inoculum, and seed borne diseases constitute the major problem because of the use of farmer’s seed for planting. The most important bean diseases constraining bean production in Rwanda and across Eastern African region in general include angular leaf spot (Phaeoisan'opsis griseola), anthracnose (Colletotn’chum lindemuthianum), bean viruses, common bacterial blight (Xanthomonas axonopodis pv. phaseolr), and rust (Uromyces appendiculatus) (Table 1). Table 1. Disease and abiotic constraints affecting beans in East Africa Constraints Hectares % Acreage affected affected Angular leaf spot (Phaeoisan’opsis gn'seola) 2,567,000 67 Low nitrogen 2,015,000 53 Anthracnose (Colletotn'chum lindemuthianum) 1,782,000 47 Bean fly (Ophiyomyia phaseoli) 1,681,000 46 P-deficiency 1 ,667,000 46 Bean Common Mosaic Virus 772,000 22 Aphids (Aphis fabae) 613,000 18 Common bacterial blight (Xanthomonas 607,000 18 axonopodis) ' Rust (Uromyces appendiculatus) 600,000 17 Source: Buruchara et al., 1998. Among the most serious diseases that attack beans in Rwanda, bean common mosaic virus (BCMV) and bean common mosaic necrosis virus (BCMNV) are the most important seed borne viral diseases that constrain bean production with possibility of 100% yield loss depending on the incidence of aphids vectors (ASARECA, 2002) Bean common mosaic vims (BCM V) and been common mosaic necrosis virus (BCMNV) BCMV and BCMNV belong to the potyvirus group, which is the largest group of plant viruses. The potyvirus group contains many other economically important viruses, such as potato virus Y (PW), tobacco etch virus (T EV), zucchini yellow mosaic virus (ZYMV), plum poxvirus (PPV), soybean mosaic virus (SMV), blackeye cowpea aphid-bome mosaic viruses (BICMV and CAM), and watermelon mosaic virus (WMV). Potyviruses are filamentous viruses with a linear, unipartite single—stranded RNA of approximately 10 kb that is terminated by a genome—linked protein (VPg) at the 5’ terminal and a poly (A) tail at the 3’ end (Figure 1). The genome contains one continuous open reading frame (ORF) that is translated into a large protein, which is subsequently cleaved into mature polypeptides by virus encoded proteases P1, HC-PRO, and Nla (Fang et al., 1995; Revers et al., 1999). VPg .- 1 HC-Pro P3 CI Nla Nlb CP Poly (A) (pro) (pro) (?) (hel) (yPg/pro) (rep) (coat) ~10 kb RNA Figure 1. General organization of a potyvirus genome. Poly (A): polyadenyl at the 3’ end, CP: coat protein, Nla and Nlb: nuclear inclusion a and b, Cl: cylindrical inclusion, P1 and P3: protease, HC-Pro: helper component protease, VPg: virus-encoded genome linked protein linked to the 5' end, prozprotease, hel: helicase, rep: replicase. BCMV and BCMNV are perhaps the most common and destructive viruses known to naturally infect common bean worldwide (Wang, 1983; Drijfhout, 1978). Those two viruses are seed borne and they cause serious losses in subsistence farming systems where farmers do not have access to disease-free seed. The economic damage caused by these viruses consists of severe reduction of crop yield and the quality of harvested seeds. BCMV and BCMNV have been reported in many parts of the world including Europe (Drijfhout, 1978), U. S. (Kelly et al., 1983, Prowidenti et al., 1984), and Africa (Silbemagel et al., 1986; Spence and Walkey, 1995; Sengooba et al., 1997; Njau and Lyimo, 2000). Seed transmission of these viruses depends on the stage of growth of the host at the time of infection, cultivar, virus strain, and sometimes temperature (Drijfhout, 1978). When plants are infected before flowering, the rate of infection can reach 95% (Galvez and Morales, 1989). If infection occurs after flowering, the virus does not usually reach the seed (Drijfhout, 1978; Morales, 1989). In addition to the seed-bome nature of these viruses, BCMV and BCMNV are transmitted by several aphid species in a non-persistent manner. From studies conducted at ClAT, relatively high aphid population can cause 100% plant infection from a source of only 2% to 6% infected seed (Galvez and Morales, 1989). Two viral proteins, helper component and coat protein, encoded by their respective genes in the viral genomes are involved in aphid transmission (Revers et al., 1999; Pirone and Blanc, 1996). The helper component assists in transmission of the virus by acting as a “bridge” between aphids and virus particles (Pirone and Blanc, 1996). Specifically, the helper component assures the virion retention in the mouthparts of the aphid from which it can subsequently be injected into the plant (Pirone and Blanc, 1996). The capsid protein seems to facilitate the interaction of the virion with the helper component as mutations in the capsid protein resulted in loss of aphid transmissibility (Pirone and Blanc, 1996) Similar to other viral diseases, chemical control of BCMV and BCMNV is ineffective. The most effective way to prevent bean plants from becoming infected with BCMV and BCMNV is the production of virus-free seed stocks (Hart and Saettler, 1981) and the use of genetic resistance (Drijfhout, 1978). Among the methods available to prevent virus infection in common bean, genetic resistance is the most cost effective and durable (Drijfhout, 1978; Kelly et al., 1995; Miklas et al., 1998). In East Africa where BCMNV is found in wild legumes, the only effective way to control these viruses is to plant resistant varieties, otherwise the bean plants would be reinfected from alternate hosts that might be growing adjacent to the bean field (Coyne et al., 2003). Pathogen variability of BCMV and BCMNV Extensive pathogenic variability exists in BCMV and BCMNV as showed by the large number of different strains reported throughout the world. Strains such as NY15, Idaho, Florida, and Western were reported in the U. S. (Dean and Hungerford, 1946; Dean and Wilson, 1959; Zaumeyer and Goth, 1964), Vordagsen, Marienau, Westlandia, lmmuna, Michelite, Great Northern, Jolanda, Colana, NL7, and NL8 in Europe (Frandsen, 1952; Hubbeling, 1963; Drijfhout and Bos, 1977), Mexican strain in Mexico (Silbemagel, 1969), Peru strain in Peru (Gamez et al., 1970), Costa Rica strain in Costa Rica (Moreno et al., 1968), Puerto Rico strain in Puerto Rico (Alconero et al., 1972), and TN strain in Tanzania (Silbemagel et al., 1986). After studying the pathogenicity spectra of all these strains, Drijfhout (1978) showed that certain strains were identical to isolates of other strains. He grouped the strains based on reaction to nine differential cultivars and proposed the international nomenclature for virus strains which is two letters identifying the country in which they were reported followed by the number indicating their sequence of discovery in that country. In addition, differences between strains have also been demonstrated by serological (Wang, 1983) and molecular studies (Berger et al., 1997). Based on the symptoms induced by the virus strains on nine host groups, Drijfhout et al. (1977) grouped BCMV strains into seven pathogenicity groups (I to VII) (Table 2). All these pathogenicity groups were considered to be specific to BCMV, but recent work based on serological interactions, and coat protein sequence data demonstrated that the pathogenicity groups Should be Classified into both BCMV and BCMNV vinIses (Wang, 1983; Vetten et al., 1992; McKem et al., 1992; Mink and Silbemagel, 1992; Khan et al., 1993; Berger et al., 1997). BCMV and BCMNV are also reorganized as serotype B and serotype A, respectively (McKem et al., 1992) based on their reaction to monoclonal antibodies (Wang et al., 1983). BCMNV differs from BCMV in that it has a shorter particle and a smaller capsid protein compared to BCMV. Cells infected with BCMNV strains show a specific type of proliferated endoplasmic reticulum, which is not found in cells infected with BCMV strains (Vetten et al., 1992; McKem et al., 1992). Each pathogenicity group comprises many virus strains depending on the way the pathogenicity genes that they carry react with host resistance genes in the gene for gene model to produce disease (Table 2). Table 2. Interaction of BCMV and BCMNV pathogenicity groups, strains, and pathogenicity genes with bean genotypes in the differential series. Pathogenicity groups, virus strains, and pathogenicity genes I“ II III IV V VI VII NL6, NL2, Host genes NL1§ NL7 NL8 035 032 NL5, NL3 NL4 , 2 P1, P12, P1, P12, P0 P1 P2 P1, P1 P1, P2 P2 P22 I, - - TN - - TN - ., + + + + + + + II bc-1bc-1 ' + ' i + + + bc-12 bc-12 ' ‘ + ‘ i + bc-2 bc-2 ' ' + ' + + ' bc-22 bc-22 ' ' ‘ ‘ ' ' * bc-3 bc-3 ' ' ' ' ' ' Source: Drijfhout, 1978. The presence of bc-u is assumed. TN: top necrosis, +: mosaic, -: non reaction; #: Pathogenicity groups, §: strains, 1': pathogenicity genes. BCMNV The common strains of BCMNV NL3, NL5, and NL8 were first reported by Drijfhout (1978). In addition to these strains, Silbemagel et al. (1986) reported a Tanzanian strain (TN) that behaved pathologically and serologically exactly as other strains of BCMNV. The NL8 strain belongs to the pathogenicity group III and carries the P2 pathogenicity gene, while pathogenicity group VI comprises the NL3 and NL5 strains that possess the P1, P12, and P2 pathogenicity genes (Drijfhout et al., 1978; Table 2). The NL3 and NL5 strains differ in their reaction to specific bean cultivars. ln Amanda, NL5 strain was more effective than NL3 in inducing systemic necrosis (Drijfhout, 1978). Berger et al. (1997) suggested that NL8 might diverge slightly from other strains of BCMNV at the 3’-end non-coding region. In studying the variation of pathogenicity among isolates of BCMV in Africa, Spence and Walkey (1995) showed that more than 53% of BCMV isolates collected from P. vulgaris were the NL3 strain. The presence of serotype A with predominance of the NL3 strain in wild legume species in Africa, and the tolerance of some of these species to the NL3 strain suggests that BCMNV evolved in the African continent probably as a result of recombination among BCMV strains and local strains (Spence and Walkey, 1995). BCMNV strains were introduced more recently into Europe and the US. through trade of infected seeds of susceptible bean varieties. The common feature of BCMNV strains is that they cause a hypersensitive reaction in bean genotypes carrying the dominant I resistance gene regardless of the ambient temperature. Strains of BCMNV are therefore known as temperature insensitive necrosis inducing (TINI) 10 strains. Symptoms of the hypersensitive reaction begin as small, red-brown spots on inoculated leaves that expand into a vascular necrosis. Response compounds in an infected plant expand into the phloem tissues causing wilting, death of young leaves, and eventually death of the whole plant. TINI strains are predominant in Central, Eastern, and Southern Africa where they cause severe yield losses in the bean crops due to systemic necrosis (Spence and Walkey, 1995; Sengooba et al., 1997). In East Africa, the effect of TINI strains has been so serious that local bean programs are reluctant to release bean varieties carrying the I gene because of the potential crop losses that indigenous farmers could suffer (ASARECA, 1998). BCMV Common mosaic inducing strains, or serotype B of BCMV probably originated and co-evolved with bean in Central and South America (Spence and Walkey, 1995). BCMV and BCMNV are very complex in their interactions with different host genotypes. In susceptible genotypes, symptoms appear as a light— green or yellow and dark green mosaic pattern usually accompanied by puckering, distortion, and rolling of the leaves. These symptoms are the result of systemic infection of the plant by the virus. The most common strains of BCMV include NL1, NL2, NL4, NL6, NL7, U82, and U85, of which NL2 and NL6 are temperature dependent and induce necrosis only at high temperatures (Drijfhout, 1978). NL1 classified as belonging to pathogenicity group I with the pathogenicity gene (P0) can only infect bean genotypes lacking resistance genes. The NL2 11 and U82 strains belong to pathogenicity group V and have pathogenicity genes P1 and P2. NL4 is in the pathogenicity group Vll possessing P1, P12 and P22 genes; NL6 and USS belong to group IV and possess P1 and P12 genes, and NL7 is classified in pathogenicity group II and only possesses the P1 pathogenicity gene. Based on virus surveys in the US. and Africa, strains of both BCMV and BCMNV can sometimes occur in the same area and occasionally in the same plant (Spence and Walkey, 1995; Silbemagel et al., 1986; Sengooba et al., 1997; Hampton et al., 1983; Njau and Lyimo, 2000), suggesting the opportunity for possible recombination and formation of new virus strains (Revers et al., 1999; Silbemagel et al., 2001 ). Resistance genes Two types Of genes conditioning resistance to BCMV and BCMNV have been identified (Ali, 1950; Drijfhout, 1978) and they are referred to as dominant and recessive resistance genes. These types of resistance genes have distinctly different mechanisms of resistance. The dominant resistant gene operates as a classical hypersensitive resistance gene whereas the recessive genes are constitutively expressed and act by restricting virus replication or movement within the plant (Kelly et al., 1995; Kelly, 1997; Kelly et al., 2003). Dominant resistance gene The dominant form Of resistance to BCMNV and BCMV is conferred by the inhibitor I gene (Ali, 1950). The dominant I gene confers a typical hypersensitive 12 resistance and has been a source of resistance against a wide range of BCMV strains for over half a century (Kelly, 1997). However, in the presence of TINI strains, the I gene is a liability because these strains stimulate a systemic hypersensitive response resulting in a vascular necrosis and the death of the plant, a phenomena known as top necrosis or black root in the field. As a result of the hypersensitive reaction, the I gene prevents the spread of the virus to neighboring healthy plants and the production of infected seed. The same lgene genotypes infected with common mosaic inducing strains, exhibit a typical resistant reaction without necrosis or systemic mosaic. The I gene is frequently found in Central American black beans (Beebe and Pastor Corrales, 1991 ). The I gene has been mapped to the linkage group 82 where it is tightly linked to the seed coat color intensifying gene 8 (Temple and Morales, 1986; Freyre et al., 1998). The B gene enhances the darkening Of testa color in Mesoamerican beans (Beebe and Pastor Corrales, 1991). The linkage between the B and l genes has prevented the widespread deployment of the I resistance gene in breeding programs where Mesoamerican bean varieties with a light colored seed coat are preferred (Temple and Morales, 1986). Fortunately, the lgene in certain Andean genetic background does not appear to be associated with a color intensifying gene. For example, some beans of the cranberry market class carry the I gene and have the red mottled seed color preferred in commerce. The presence of the I gene in some bean cultivars in Central and Eastern Africa where BCMNV strains exist has been such a serious problem that breeders in those regions select against the I gene (Silbemagel et al., 1986). Another 13 alternative is to protect the I gene by incorporating into new cultivars a series Of recessive genes that confer resistance to the specific necrotic inducing strains of BCMNV (ASARECA, 1998). Recessive resistance genes A large number of genes conditioning resistance to potyviruses are recessive in their mode of action. For example, among all known potyvirus resistance genes, 40% are recessive whereas only 20% are recessive genes for other virus groups (Fraser, 1992). Recessive resistance has been demonstrated to be effective and long lasting (Fraser, 1992; Johansen et al., 2001, Harrison, 2002). While little is known about the nature of resistance conferred by recessive genes, two hypotheses could explain the role of recessive genes in virus resistance (Revers et al., 1999): 1) The resistant host may lack a host function essential for particular steps in viral pathogenesis, and consequently the dominant allele encodes a host factor, which is required for the virus to replicate and/ or move in the susceptible host; 2) The susceptibility allele encodes a dominant negative regulator of resistance. These hypotheses may explain why vimlence against recessive resistance is rare since it is unlikely that a virus could mutate to overcome a missing or defective host function. Three strain specific recessive loci bc-1, bc-12, bc-2, bc-22 and bc-3 control resistance to BCMV and BCMNV in common bean (Drijfhout, 1978). The bc-1 and bc-2 loci are independent loci and are allelic to bc-12 and bc-22, respectively (Drijfhout, 1978). The strain specific recessive genes interact with virus pathogenicity genes in the 14 gene for gene model (Drijfhout, 1978). The forth-recessive locus conditioning resistance to these viruses is a strain unspecific bc-u gene. The bc-u gene is necessary for the full expression of all strain specific recessive genes (Drijfhout, 1978). With the exception of the bc-2 locus, the other recessive resistance genes have been mapped to independent loci: bc-12 allele has been mapped to the linkage group 83 (Kelly et al., 2003) and this suggests that bc-u also resides on B3 due to its loose linkage with bc-1 (Strausbaugh et al., 1999). The bc-3 gene has been mapped to B6 linkage group (Gepts, 1999; Kelly et al., 2003). Breeders could exploit the independence of these resistance genes to develop cultivars with pyramided resistance as a strategy in breeding for durable resistance to BCMV and BCMNV. BCMV and BCMNV resistance genes show epistatic interactions when plants are inoculated with the NL3 strain of BCMNV: bC-3 masks bc-22 and bc-12; bc-2 2 masks bc-12. The bc-3 is epistatic to the I gene while all other recessive genes are hypostatic to the I gene (Kelly et al., 1995). Because of the epistatic interaction between bc-3 and I, the effect of the I gene cannot be detected phenotypically in the presence of the bc-3 gene. This prevents breeders from distinguishing and selecting the desirable Ibo-3 recombinants from the single bc- 3 genotypes (Kelly, 1997). Bean breeders must use molecular markers linked to the I gene (Haley et al., 1994b; Melotto et al., 1996) to select those genotypes carrying the I and bc-3 gene combination. The bc-3 gene originated from Pl 181954 accession. Analysis of the seed protein of this accession revealed the ‘8’ type phaseolin suggesting that this 15 gene is Middle American in origin (Johnson and Gepts, 1994). The bc-3 gene confers resistance to all known strains of BCMNV and BCMV depending on the state of the I gene background. In a recessive igene background, bc-3 requires the presence of bc-u to be fully expressed while bc-u is not needed in a dominant lgene background. In fact, Miklas et al. (1998) demonstrated that bean cultivars with Ibo-3 and bc-ubc-3 genotypes were immune to all strains of BCMV and BCMNV while genotypes ibc—3 were resistant to all strains of BCMNV but susceptible to some strains of BCMV from pathogroups I, II, and IV. Pyramiding genes for resistance to BCMV and BCMNV into single cultivar has proven to be the best strategy to achieve durable resistance to these viruses (Kelly et al., 1995). The combination of the dominant hypersensitive I gene with strain specific recessive genes should provide a long lasting resistance given that these genes possess different modes of action. The I gene confers a typical hypersensitive resistance while the recessive genes act by restricting virus movement within the plant (Kelly, 1997). The combination of the I gene and recessive genes offers a good strategy for protecting bean plants against the TINI strains as the plant is protected against the hypersensitive reaction. The use of these gene combinations to improve resistance to BCMV and BCMNV in bean cultivars has been and Should continue to be one of the objectives in breeding for resistance to these Viruses (Kelly et al., 1994). As many of the genes that confer resistance to BCMV and BCMNV are recessive in nature (Drijfhout, 1978), their introgression through traditional breeding is laborious and time consuming because of the need for progeny tests 16 to identify the carriers of the desired genes. Moreover, these resistant genes show epistatic interactions, which prevent direct selection for hypostatic genes. Testcrosses are required to identify many useful resistance genes in combination with other hypostatic genes. The use of indirect screening methods with molecular markers linked to the genes of interest would facilitate their introgression into a number of been cultivars susceptible to BCMV and BCMNV. The use of molecular markers tightly linked to resistance genes to assist in breeding for resistance is known as marker-assisted selection (MAS). The recessive gene bc-3 is particularly useful in areas where BCMNV is prevalent Since it confers resistance to all known strains of BCMNV and many strains of BCMV (Miklas et al., 1998). The use of MAS to transfer this gene into susceptible genotypes would greatly facilitate the breeding for resistance to BCMV and BCMNV especially in areas where breeders are reluctant to introduce necrotic strains of BCMNV for screening purposes. A number of DNA markers linked to the bc-3 gene have been found. Haley et al. (1994a) identified two Random Amplified Polymorphism DNA (RAPD) markers linked to the bc-3 gene in near- isogenic lines (NlLs) of common bean: OAD19690 (1 .9:l: 1.4 cM) linked in coupling and 0813660 (7.11: 2.6 CM) linked in repulsion phase. In addition, Johnson et al. (1997) developed OC11350,420 and OC20460 RAPD markers linked to the bc-3 gene that were converted later into SCAR markers to improve their utilization. The use of these RAPD and SCAR markers in MAS has been limited due to a lack of reproducibility and consistency among diverse gene pools of common bean. Direct screening with strains of BCMV and BCMNV is still required to 17 confirm the presence of the bc-3 gene. To efficiently introgress the bc-3 gene for resistance to BCMV and BCMNV into susceptible bean cultivars, there is a need to discover a more robust DNA marker linked to bc-3 that can be used in diverse genetic backgrounds and displays consistent amplification across laboratories. Genetic markers Genetic markers are heritable entities that are associated with economically important traits (Staub et al., 1996). Plant breeders can use these genetic markers to facilitate cultivars’ identification, the determination of genetic similarities among breeding stocks, calculation of level of polymorphism, heterozygosity or cross-pollination rates (Staub et al., 1996; Masojc’, 2002). However, the main expectation with respect to genetic markers is their potential use as indirect selection tools in MAS. Markers tightly linked to economically important traits that are under the control of single genes have the potential for immediate utility in plant improvement programs. Breeding for disease resistance is one of the main objectives of plant breeding programs. Markers tightly linked to disease resistance genes allow selection of the gene indirectly Since the expression of molecular markers is not masked by epistatic interaction that occurs between resistance genes. Moreover, markers for disease resistance offer the advantage of selecting for resistance in the absence of the pathogen or a variant of the pathogen, and importantly, MAS offers breeders a viable approach to developing cultivars with pyramided resistance. When screening conditions are not ideal or when working with 18 mixture of a pathogen races, MAS provides a precise way to ascertain the presence of the desired gene. MAS has shown to provide a gain in selection time of about two generations (Hospital et al., 1992). Types of genetic markers There are several types of genetic markers such as morphological, biochemical, and molecular or DNA markers such as Restriction Fragment Length Polymorphisms (RFLP), RAPD, Amplified Fragment Length Polymorphisms (AFLP), Simple Sequence Repeat (SSRs) and Single Nucleotide Polymorphism (SNPS). Morphological markers Morphological markers are those controlled by a single locus. The morphological markers can be used as genetic markers if their expression is reproducible over a range of environments. Morphological markers are usually dominant and they may have negative pleiotrophic effects. The fact that morphological markers can be influenced by environment and genetic interaction (epistasis) limits their usefulness in indirect selection. Biochemical markers Biochemical markers represented by isozymes were used extensively in crop characterization (Gepts, 1988). lsozymes are allelic variants of enzymes that are differently charged and can be separated using electrophoretic 19 procedures (Staub et al., 1996). The visualization from a specific enzyme represents protein products. lsozymes are codominant markers. Despite the fact that isozymes are limited in number and are species, tissue, and developmental stage specific (MaCMillin, 1983), they provided a good alternative to overcome the limitation of morphological markers (T anksley and Jones, 1981). lsozymes have been a useful genetic marker in studies of seed storage proteins in bean and have provided valuable insight in the classification of the gene pools of been (Gepts, 1988). RFLP markers RFLP markers are detected by scoring DNA fragments generated by hybridization of restriction length fragments produced by different restriction enzymes. RFLP markers are co dominant and abundant in most crops. RFLP are markers of choice in genetic mapping, classification, and cloning. RFLP markers are important as anchor points in a single high-density map developed from linkage maps established in different populations and laboratories. The use of RFLP clones facilitated the construction of a core genetic linkage map in common bean (Freyre et al., 1998) and the integration of genetic linkage and physical chromosomal maps in common bean (Pedrosa et al., 2003). The use of RFLP in indirect selection has been limited because of costly and sophisticated techniques involved in their generation and utilization. 20 RAPD markers RAPD markers are generated by PCR amplification of random genomic DNA segments with single primers of arbitrary sequence. RAPD found many applications in diversity studies (Beebe et al., 2000), mapping studies (Freyre et al., 1998), and MAS (Kelly, 1995) in common bean. RAPD markers are dominant and they possess several advantages in genetic mapping and gene tagging: 1) a universal set of primers can be used and screened in a short period of time, 2) no isolation of cloned DNA probes or preparation of hybridization filters is required, 3) only small amounts of DNA are needed. Moreover, RAPD markers are valuable in the construction of intraspecific genetic maps and in gene tagging research. Compared to RFLP markers, RAPD are more useful for detecting polymorphism within gene pools in self-pollinated species (Kelly, 1995). RAPD markers combine the advantages of being rapid and cost effective. As an extension of single gene tagging, linkages between RAPD markers and quantitative trait loci (QTL) controlling complex traits have also been developed (Schneider et al., 1997; Kolkman and Kelly, 2003). Concerns with reproducibility in the development of RAPD markers across laboratories have been reported (Jones et al., 1997; Weeden et al., 1992). The variability lies in the choice of thermal—stable DNA polymerase and thermal cycler used in PCR amplification, or in imprecise matches between Short oligonucleotide primers (decamer) and the template DNA at the low annealing temperatures (35-40 °C) typical of these studies (He et al., 1994; MacPherson et al., 1993; Meunier et al., 1993). To 21 address these concerns, sequence characterized amplified region (SCAR) and allele-specific associated primers (ASAPs) have been developed (Kesseli et al., 1992; Paran and Michelmore, 1993; Weeden et al., 1992). Reproducibility is increased by sequencing the two ends of the linked RAPD fragment and synthesizing two longer primers (~24 base pairs) homologous to each end. These two primers are used in the same PCR protocol, but at elevated annealing temperatures (50-65 °C), and generally produce a single fragment equivalent to that previously sequenced. The majority of SCAR markers amplify single major bands of the same size as the original RAPD. Since only one band is amplified, the electrophoresis step can be eliminated and visualizing directly the amplified DNA under UV light by staining with ethidium bromide (Melotto et al., 1996). The presence or absence of the amplified DNA indicates the presence or absence Of the linked resistance gene. SCAR markers are the markers of choice for indirect selection in beans (Melotto et al., 1996; Melotto et al., 1998; Johnson et al., 1997; Awale and Kelly, 2001; Sartorato et al., 1999, Vallejo and Kelly, 2001). AFLP markers AFLP markers (Vos et al., 1995) combine some features of RFLP and RAPD technologies. This technique is based on the PCR amplification of the restriction fragments resulting from the digestion of genomic DNA. The principle steps are: 1) restriction of DNA and ligation Of oligonucleotide adapters 2) amplification of sets of restriction fragments, and 3) gel analysis of the amplified fragments. The restriction fragments are generated by a combination of two 22 restriction enzymes, an infrequent cutter and a frequent cutter. The resulting fragments terminate with rare cutter sequences at one end and the frequent cutter at the other end. The advantage of using both enzymes is to reduce the number of fragments to be separated on the denaturing gel given that only rare lfrequent cutter fragments are amplified and this limits the number of selective nucleotides needed for selective amplification. The combination of ECORI and Msel is used in many laboratories. Msel is preferred as a four bases cutter because its recognition sequence (TTAA) is frequent in eukaryotic (AT-rich) genomes. Consequently, it cuts frequently in most eukaryotic genomes yielding fragments of optimal size range for both PCR amplification and separation on polyacrylamide gels. ECORI is a reliable six-cutter enzyme and it has shown the advantage of limiting the problems associated with partial restriction (Vos et al., 1995). Pst is being used more recently to provide fragments in regions of active transcription, as Pst does not cut methylated sites. The double stranded adapters ligated to the ends of the restriction fragments help in creating sites for primer annealing in fragment amplification. AFLP primers consist of three parts: the 5’ part corresponding to the adapter, the restriction site sequence and the 3’ selective nucleotides. One to three selective nucleotides at the 3’ end provide an effective way to select the desired number of fragments for amplification (Vos et al., 1995) but three selective 3’ nucleotides are more common (Hazen et al., 2002; Vuylsteke et al., 2000). The DNA amplification requires two steps of amplification in the AFLP: First the genomic DNA is pre-selectively amplified with AFLP primers both having a single selective nucleotide. The pre-selective 23 amplification step has the purpose of reducing the smear backgrounds on the gel, and of course provides unlimited amount of template DNA for AFLP reactions. The pre-selective amplification products are then diluted and used as template in the selective amplification. Selective amplification is carried out using primer combinations having one to three selective nucleotides. Some advantages of AFLP are that prior knowledge of the genomic sequence and the construction of CDNA library are not needed (Bradeen and Simon, 1998). AFLP technique generates fingerprints of any DNA regardless of origin or complexity. Compared to RAPD, AFLP technique is more robust and reliable (Jones at al., 1997) because more stringent reaction conditions are used for primer annealing (Vos et al., 1995). In addition AFLP is insensitive to DNA concentration given that only variations are seen on level of intensity of the bands and the difference in thermocyclers does not affect the band pattern (Vos et al., 1995). However, AFLP is sensitive to restriction conditions given that incomplete restriction of DNA may result in partial fragments with the consequence that difference in band patterns detected do not reflect true DNA polymorphism (Vos et al., 1995). AFLP is a powerful tool to reveal restriction fragment polymorphisms (markers) and it is a suitable technique for several genetic studies. AFLP markers were used in construction of genetic maps in different crops such as sweet potato (lpomoea batatas) (Kriegner et al., 2003) and common bean (Tar’an et al., 2002), estimation of genetic diversity in wheat (Triticum aestivum) (Hazen et al., 2002) and maize (Vuylsteke et al., 2000), identification of DNA markers linked to important traits in different crops such as 24 disease resistance in cowpea (Vigna unguiculata) (Ouédraogo, et al., 2001), wheat (Guo et al., 2003), common bean (T ullu et al., 2003; Kolkman and Kelly , 2003), and soybean (Glycine max) (Meksem et al., 2001; Mienie et al., 2002), identification of markers linked to sex expression in asparagus (Asparagus Officinalis) (Reamon-Btittner et al., 1998), and the identification of markers linked to physical and chemical components in common bean (Guzman - Maldonado et al., 2003). AFLP markers are dominant (Bradeen and Simon, 1998) in nature and their polymorphisms are usually generated SNPS, insertions or deletions, and SSRS (Meksem et al., 2001; Hazen et al., 2002). Because of the high sensitivity and level of resolution for difference in band size, the AFLP technique is able to detect small insertion/deletion type mutations of only few bases. AFLP generate hundreds of informative genetic markers with a high number of polymorphic loci (Jones et al., 1997). AFLP markers have received limited use in MAS because of its technical complexity, relatively high cost of sample preparation, use of polyacrylamide gel electrophoresis, and marker detection (Prins et al., 2001). The conversion of AFLP markers into sequence tagged Site (STS) markers, which are technically simpler, should enhance their use in MAS. The conversion of AFLP markers into STS facilitated identification of the sex trait in Asparagus officinalis (Reamon- Btittner, 2000) and breeding for Fusan'um head blight resistance in wheat (Guo et al., 2003). 25 Simple sequence repeat (SSRS) and single nucleotide polymorphism (SNPS). The microsatelites or simple sequence repeats (SSR) are highly mutable loci that may be present at many sites in the genome (Jones et al., 1997). This class of markers is hypervariable, codominant and detects a high level of allelic variation associated with economic traits. Comparing the degree of information content (expected heterozygozity) of RAPD, AFLP, RFLP, and SSR markers, the SSR markers proved to contain the highest expected heterozygosity (Wayne et al., 1996; Pejic et al., 1998). SSR markers also are highly reproducible across laboratories (Jones at al., 1997). Single nucleotide polymorphisms (SNPS) are mostly used in human genome studies (Landergren et al., 1998). Currently SNPS are also gaining popularity in crop research mostly in mapping studies and genetic diversity studies due to its precision in genotyping. SSRS (Yu et al., 2000) and SNPS (Melloto and Kelly, 2001) markers have received limited use in common bean breeding as few have been developed probably due to the sequencing costs associated with their generation. Selection efficiency of indirect selection The efficiency of indirect selection depends on the nature of the marker, linkage distance, the orientation of marker linkages, and breeding method used. The successful use of MAS requires tight linkages (<5 0 M) between markers and genes of interest or the use of two loosely linked markers flanking the desirable locus (Tanksley, 1983). A dominant marker directly linked to the susceptibility 26 allele allows selection against susceptibility. This type of linkage increases the selection efficiency by allowing the selection against the heterozygote and this is particularly useful when selecting for recessive genes. For the identification of homozygous resistant bc-3/bc-3 individuals, Haley et al. (1994a) demonstrated 82% selection efficiency by selecting against a susceptibility-linked marker versus 26% efficiency in selecting a marker linked to the resistance. Markers flanking the resistance gene significantly increase selection efficiency of resistant individuals when loosely linked (>5 cM) markers are used. The selection efficiency was increased from 94% to 99% when two flanking markers were used to select for the 00-2 gene for anthracnose resistance in common bean (Young and Kelly, 1996). The information generated by a pair of markers, one linked in coupling and other in repulsion to the target locus, is genetically equivalent to that obtained from a single codominant marker. Selection efficiencies of 19%, and 77% were obtained using coupling and repulsion markers linked to homozygous resistant Ur-11/Ur-11 gene, respectively and it increased to 83% using both markers as codominant pair (Johnson et al., 1995). A codominant pair of flanking markers increases selection efficiency since they combine the advantages of flanking markers with the ability to identify heterozygotes. The selection efficiency was increased from 33% for the coupling-phase markers to 92% for the repulsion phase markers to 95% for the co dominant flanking markers in the selection of Co-6/Co-6 homozygote individuals for resistance to anthracnose in common bean (Young and Kelly, 1997). 27 Resistance genes p yramiding Breeding for durable pest resistance challenges both plant breeders and pathologists because highly variable plant pathogens and insects do not remain static. As plant breeders work with variable plant pathogens, constant changes lead to increased variability and complexity among plant pathogens that can be controlled by different genes. Even though race-specific genes are recognized to be less durable, pyramiding genes based on their complementarities offers a good Opportunity to achieve more durable disease resistance (Duvick, 1996). Obtaining stable rust resistance in beans necessitated pyramiding three genes Up-2, B-190, and Ur-3 against 63 of the 65 bean rust races characterized by USDA- ARS in the US. (Kelly et al., 1994). The landrace Colorado de Teopisca (62333) from Chiapas, Mexico that carries the CO-42, 00-5, and 00-7 genes conditioning resistance to anthracnose in common bean is highly resistant to more than 380 isolates of C. lindemuthianum (Young et al., 1998). The most durable approach to achieve resistance to all strains of BCMV and BCMNV is through the combination of I bc--12 bc-22 bc—3 genes (Drijfhout, 1978). Currently, the resistance to all known strains of BCMV and BCMNV is achieved by combining I gene and bc-3 genes. The dominant I gene is hypostatic to bc-3 and the chance of loosing the I gene through selection is increased. Linked markers offer the only realistic way to maintain and continue to utilize the lgene in gene pyramids in future cultivars. Pyramiding race-specific genes has been successful in controlling stem rust (Puccinia graminis) in spring wheat in North America 28 since 1955 because contemporary spring wheat cultivars possess as many as six resistance genes (Schafer and Roelfs, 1985). The Objectives of the present study were to: (1) Assess disease resistance in local Rwandan bean varieties using a combination of direct and indirect screening methods, (2) transfer resistance to BCMNV into Rwandan bean varieties through backcrossing with US. lines known to carry the bc—3 gene, and (3) initiate a search for a more robust DNA marker linked to the bc-3 gene in common bean. 29 Chapter 2. Assessing disease resistance In local Rwandan bean varieties using a combination of direct and indirect screening methods and transferring BCMNV resistance into susceptible Rwandan bean cultivars by backcrossing Introduction Bean diseases are one of the most important factors contributing to low yields in most been producing regions of the world. Multiple pathogens have been reported to attack beans and reduce yield and quality (Beebe and Pastor Corrales, 1991). Some diseases such as anthracnose (Colletotn'chum lindemuthianum), angular leaf spot (Phaeoisan’opsis gn‘seola), mst (Uromyces appendiculatus), common bacterial blight (Xanthomonas axonopodis pv. phaseoli), bean golden mosaic virus (BGMV), and bean common mosaic virus (BCMV) are distributed worldwide while others important diseases such as web blight (Thanatephonrs cucumen's) and white mold (Sclerotinia sclerotiorum) are restricted to specific bean growing regions where environmental conditions favor their development (Beebe and Pastor Corrales, 1991). Bean diseases constitute a major problem for bean production in Rwanda. The most important bean diseases constraining bean production in Rwanda and across the East African region in general include angular leaf spot (ALS), anthracnose, BCMV, BCMNV, common bacterial blight (C83) and nIst (Table 1). 3O Phaeor’san'opsis griseola, the causal agent of ALS has shown to have an evolutionary pattern similar to that of the common bean host (Pastor-Corrales et al., 1998). The Andean isolates of P. griseola have a narrower virulence range while the Middle American isolates are capable of attacking beans from both Andean and Middle American gene pools (Pastor-Corrales et al., 1998). Since P. griseola is a highly variable fungus, breeders need to employ diverse sources of resistance (Busogoro et al., 1999a). The diversity of this pathogen in Central Africa may have resulted from selective effects of bean mixtures grown in the region (Busogoro et al., 1999b). Resistance to the ALS has been shown to be qualitatively inherited and conditioned by both dominant and recessive genes (Busogoro et al., 1999b; Ferreira et al., 2000; Nietsche et al., 2000). Breeding strategies for resistance to ALS should employ the intercrossing of Andean and Middle American gene pools and pyramiding resistance genes from different genetic sources (Nietsche et al., 2000). Bean anthracnose caused by C. lindemuthianum is a destmctive seed borne disease of common bean (Beebe and Pastor Corrales, 1991). In tropical and subtropical areas of Latin America, Central and East Africa where farmers save their own seed, anthracnose is a particular constraint of been production (Pastor Corrales and Tu, 1989). In these areas environmental conditions favor the development of the pathogen. ln Rwanda and throughout East Africa, anthracnose is regarded as the second major biotic cause of yield loss after ALS (lSAR, 2001). 31 BCMV and BCMNV are economically important constraints of bean production in different parts of the world (Drijfhout et al., 1978; Kelly et al., 1983; Prowidenti et al., 1984; Spence and Walkey, 1995). In Eastern and Central Africa, yield losses due to BCMV are estimated at 772,000 tones per year. BCMNV is a particular threat to bean production in this region because of the prevalence of subsistence agriculture as farmers save their own seed for the next growing season, increasing the chance of planting seed infected with virus. For example, a field survey conducted by ASARECA (2002) Showed that all fields surveyed had these viruses and in some fields the incidence was as high as 100%. As BCMV and BCMNV symptoms mimic soil nutrients deficiency, many farmers in Central and East Africa are unaware of the problem. Developing varieties with increased levels of vims resistance constitutes a good strategy for increasing bean production and enhancing economic opportunities for growers. Many BCMV resistant varieties used in Rwanda were developed by the Centro lntemacional de Agriculture Tropical (CIAT) in Colombia. Many of these varieties carry the dominant I gene which is a liability with the widespread distribution of BCMNV strains in Eastern Africa (Spence and Walkey, 1995). In order to prevent the loss of the crop due to systemic necrosis known as black root that result in the death of bean plants, national breeding programs are reluctant to release varieties carrying the lgene. Currently, the bc-3 gene which confers resistance to all known strains of BCMNV and many BCMV strains (Drijfhout, 1978; Miklas et al., 1998) is highly desirable. The incorporation of this gene into Rwandan material would minimize the damage of these two viruses in local bean cultivars. 32 Moreover, the availability of adapted cultivars carrying this gene would be a source of bc-3 gene for improving future varieties for local production. Common bacterial blight (CBB) caused by X. axonopodis pv. phaseoli is a widely distributed bacterial disease of common bean. C88 is a seed-bome disease and its negative effect on both yield and quality of beans is serious in a country such as Rwanda where farmers save seed for next growing season. Genetic resistance for CBB is quantitatively inherited and its expression differs over environments making breeding for resistance a challenge. Bean rust caused by U. appendiculatus fungus results in serious yield losses of bean crop due to premature defoliation. The fast spread of the disease is favored by the windblown nature Of the spores between and within bean fields. U. appendiculatus is a very highly variable fungus suggesting the risk of the loss of resistance when deploying single major resistance genes. The use of pyramided resistance genes in a single variety would be the best strategy when breeding for rust resistance. Resistance genes conditioning resistance to rust appear to be clustered in common bean (Stavely, 1984). For example the resistance to individual rust races in the bean genotype B190 (Ur-5 locus) is conditioned by single dominant genes that appear to be inherited as a complex linkage block (Stavely, 1984). Integrated disease management is an important strategy to control bean diseases in the U. S. (Schwartz and Peairs, 1999), whereas the majority of been improvement programs especially in developing countries use plant resistance aS the most important method of the disease management because it is the most 33 practical and easy to adopt. Thus, breeding for disease resistance is the main goal for all plant breeding programs in developing countries such as Rwanda. Selection of the resistant parents carrying resistance to a given pathogen is the first step in disease resistance breeding. One of the methods to identify resistant parents to a given pathogen is to expose the potential genotypes to the pathogen variability and evaluate the host reaction. In certain cases direct evaluation is hampered by the lack of appropriate pathogen strains and unfavorable screening conditions. Alternatively, breeders may choose to use indirect screening using molecular markers linked to resistance genes to identify genotypes that are potential carriers of desired genes. The evaluation of Rwandan bean varieties for markers linked to resistance genes is important due to the lack of opportunity to screen directly due to the unavailability of certain pathogens and races of different pathogens in the U. S. Information gained about the resistance genes carried by these varieties and the identification of potential sources of resistance genes could be used to improve local commercial varieties. The objectives of this section of the study were to: 1) assess the disease resistance in local Rwandan bean varieties using a combination of direct and indirect methods 2) transfer resistance to BCMNV into susceptible varieties from Rwanda by the backcross method. 34 Materials and methods Phenotypic and molecular evaluation of Rwandan cultivars for disease resistance All analyses were conducted on 10 elite Rwandan cultivars: RWR1802, RWV524, RWR13121, RWK10, NGZZ4-4, CAB19, SCAM800M/15, RWV167, RAB487, and G2331 released in 2001 by Rwanda’s Institute of Agronomic Sciences (ISAR). Cultivars RAB487, CAB19, G2331, and SCAM800M/15 originated from CIAT and were tested and released in Rwanda by ISAR while the other cultivars were developed by ISAR (Table 3). Characteristics such as seed color, seed size, growth habit, photoperiod sensitivity, and days to flower were recorded prior to the study. Seed color was determined by visual observation of the seed coat color. The seed size was determined by weight of 100 seeds. The growth habit was recorded as determinate or indeterminate based on the absence or presence of the terminal reproductive bud. For photoperiod response, the seeds were planted in the greenhouse (16h and 25 °C) and the plants were evaluated for flowering after a period of three months. Photoperiod sensitive cultivars did not flower while photoperiod insensitive plants flowered and produced seeds. Days to flower were number of days after which 50% of the plants were flowering. For photoperiod sensitive genotypes, the days to flowering were determined as described above but flowering was later induced by covering the plants with a black cloth to exclude the light. 35 Table 3. Characteristics of Ten Rwandan Bean Cultivars Gene Growth Photo Days to Varieties pool+ Seed color Seed Size” habit++ period I flower CAB19* MA White 27.4 lnd. PS 50 c2331I MA Yellow 34.8 Ind. PS 65 RAB4871 MA Red 26.8 Det. PI 39 NC 224-4 A Black-white 45.9 Ind. PS 59 RWK10 A Red-white 62.9 Det. PI 29 RWR 1802 A Yellow 49.7 Det. PI 30 RWR1312| A Red-white 49.3 Det. PS 30 RWV167 MA Red 26.6 Ind. PS 65 RWV 524 A Purple-white 48.3 Ind. PS 65 SCAM80CM/1 5* MA Red-white 40.3 Det. PI 39 *CIAT origin; 7 weight of 100 seeds in g; i A: Andean gene ooI, MA: Middle America gene pool; ” lnd.: lndeterrninate, Det.: determinate; PS: Photoperiod sensitive, Pl: Photoperiod insensitive (under conditions of East Lansing latitude: 43 °N). 36 BCMNV and Anthracnose evaluation For BCMNV screening, the NL3 strain was used. This strain was found to be more prevalent in Rwanda (Spence and Walkey, 1995). NL3 strain was isolated from the cultivar ‘Michelite’ identified and catalogued in the pathogenicity group Vla by Drijfhout (1978). Virus was maintained on the universal susceptible cultivar ‘Sutter Pink’ in a screened greenhouse. Young leaves with mosaic symptoms collected from the universal susceptible cultivar ‘Sutter Pink’ were used as the inoculum source of NL3 strain. lnoculum solution was prepared by grinding young leaves in a mortar, homogenized with carborundum powder and 10mM phosphate buffer (3 mM K2PO4, 7 mM NaOzHPO4) pH 7.2. Six seedlings per pot from each genotype plus two controls, Sutter Pink and Black Hawk that, when inoculated with NL3, developed mosaic, and lethal systemic necrosis respectively (Drijfhout, 1978) were grown in the greenhouse (25 °C and 16 h photoperiod). Nine days after planting, seedlings were rub-inoculated with the viral homogenate at primary leaf stage, and the two primary leaves were inoculated. Disease reaction was recorded at 7 days and 21 days. For anthracnose screening, Rwandan cultivars and checks were grown in the flats containing Baccto planting mix (Michigan Peat 00., Houston, Texas) and 6 nine days old seedlings of each cultivar were inoculated with races 7, 73, and 256 of anthracnose as described by Balardin et al. (1998). Races 7 and 73 were from Central and North America and they were used because they are known to be more common and to have a wide spread geographic distribution (Balardin et al., 1997). Race 7 is of Andean origin and it is virulent to genotypes carrying the 37 Co-1 resistant gene such as Michigan Dark Red Kidney (MDRK) and Montcalm (Balardin et al., 1997). The Middle American race 73 is virulent to genotypes possessing both Co-2 and 00-3 resistant genes (Balardin et al., 1997). Race 256 is a Middle American race virulent to genotypes carrying 004 gene such as T0 (Balardin et al., 1997). Race 256 was chosen because this isolate came from Kenya (Kelly, personal communication) suggesting that it may be present in Rwanda. Disease reaction was recorded one week after inoculation. Cultivars showing any lesions on stems or leaves were considered as susceptible and varieties with clean foliage and stems were recorded as resistant. Marker evaluation Leaf DNA was used for indirect screening of the presence of disease resistance genes. Leaf samples were collected from seedlings of the 10 Rwandan cultivars grown in the greenhouse. The DNA was extracted using the miniprep procedure described by Afanador et al. (1993). The extracted DNA was quantified using a fluorometer (Hoefer TKO100, Hoefer scientific, San Francisco, CA) and standardized to uniform concentration (10 ng/pl) by dilution. Amplification reactions were made of 17.85 ul H20, 3 pl Of template DNA (10 nglul), 3 pl Of primer (10 ng/pl, Integrated INA Technologies, Coralville, IA), 1X PCR buffer, 5 mM of MgClz, 2 U of Taq DNA polymerase, and 0.5 mM of each dNTP (lnvitrogen, Carsbad, CA) for SCAR and STS markers tested (Table 4). The amplification of DNA was performed by polymerase chain reaction (PCR) using a Perkin Elmer Cetus DNA Thermal cycler 480. The PCR files were as suggested by the developers (Melotto et al., 1996; Melotto et al., 1998; Awale 38 and Kelly, 2001; Vallejo and Kelly, 2001; Vallejo and Kelly, 2002; Sartorato et al., 1999). The amplified DNA was run on 1.4 % agarose gel containing 0.5 pg. ml'1 ethidium bromide, 40 m M Tri-acetate, and 1mM EDTA and photographed under ultraviolet light for SCAR markers and on 6 % polyacrylamide gel with silver staining method (Promega, WI). 39 Table 4. Description of molecular markers linked to different disease resistance genes in common bean. Diseases Markers Size (bp) Resistance Reference Genes Anthracnose SE ACTMCCA 108 Co-12 Vallejo and Kelly, 2002 Anthracnose SAS13 950 00-42 Melotto and Kelly, 1 998 Anthracnose SH18 1150 00-42 Awale and Kelly, 2001 Anthracnose SBB14 1150/1050 Co-42 Awale and Kelly, 2001 Anthracnose SAB3 450 00-5 Vallejo and Kelly, 2001 Anthracnose 8812 350 00-9 Mendez de Vigo et al., 2002 ALS# SN02 890 Pth Nietsche et al., 2000 BCMV“ sw 13 690 I Melotto et al., 1996 CBB" SAP6 820 Major QTL“ Ariyarathne et al., 1 999 CBB" BC490 1250 Major QTL” Ariyarathne et al., 1 999 Rust Sl19 460 Ur-5 Melotto and Kelly, 1 998 Rust SK14 620 Ur-3 Nemchinova and Stavely, 1998 # ALS: Angular leaf spot, BCMV: Bean common mosaic virus, CBB: Common bacterial blight, i QTL: Quantitative trait loci. All molecular markers tested were linked in coupling with resistance genes. 40 Transferring resistance to BCMN V by introgressing bc-3 gene into Rwandan cultivars using the backcross method The genetic material used in this study consisted of 8 BCMNV susceptible Rwandan cultivars: RWR1802, RWV524, RWR1312I, RWK10, NG224-4, CAB19, SCAM80CM/15, and RAB487. Among these cultivars RWV524, CAB19, and NG 224-4 were photoperiod sensitive under long day conditions in Michigan. These genotypes were induced to flower by placing them under short day conditions (12 hours) in the greenhouse by covering them with a black cloth to exclude light along with RWV167 and 62331 recommended for Rwandan highland regions (>2300m). RWV167 and G2331 were highly photoperiod sensitive, which prevented flower initiation and they were not used in the backcross program. When inoculated with NL3 strain of BCMNV, three Rwandan cultivars (RWR1802, RWV524, NG224-4) showed systemic mosaic symptoms suggesting that they do not possess any resistance genes for BCMNV, two genotypes (RAB487, and RWK10) showed top necrosis suggesting that they carried the I gene. The SW13 marker linked to the lgene was used to confirm the presence of I gene in these cultivars. The other varieties (SCAM18CM/15, CAB19, and RWR1312l) had a mild mosaic reaction suggesting that they carried the bc-12 resistance gene. None of the Rwandan cultivars showed a complete resistance (no-reaction) after inoculation, indicating the need to introgress the bc-3 gene into these bean varieties. The Rwandan cultivars were used as recurrent parents in the backcross process to introgress bc-3 gene. Four U.S. breeding lines USDK-4, USWK-6, USCR-7, and l99530 (Table 5) served as donor parents for the bc-3 gene. USDK—4, USWK-6, and USCR-7 lines were developed and 41 released by USDA Agriculture Research Service in collaboration with Washington State University (Miklas et al., 2002a, 2002b). Those lines combined bc-3 and l genes conferring resistance to all known strains of BCMV and BCMNV (Miklas et al., 2002a, 2002b). The resistance reactions of these donor parents to BCMV and BCMNV were confirmed before crossing by inoculation with the NL3 strain of BCMNV. The presence of the I gene was also confirmed in all three parents by the presence of SW13 marker linked to I gene. The breeding line l99530 was a small red derived from PR9357-107 (Beaver et al., 1998) that carried the bc-3 gene conferring resistance to BCMNV in the absence of I gene. The absence of I gene was confirmed by the no reaction upon the inoculation with NL3 strain and the absence of the SW13 SCAR marker linked to the lgene. The initial crosses were made in the greenhouse (25°C and 16h photoperiod) between BCMNV susceptible Rwandan cultivars and US. cultivars possessing the bc-3 gene as the donor parents. Crosses were designed between the parents from the same gene pool having similar seed coat pattern and size (Table 6). This crossing design was used to prevent lethality of F1 Middle American/Andean hybrids (Haley et al., 1994d) and to reduce the time to recover commercial traits of the recurrent parents. 42 Table 5. Characteristics Of the donor parents used in the backcross programs. Genotypes Gene Seed Seed Days Resistance genes to pool’ color size“ to flower BCMNV USWK-6 A White 58 33 Ibo-3 USDK—4 A Red 54 33 Ibo-3 Ibc-3 USCR-7 A Red- 52 30 white l99530 MA Red 27 37 bc-3 * A: Andean gene pool, MA: Middle America gene pool; ” weight of 100 seeds in 9; all donor parents were photoperiod insensitive. - Table 6. Crossing block between Rwandan genotypes and US lines. Male parents Female parents l99530 USDK-4 USWK-6 USCR-7 CAB1 9 X NG224-4 X RAB487 X RWK10 X RWR1 802 X RWR1 3121 X RWV524 X SCAMBOCM/15 X X: Actual crosses made. 43 After the initial crosses, the F1 populations were backcrossed to Rwandan cultivars to produce the BC1F1. The BC1F1 should segregate on a 1:1 ratio with 50 % BC-3BC-3 and 50 % BC—3bc-3. Phenotypically, all the BC1F1 progeny were susceptible. BC1F1 plants were self-pollinated for progeny testing and to fix the recessive gene in the BC1F2 generation that segregated 1 resistant to 7 susceptible (1:7). Since the desired gene could be assessed before flowering, BCIF2 seed were planted and the seedlings were inoculated with NL3 strain. The inoculation conditions were as described earlier except that in this case all BC1F2 plants were rub-inoculated with the virus homogenate instead of a sample of six seedlings. The resistant plants from this inoculation were backcrossed to the recurrent parents to produce the BC2F1 generation. The BC2F1 was the last generation produced in Michigan and they will be further backcrossed to the recurrent parents to produce BC3F1 generation, These individuals will be self- pollinated and the resulting BC3 F2 generations will be inoculated with the NL3 strain. The resistant plants from the inoculations will continue to be self-pollinated to produce BCaFa progeny for field trial evaluation of traits similar to those of the recurrent parents. The summary of the backcross program is shown in figure 2. § Rwandans lines X US lines —> Initial crosses in the greenhouse l §F1 X RP —> First backcross initiation l §BC1F1 __> Plants from the first backcross selfed €19 §BC1 F2 x RP—> BC1F2 subjected to rub inoculation with NL3 Resistant plants crossed to the recurrent parents BC2F1 X RP_, Subsequent crosses to the recurrent parents l BC3F1 _> Plants selfed for the progeny test lea BC3F2 —> Inoculation with NL3 and self-pollination lea BC3F3 —’ Resistant lines ready for field trials Figure 2. Backcrossing Scheme. 69 = Self-pollination, RP= Recurrent parent= Rwandan lines, §. Flowering induced under short days. 45 Results and discussion Phenotypic evaluation Genotypes RWV167 and 62331 were resistant to all anthracnose races tested while RWR1802 and RWV524 were resistant to both race 7 and 73 (Table 7). The resistance of RWV167, 62331, RWR1802, and RWV524 to both Andean and Middle American isolates suggests the presence of durable resistance to anthracnose in these varieties (Balardin and Kelly, 1998). Varieties NG224-4 and RWK10 were susceptible to race 7 but resistant to both 73 and 256 races while CAB19 was only resistant to race 7. RWR1312| was resistant to race 73 and susceptible to both races 7 and 256. Variety RAB487 was heterogeneous to races 7 and 73 and resistant to 256 whereas SCAM80CM/15 was susceptible to race 7, resistant to 73, and heterogeneous to race 256. Rwandan cultivars were grouped into three categories based on their reaction to NL3 strain of BCMNV: G2331, NGZZ4-4, RWR1802, RWV167, and RWV524 Showed typical mosaic symptoms suggesting that these genotypes do not possess any gene conferring resistance to BCMV and BCMNV. Cultivars CAB19, RWR1312l, and SCAM8OCM/15 showed mild mosaic reaction which is a partial resistance conditioned by the bc-12 gene (Kelly, 1997). RAB487 and RWK10 expressed top necrosis reaction, which resulted in the death of the plants. This reaction was indicative of the presence of I gene in these varieties. None of the Rwandan bean varieties expressed a no-reaction to NL3 indicating 46 the absence of the bc-3 gene and this suggested the need to introduce this valuable gene into Rwandan cultivars. Table 7. Phenotypic reaction of Rwandan bean cultivars to races 7, 73, and 256 of C. lindemuthianum and NL3 strain of BCMNV. Disease reaction Anth racnose races # BCMNV§ Varieties 7 73 256 CAB1 9 R S S MM (32331 R R R M NG224-4 S R R M RAB487 R/S R/S R TN RWK1 0 S R R TN RWR1 31 2| S R S MM RWR1 802 R R S M RWV167 R R R M RWV524 R R S M SCAMSOCM/‘I 5 S R R/S MM 5 BCMNV: Bean Common Mosaic Necrosis Virus; M: Mosaic (susceptible systemic mosaic symptom indicating the lack of resistance genes); MM: Mild Mosaic (partial resistance conferred by the presence of bc-12 gene); TN: Top Necrosis (hypersensitive resistance due to the presence of the lgene); # R: Resistant, R/S: heterogeneous. S: susceptible. 47 Marker evaluation Six makers were tested to detect the nature of resistance to anthracnose in the ten Rwandan cultivars (Table 8). None of the Rwandan cultivars appeared to carry the $812 marker linked to 00-9 gene (Mendez de Vigo et al., 2002). The cultivar RWV167 possessed the SCAR markers linked to 00-5 and 00-42 genes and the STS SEACTMCCA marker linked to the 00-12 allele. The presence of those markers linked to important genes conferring resistance to anthracnose may explain the high phenotypic resistance of RWV167 to anthracnose races tested since each of those genes confers resistance to races 7, 73, and 256. The Co-12 allele was found in Kaboon and demonstrated to confer resistance to races 7 and 73 (Melotto and Kelly, 2000). The CO-5 gene is an important gene in anthracnose resistance and it has been shown to confer resistance to many races including 7, 73, and 256 races (Balardin et al., 1997). The allele Col-42 has also been found in 62333 variety and SEL1308 line (Young et al., 1998) and is recognized as the most broadly based resistance gene since no known race of anthracnose has overcome this resistance (Balardin and Kelly, 1998, Melotto and Kelly, 2001; Young et al., 1998). The presence of CO42 allele may confer broad resistance to many anthracnose races including 7, 73, and 256 races in RWV167 cultivar. The combination of 00-12, 00-42, and 00-5 genes in RWV167 should confer useful resistance to many races of anthracnose and serve as a valuable source of genes in breeding for anthracnose resistance in Rwanda. 48 Table 8. Molecular markers linked to genes conditioning resistance to anthracnose in ten Rwandan bean cultivars. Genes Co-12 00'5 Co-4, 00-42 Co-42 Co-42 Co-9 Varieties Markers SEACTMCCA SABB SAS13 SH18 83314 8312 CAB19 + - - - - - G2331 - + - - - - N6224-4 - - .. - - - RAB487 +l- - + - - - RWK10 + - + - - - RWR1312I + - - - - - RWR1802 + - - - - - RWV167 + + + + + - RWV524 + - + - - - SCAM80CM/1 5 + - - - - - Th Presence of the marker -: absence Of the marker +/-: Heterogeneous for the marker. Cultivar 62331 carried the SAB3 marker linked to Co-5 but lacked markers linked to CO-42 and Co-12 alleles. The presence of the marker linked to Co-5 suggests that this cultivar carries the 00-5 gene similar to 62333 cultivar (Young et al., 1997). 62331 shares genetic Similarity with 62333 as both accessions came from the same region of Mexico. The Co-5 gene may be responsible for the resistance to races 7 and 73 observed in 62331. 49 The STS SEACTMCCA marker linked to the Co-12allele was also present in CAB19, RWK10, RWR1312I, RWR1802, RWV524, and SCAM80CM/15. Among these cultivars RWR1802 and RWV524 were resistant to both 7 and 73 races suggesting that they may carry the Co-12 allele similar to the one in Kaboon (Melotto and Kelly, 2000). The fact that cultivar RWV524 was susceptible to race 256 and also possessed the SAS13 marker that detects both Co-4 and Co-42 alleles suggests that this genotype may possess the Co-4 gene similar to the one also found in TO differential (Young et al., 1997). RAB487 was heterogeneous when inoculated with races 7 and 73. This heterogeneity was also seen when the marker linked to Co-12 allele was tested. Cultivars RAB487 and RWK10 possessed SAS13 marker but they did not appear to possess either SH18 or 88314 markers specifically linked to the Co-42 allele. The presence of the SAS13 marker that also amplifies the CO4 locus combined with the phenotypic resistance of RWK10 and RAB 487 to race 256 suggests that these genotypes may carry another allele at the 00-4 locus, probably the 00-43 allele (Alzate- Marin et al., 2002). The Co-4 locus in common bean has been shown to be multi- allelic (Melotto and Kelly, 2001). When markers linked to ALS, BCMV, BCMNV, C83 and Rust were tested (Table 9), six of the Rwandan cultivars N6224-4, RAB487, 62331, RWV524, RWK10, RWR1802 carried the SCAR marker SN02 linked to Phg 2 gene conditioning resistance to both races 63.19 and 31.17 of ALS. The SN02 marker was found in Mexico 54 and Cornell 49-242 (Sartorato et al., 1999; Nietsche et 50 al., 2000). Those genotypes could be valuable sources of resistance genes in breeding for resistance to ALS resistance. Cultivar 62331 possessed SCAR markers SAP6 and BC409 linked to two major QTL conditioning resistance to C88. This genotype was the only variety that combined both linked QTL conferring resistance to C88. RAB 487 and RWV167 possessed the QTL BC409 but they did not carry SAP6, whereas RWR1312I and CAB19 possessed only the SAP6 QTL. The other cultivars did not Show the presence of any marker linked to QTL conferring resistance to C88 tested. SAP6 and BC409 QTL are linked and mapped on bean linkage group B10 (Miklas et al., 2000b). The finding of both QTL in 62331 is a good indication that this genotype may carry the genomic region conferring resistance to C88. The presence of only one QTL in genotypes RWR1312I, CAB19, RWV167 and RAB487 was not well understood given that both QTL. are linked (Miklas et al., 2000b). The usefulness of the QTL for resistance to CBB on linkage group B10 was suggested to be limited to the Andean gene pool, Since many susceptible Middle American genotypes have been found to possess both markers (Kelly et aL,2003) 51 Table 9. Molecular markers linked to the genes and quantitative trait loci (QTL) conditioning resistance to angular leaf spot (Phg 2), been common mosaic necrosis virus (I), common bacterial blight (QTL), and rust (Ur-3 and Ur-5) in ten Rwandan bean cultivars Genes Phg 2 I QTL'“ QTL“ Ur-5 Ur—3 Markers Varieties SN02 SW13 SAP6 BC409 Sl19 SK14 CAB19 - - ~I- - .. + 62331 + - + + + _ N6224-4 + .. .. .. + .. RAB487 + + - + + + RWK10 + + - - + + RWR1312I - - + - + _ RWR1802 + - - _ + _ RWV167 - - - .j. - _ RWV524 + - - _ + _ SCAM80CM/1 5 - - - - - + +: Presence of marker; -: absence of the marker; '" QTL: Quantitative trait loci conferring resistance to bacterial common blight on linkage group B10. Varieties RAB487 and RWK10 were the only Rwandan cultivars possessing the hypersensitive I gene conditioning resistance to BCMV and BCMNV. The presence of the I gene in RAB487 and RWK10 was confirmed by their top necrosis reaction after inoculation with the NL3 strain of BCMNV. To be effective in those regions where strains of BCMNV are prevalent, the I gene present in RWK10 and RAB 487 needs to be protected by recessive genes 52 particularly the bc-3 gene. The protection of the I gene in these genotypes will confer resistance to all known strains of BCMV and BCMNV since currently no known strain has overcome the Ibo-3 pyramid. In the areas where phenotypic screening is the only available method, the use of the donor parents having both land bc-3 is the best strategy to ascertain the fixation of the lgene in successive generations. Rwandan cultivars RWR1312I, N6224-4, RAB487, 62331, RWV524, RWK10, and RWR1802 possessed the Sl19 marker linked to Ur-5 locus conferring resistance to U. appendiculatus. In addition, cultivars RAB487, RWK10, SCAM 80CM/15, and CAB19 possessed the SK14 linked to Ur-3 gene. RWV167 did not possess any marker linked to rust resistance. The Sl19 SCAR marker (Melotto and Kelly, 1998) was developed from a RAPD marker identified by Haley et al. (1993) and was found in B-190 breeding line, which had a broad- based resistance to many races of bean rust. This broad based response was found to be controlled by several dominant genes tightly linked in coupling and inherited as a major block (Stavely, 1984). The RAPD marker K14, from which the SCAR SK14 marker was developed, showed a broad application across gene pools and beans market Classes (Haley et al., 1993). The Ur-3 gene is an important resistance gene effective against 43 of 87 races of bean rust characterized in the US. collection (Stavely, 1998). The pyramid of the markers linked to broad based rust resistance genes in RWK10 and RAB487 might protect these varieties against a range of bean rust races and could furnish 53 valuable sources of resistance genes in breeding for rust resistance in future Rwandan bean cultivars. The presence of a marker in a given genotype does not always imply the presence of the linked resistance gene since some markers were shown to be gene pool or population specific in the host (Johnson et al., 1997; Miklas et al., 20003, Urrea et al., 1996). Recombination between the gene and marker can also lead to the presence of the marker that is not linked to the gene of interest. This suggests that breeders must treat marker data for gene detection with a level of caution and wherever possible, inoculations with races of the pathogen must be performed to validate marker data before using the selected genotypes for breeding purposes. The transfer of the bc—3 gene After the first hybridization, all the hybrid progenies were crossed to the recurrent parents. These crosses produced B61F1 seeds, which were susceptible and were genotypically 50% BC-38C—3 and 50% BC-3bc-3. The reaction of these seeds to NL3 was not tested because the inoculation would result in the death of all of the BC1F1 materials. Instead, all the BC1F1 plants were self-pollinated to produce the BC1F2 generation that segregated 1 resistant: 7 susceptibles after inoculation with NL3 strain of BCMNV. The BC1F2 seeds were also segregating for seed size, shape and color. As many populations were being tested simultaneously, where seeds were adequate, only those with characteristics similar to the female parents were planted and tested with BCMNV. The result of these inoculations is shown in Tables 10, 11, 12 and 13. The data illustrates the 54 number of plants observed in different Classes of reactions, whereas the Chi- squared (X2) was calculated for only two Classes; resistant versus susceptible since only resistant plants carrying the bc-3 gene were desired. The susceptible class comprises plants with mosaic indicating the absence of resistance genes, mild mosaic resulting from the presence of the single bc—12 gene, top necrosis suggesting the presence of the single I gene, and vein necrosis reaction indicating the presence of the combination ofbc-12 and I resistance genes. After the screening, the resistant plants were used as male parents to provide pollen for crossing with the susceptible Rwandan recurrent parents. These crosses resulted in the BC2F1 generation, which was the last generation to be completed at MSU. The next backcross generations will be continued in Rwanda. Table 10. Observed reactions in IC1F2 populations after inoculation with NL-3 strain of BCMNV. Top NO-reaction Total Crosses” Mosaic Necrosis RWV524/ USDK-4 23 20 4 47 N6224-4/ USWK-6 21 28 1 50 RWR1802/ USDK-4 26 36 4 66 ‘ Recurrent parents lacked resistance genes for BCMNV. 55 Table 11. Resistant vs susceptible, segregation ratios of BC1F2 populations, _goodness of fit to expected ratios, and p value Crosses“F Resistant Susceptible Expected X7 P value plants plants ratios R: S“ RWV524 / USDK-4 4 43 1:7 0.68 0.41 N6224-4 / USWK-6 1 49 1:7 5.04 0.02 RWR1802/ USDK-4 4 62 1:7 2.50 0.11 rRecurrent parents lacked resistance genes for BCMNV, " R= Resistant, S= Susceptible Table 12. Observed reactions in BC1F2 populations after inoculation with NL-3 strain of BCMNV Crosses” Mosaic mgiaic xgtigosis Nggrosis lrig-won Total RAB487/ l99530 1 0 0 43 2 46 RWK10 I USCR-7 0 0 0 60 8 68 RWR1312I / USCR-7 4 10 10 1O 3 37 SCAM15/ 18CM/l99530 11 2 0 0 5 18 CAB19 / l99530 20 7 0 0 3 30 ”Recurrent parents carried I or bc-1Tresistance genes for BCMNV. 56 Table 13. Resistant vs susceptible, segregation ratios of BC1F2 populations, goodness of fit to expected ratios, and pvalue. Crosses++ Rgfaiittint Suzcaerpttsible ExE-ig'ed X2 P value RAB 487/ l99530 3 43 1': 7 1.50 0.22 RWK10 / USCR-7 8 60 1: 7 0.03 0.85 RWR13121/ USCR-7 3 34 1: 7 0.652 0.40 SCAM15/18CM/l99530 5 13 1: 7 3-84 0.05 CAB 19 / l99530 3 27 1: 7 0.171 0.60 Ti Recurrent parents carried I or bc-1z resistance genes for BCMNV " R= Resistant, S= Susceptible. The inheritance of the bc-3 gene in RWV524/USDK-4 and RWR1802/USDK-4 populations segregated as expected. In fact, these two Rwandan varieties lacked all resistance genes for BCMNV and they were crossed with genotypes carrying both bc-3 and I genes. The observed ratios in the BC1F2 generation were as expected (1 R:7S; Table 11). For the cross N6224- 4/USWK-6, two genes I and bc-3 were segregating but the observed ratios deviated from the expected values. It seems that in this cross, the bc-3 gene was not transferred as expected because out of 50 plants tested, only one was resistant. The small size and the non random sample of seeds tested may be the cause of the poor segregation of the bc-3 gene. Otherwise, the cross was successful given that plants showing the top necrosis were observed indicating that the I gene was inherited and was present in successive generations. In the cross RAB487/I99530, both I gene from RAB487 and bc-3 gene from l99530 segregated (Table 13). The observed ratios matched the expected when two 57 genes are segregating. For the cross RWK10/USCR-7, the I gene was fixed since both parents carried the I gene and only the bc—3 gene was segregating. This cross was the most successful of all crosses with a high significance value (p= 0.85) of the observed ratios meeting the expected ratios. The fact that these genotypes had a short growing season, large flowers and the presence of I gene facilitated crossing and selection. As expected, only two phenotypes: no-reaction and top necrosis were observed, which reduced the chance of errors when scoring phenotypes or confusion of the different reactions to the virus. In the crosses SCAM15/18CM/l99530 and CAB19/l99530, two recessive genes bc-12 and bc-3 were segregating. These crosses were not easy to score because the readings were delayed to determine which genotypes display a no-reaction, mosaic or mild mosaic reactions and this delay might have contributed to incorrect scoring. The range of phenotypes was more problematic for the cross RWR13121/USCR-7 in which bc-12, bc-3 and I genes were segregating. In this cross, misclassification could happen especially when vein necrosis, mild mosaic and no-reaction were expected because the inoculated leaves could senesce and fall off from the plant before the phenotype could be recorded. In this instance, the second or third trifoliate leaves were also inoculated and were used to verify the reaction of any unknown plants. In this cross some of the plants showed top necrosis reaction and this provided additional evidence for the segregation of the lgene in those populations. The transfer of the bc-3 gene into Rwandan cultivars was successful in early successive segregating generations. The availability of this wide array of 58 bean seed types preferred in Rwanda carrying the bc-3 gene will offer an excellent assurance for Rwandan growers to use their preferred seeds with protection against local strains of BCMNV. Those varieties will also be sources of bC-3 gene for the National bean program in ISAR to develop protected lgene resistance for successful protection of future bean cultivars against a large array of strains of BCMV and BCMNV present in Rwanda. 59 Conclusion The objectives of this study were to evaluate Rwandan genotypes for the presence of disease resistance genes and transfer the bc-3 gene conferring resistance to BCMNV into local Rwandan bean cultivars. For the evaluation of the presence of disease resistance genes, cultivar RWV167 appeared to have a phenotypic resistance to the races of anthracnose tested and to posses a pyramid of five markers linked to anthracnose resistance genes. These data suggested that RWV167 might have a broad resistance against a large number of anthracnose races and valuable as a potential source for anthracnose resistance genes. Cultivars RAB487 and RWK10 possessed a large number of markers tested linked to almost all the disease pathogens tested. The same varieties carry the lgene-conditioning resistance to BCMV and BCMNV. To be effective in Central and East Africa, this I gene needs to be protected by the bc-3 and this combination offers complete resistance against all known strains of BCMV and BCMNV reported to date. Variety 62331 possessed the phenotypic resistance to races of anthracnose tested and it appeared to have the marker linked to Co-5 gene. This cultivar has the combination of SAP6 and BC409 markers linked to major QTL conditioning resistance to C88, molecular marker linked to Phg 2 gene conferring resistance to angular leaf spot, and Ur-5 gene for rust resistance. The backcross breeding program conducted in this study to introduce the bc—3 gene into bean seed types preferred in Rwanda was successful. The 60 availability of the local bean genotypes protected against BCMNV will provide a good opportunity for ISAR to use this germplasm as sources Of this gene for future development of BCMNV resistant varieties in Rwanda. 61 Chapter 3: Search for a new marker linked to bc-3 gene Introduction In plant breeding, molecular markers tightly linked to genes of interest (3 5 CM) are needed for implementing MAS strategy (T anksley, 1983). However, many markers have been shown to be restricted to specific genetic populations limiting their usefulness in a large range of genetic materials. In common bean, Middle America and Andean gene pools have been established based on data from morphological traits, isozymes, seed proteins and DNA markers (Gepts, 1988; Singh et al., 1991). These gene pools reflect multiple domestication events within distinct wild populations of common bean (Gepts and Debouck, 1991). Molecular markers that are gene pool specific have been demonstrated with RAPD markers linked to Up 2 gene for rust resistance, bc-12, and bc—3 genes for BCMV and BCMNV resistance, and bgm-1 gene for reSistance to bean golden mosaic virus (Miklas et al., 1993, Haley et al., 1994a, Miklas et al., 2000a; Urrea et al., 1996). In contrast, RAPD markers linked to Ur-3, I, and C042 genes for rust, BCMNV and BCMV, and anthracnose resistance, respectively (Haley et al., 1994c; Haley et al., 1994b; Young et al., 1998) are known to have a broad application across a wide range of germplasm from both gene pools. Different techniques such as the use of NlLs, bulked segregant analysis (BSA), and DNA bulks from different gene pools in early screening, have been utilized to search for RAPD and AFLP markers linked to genes of interest in common bean (Haley et al., 1994a; Haley et al., 1994c; Young and Kelly, 1996; Vallejo and Kelly, 2002). In these studies, RAPD markers linked to recessive 62 genes such as bgm—1 conditioning resistance to been golden mosaic virus (Urrea et al., 1996), bc-12 for BCMV and BCMNV (Miklas et al., 2000a) proved to be useful. Where RAPD markers are ineffective, AFLP markers represent a powerful technique for studies in crop improvement such as genetic diversity, genetic linkage and gene mapping, map based cloning, and genetic homogeneity studies (Hazen et al., 2002; Vuylsteke et al., 2000; Lombard et al., 2002). Dominant and recessive genes condition resistance to BCMV and BCMNV in common bean (Drijfhout, 1978). The dominant resistance is conferred by the I gene and the recessive resistance is provided by a series of be recessive loci. Among the recessive genes, bc-3 confers resistance to all known strains of BCMNV and many strains of BCMV (Miklas et al., 1998). The transfer of this gene into susceptible bean genotypes has proven difficult due to its recessive nature. Progeny tests are always required to ascertain that the gene is being transferred in the successive breeding generations. In addition direct screening with strains Of BCMNV is needed to confirm the presence of bc-3 gene. Markers linked to bc-3 gene are needed to facilitate the introgression and gene pyramiding of bc-3 and I gene, which confer resistance to all known strains of BCMV and BCMNV. A number of DNA markers linked to the bc—3 gene have been identified (Haley et al., 1994a; Johnson et al., 1997) but their use in MAS has been limited because of their lack of reproducibility across laboratories and their specificity to bean gene pools and populations. 63 1“,...“ o". T" ' X To efficiently introgress the bc—3 gene for resistance to BCMV and BCMNV into susceptible bean cultivars, there is a need to identify a new robust DNA marker that will be more broad based in different genetic backgrounds and reproducible across laboratories. In this work, NILS and DNA from different gene pools were combined with the BSA (Haley et al., 1994c; Young and Kelly, 1996; Michelmore et al., 1991) to increase the opportunity of finding a robust marker linked to bc-3 gene using both RAPD and AFLP analyses. Materials and methods Mapping population. Two mapping populations representing different gene pools were used to screen for a marker linked to bc-3. The Middle American gene pool was represented by a population derived from a cross between Bunsi and Raven cultivars while the Andean population was derived from a cross between the Rwandan variety RWK10 and a U. S. breeding line USCR-7. Bunsi/Raven population consisted of a group of F47 recombinant inbred lines (RILs) developed at Michigan State University. The F5 progeny from a single F4 plant that segregated for hypersensitive top necrosis and no reaction after inoculation with the NL3 strain were chosen as they were segregating for the bc-3 gene. The F5 progeny were advanced in bulk to the F7 generation. RlLs that were near- isogenic for the bc-3 locus after inoculation with NL3 strain of BCMNV were used in this study. In the second population, the cross was made between two 64 Andean RWK10/USCR-7 breeding lines after recording their reaction to NL3 strain of BCMNV. The RWK10 is a large seeded dry bean cultivar carrying the hypersensitive lgene conferring resistance to BCMV and BCMNV (ISAR, 2001). The USCR-7 is a cranberry bean genotype carrying the l and bc-3 genes combination (Miklas et al., 2002b). The F1 generation was selfed to produce F2 individuals that segregated for the bc-3 gene but were fixed for the I gene. Leaf samples for DNA extraction were collected from each F2 plant and stored at -80 °C. Each F2 plant was self-pollinated to produce F23 families that were inoculated to determine the genotype of the parental F2 plant (Figure 3) RWK10 (I gene) X USCR-7 (land bc-3 genes) F1 is F2 individuals 1 5 F2; 3 families Figure 3. Development of RWK10/USCR-7 mapping population. 81: Self-pollination To determine the reaction of individual F2 to BCMNV, an average of ten or more seedlings from each F23 family were mechanically rub inoculated with the 65 NL3 strain of BCMNV. The observed reactions in F23 families indicated the genotype of the F2 parental plants used to produce the F23 families. Identification of a marker linked to bc-3 gene. RAPD analysis The search for a RAPD marker linked to bc-3 gene began with screening random decamer primers (Operon Technologies, Alameda, CA and Integrated DNA Technologies, Coralville, IA) using DNA bulks from NILS along with the parents of the Bunsi/Raven population. DNA extractions followed the mini-prep protocol of Afanador et al. (1993). Equimolar pools of DNA from 5 resistant individuals and 5 susceptible individuals respectively were used in BSA (Michelmore et al.1991). PCR profile and agarose gel-electrophoresis conditions were those of Haley et al. (1993). The purpose of using these NILS was to speed up the screening process when the second mapping population was growing and increase the chance Of finding a robust marker that would be functional in both Middle America and Andean bean gene pools. Primers shown to be polymorphic among Bunsi/Raven NILS were run on RWK10, USCR-7, and the DNA bulks of homozygotes F2 individuals created using BSA in RWK10/USCR-7 population. Those primers that showed polymorphism between parents and bulks were subsequently tested on individual members of the bulks. 66 AFLP analysis As in RAPD analysis, BSA was combined with AFLP and silver staining techniques (Promega Madison, WI) to continue the search for a marker linked to bc~3 gene. DNA of five plants in each group (resistant and susceptible) was used in the bulks. After DNA isolation and quantification, the DNA concentration was brought to 100ng/pl by dilutions. The AFLP protocol followed those previously described (Vos et al.1995; Hazen et al., 2002) with minor modifications. DNA was digested using ECORI and Msel enzymes in a cocktail of 10 pl of genomic DNA (100 ng/pl), 32.75 pl of H20, 1X one phorall buffer (OPA, Pharmacia), 0.1 U/pl Msel, 0.1 Ulpl BSA buffer (both from New England Biolabs), and 0.1 U/pl ECORI (GibcoBRL) for each sample. The digestion was carried out in the water bath at 37 °C for 3 hours and the samples were agitated after each hour. At the completion of the digestion, the enzymes were inaCtivated by heating the samples at 70 °C for 15 minutes. At the end of the digestion, the double stranded adapters were linked to the restriction fragments by adding to each sample 10.03 pl of the mixture of 1 pg/pl ECORI adapter, 1 pg/pl Msel adapter (Integrated DNA Technologies, Coralville, IA), 1 pl of T4 DNA ligase 10X buffer, 0.33 pl of T4 DNA ligase (3 U/pl, Promega), and 6.7 pl of ddH2O and incubated at room temperature for 3 hour with the agitation after each hour. The pre amplification of the DNA was carried out using both EA and MC primer combination. Each 20 pl reaction comprised 11.9 pl of H20, 2 pl of template DNA from restriction/ligation, 25 ng of EcoRl+ A oligo, 25 ng Msel +C oligo (Integrated DNA Technologies, Coralville, IA), 0.5 mM dNTP (lnvitrogen, Carlsbad, CA), 1X 67 ra- ‘1." PCR buffer, 1.5 mM MgCl2, and 0.5 U Taq polymerase (Promega, Madison, WI). The PCR conditions for the pre- amplification were 94 °C/2 min; 26 cycles of 94 °C l1 min, 56 °C/1 min, 72 °C I1 min; and 72 °C I5 min. The preamplified PCR products were diluted by adding 100pl of sterile H20 and used as template in selective amplification. The selective amplification mixture was prepared in 20 pl consisting of 1 pl of DNA from preselective PCR, 13.8 pl of sterile water, 25 ng ECORI+ANN oligo, 30 ng Msel+CNN oligo (Integrated DNA Technologies, Coralville, IA), 0.4 mM dNTP (lnvitrogen, Carsbad, CA), 1X PCR buffer, 1.5 mM MgCl2, and 0.4 U Taq polymerase (Promega, Madison, WI). The PCR profile was 94°C/2 min; 12 cycles of 94°C/30s, 65°C/3OS, 72°Cl1 min with annealing temperatures decreasing by 0.7°C each cycle; 23 cycles of 94°C/3OS, 56°C/30s, 72°C/1 min; and an extension of 2 min at 72°C. The PCR products from this amplification were combined with 8p| of formamide loading buffer (98% formamide, 10 mM EDTA, PH8.0, 1 mg/ml bromophenol blue, and 1.0 mg/ml xylene cyanol). The samples were run on 6 % polyacrylamide gel at 85W. Silver staining was carried out following manufacturer’s instructions (Promega, Madison, WI) except that both fix/stop and developing solutions were partially frozen at —20°C. The gels were scored searching for polymorphism both between parents as well as bulks. The AFLP analysis was performed on the RWK10/USCR-7 F2 mapping population because most of the same primer combinations had been run previously on the Bunsi/Raven population (Ender, 2003) 68 Sequence tagged sites (STS) marker development. To clone an AFLP fragment was cloned as described earlier (Vallejo and Kelly, 2002). The PCR product from selective amplification with the original primer combination that generated a fragment cosegregating in the mapping population was run on 6% polyacrylamide gel leaving empty lane between successive samples. The fresh fragment of interest was excised from the gel using a sterile sharp-edged blade razor and it was soaked in 50 pl of sterile water overnight at 4°C. Five pl of the supernatant was used as template DNA in PCR reaction where the original primer combination and conditions of selective amplification were used. The resulting PCR product was run on 2% NuSieve 6T6 (FMC BioProducts, ME) agarose gel using TBE (70 Volts) for one hour at 4°C. The band was excised from the agarose gel and it was purified using QIAquickTM Gel Extraction Kit (QIAGEN lnc., Valencia, CA) following the manufacturer's instructions. After the purification, the same DNA was used in the original selective PCR and 4 pl of the PCR product was cloned into pCR®2.1 - TOPO® plasmid (from TOPO TA cloning Kit for sequencing Version E, lnvitrogen, Carlsbad, CA 92008) and transformed in Escherichia coli following the manufacturer’s instructions. The recombinants plasmids were grown overnight on a selective LB media containing ampicillin (75 pg/ml) as a selective agent, X—gal (40 mg/ml), and IPTG (100 mM). Ten transformed colonies were cultured in 3ml of magnificent broth liquid media (Mac Connell Research, San Diego, CA) and grown at 37 °C for 16 hours. The plasmid DNA was extracted using the Wizard® Plus SV Minipreps DNA purification system (Promega Madison, WI) following the 69 manufacturer‘s instructions. The insertion was verified by digesting the plasmid DNA with ECORI restriction enzyme (5 pl of plasmid DNA, 3pl 10X React® 3 buffer, 1 pl of 10 U/pl ECORI enzyme, and 21 pl of water) for 1 hour at 37 °C and run on 1.4% agarose gel. The DNA of the clone with the expected insert size was quantified and sequenced at Michigan State University Genomics Technology Support Facility. Based on the sequence, a pair of primers was designed using Primer3 software (http://biowb.sdsc.edu/C6l/BW.cgi). The synthesis of the primers was performed by Integrated DNA Technologies (Coralville, IA). To confirm the conversion of AFLP into STS, PCR reactions were performed using the new primer combination on the parents and bulks first and then it was extended to the whole mapping population. The optimal PCR amplification was conducted in a reaction containing 17.80 pl of H20, 3 pl of template DNA (10 ng/pl), 3 pl of the primer mix (10ng/pl), 0.6 pl of 5 mM dNTP mix, 3 pl of 10X PCR buffer, 2.25 pl of 5 mM MgCl2, and 2 U Taq DNA polymerase (lnvitrogen, Carsbad, CA). DNA amplification was performed using Programmable Thermal Controller (PTC)-100 therrnocycler (MJ Research Inc., Waltham, MA) and the PCR protocol consisted of 94°C for 2 min, 25 cycles of 30s at 94°C, 1 min at 53°C, and 1 min at 72°C and the final extension of 5 min at 72°C. To separate the amplified products, the PCR products were run on 6 % polyacrylamide gel and visualized with the Silver staining method (Promega Madison, WI). 70 Segregation analysis The segregation for both molecular and phenotypic data for the F2 and F23 generation was analyzed to verify their goodness of fit with 1R:3S and 1RR:2Rr:1rr segregation ratios using the X2 test. Linkage analysis The linkage analysis between the markers (AFLP and STS) and bc-3 gene was performed using MAPMAKER 3.0 software (Lander et al., 1987) with a minimum LOD of 3.0. The recombination fraction was calculated using the Kosambi function. 71 Results. RAPD analysis A total of 1009 random primers were screened on both Bunsi/Raven and RWK10/USCR-7 parents along with the respective resistant and susceptible bulks. Only four primers (0.4%) were polymorphic between the four resistant and susceptible parents (Table 14). One hundred sixty-eight primers (16.6%) were polymorphic between Bunsi and Raven and 115 primers (11.4%) were polymorphic between USCR—7 and RWK10 but none were polymorphic between resistant (USCR—7 and resistant bulk) and susceptible (RWK10 and susceptible bulk) genotypes. Four primers (0.4%) were polymorphic between parents and bulks in Bunsi/Raven population. Among these only AA19 primer appeared to cosegregate with bc-3 gene in Bunsi/Raven population when individual members of the bulks were examined. However, AA19 did not prove to be linked with the bc-3 when tested on the entire mapping population. Only two (0.2%) primers were polymorphic between parents and bulks in USCR-7/RWK10 population. Table 14. Pattern of the gimers that were polymorphic between both parents. Primer Raven (R) Bunsi (S) USCR-7(R) RWK10 (S) AC12 + - + - AD17 - + - + UBC156 + - + - UBC241 - + - .j. +2 Presence of the band. -: Absence of the band 72 AFLP analysis. A total of 42 primer pair combinations were screened against parents and bulks in the RWK10/USCR-7 population. Twelve primer combinations (20.5%) generated at least one polymorphism between parents and a total of 23 polymorphisms were generated. Out of 12 primer combinations, 5 were polymorphic between both parents and bulks. When individual members of the bulks were examined only one codominant primer combination EACAMCGG cosegregated with the disease reaction (Table 15, Figure 4). The cosegregating bands were approximately 169/172 bp in size. The smaller fragment was associated with resistance and the larger fragment was associated with the susceptibility (Figure 4). This AFLP primer combination was tested in the entire mapping population. A total of 92 F2 individuals were screened against this primer combination. As some of the F23 families had less than 10 plants, verifiable phenotypic data was not adequate to confirm the phenotype. Since the resistance was recessive, heterozygotes could still carry bc—3 allele but exhibit a susceptible phenotype. It was decided that progeny sizes under 10 plants per family was too small to decide between homozygous and heterozygous phenotypes so homozygote individuals with less than 10 plants were not included in the linkage analysis. Using this criterion, 58 F2 individuals were used to calculate the linkage between the EACAMCGG169,172 marker and the disease reaction. The linkage analysis Showed that EACAMCGG139,172 marker was linked at 4.4 cM distance from the bc—3 gene. When the X2 test was performed, the 73 results from the F2 support a segregation ratio of 1:3 resistant to susceptible individuals (T able 15). These F2 results were confirmed by the analysis of the F23 families that showed a good fit to a segregation ratio of 1:2:1 (Table 15). These segregation ratios indicated that the resistance to BCMNV in USCR-7 was conditioned by a recessive gene as reported by Drijflwout (1978). Table 15. Chi-square analysis of Observed ratios for bc-3 gene and the AFLP marker EACAMCGG169/172 segregating for resistant (R) and susceptible (S) in F2 generation and for homozygous resistant (RR), heterozygous susceptible (Rr), and homozygous susceptible (rrLfor F23 families in RWK10/ USCR-7 population Expected Observed Locus tested Generation R: S and R: S and X2 Probability RR: Rr: rr RR: Rr: rr bc-3 F2 1 :3 13:45 0.2 0.625 bc-3 F2 3 1:2:1 13:34:11 1.16 0.375 EACAMCGG F2 1 :3 10:48 1 .86 0.175 EACAMCGG F2; 3 12221 10236312 3.52 0.175 To facilitate the scoring and reduce the costs associated with the AFLP analysis, the EACAMCGG139m2 marker was transformed into the STS SEACAMCGG marker, which eliminates the digestion and ligation steps. With STS markers, selected polymorphisms are detected from direct PCR amplification of the DNA with the primer combination. 74 Figure 4. AFLP gel showing the polymorphism generated by the primer combination EACAMCGG169l172. Lanes: 1. Resistant parent (USCR-7); 2. Susceptible parent (RWK10); 3. Resistant bulk; 4. Susceptible bulk; 5, 6, 7, and 10. Resistant F2 individuals; 8.12.13, and 14. Susceptible F2 individuals; 9 and 11. Heterozygous F2 individuals; 15. Ten bp ladder. 75 The 169 bp fragment corresponding with resistance phenotype was sequenced and two primers combinations were designed to amplify codominant 134/137 bp fragments (Figure 5). The smaller sized fragment resulted from the elimination of the sequence corresponding to AFLP adapters in designing primers. The sequences of the primers synthesized were 5'-C66TCATACATTI'ATACAAAACC—3’ for the fonrvard primer and 5’-A6TTI'GACAGGTGCAA6TCT-3’ for the reverse primer. This synthesized SEACAMCGG134/137 marker generated bands of the same pattern as the original AFLP and resulted in the same number of recombinants. This primer combination amplified an additional weak band above the expected bands, but the use of completely frozen solutions in fixing, staining, and developing helped to eliminate this band. The amplification of a few additional bands by STS markers was also reported by Guo et al. (2003) in wheat and Meksem et al. (2001) in soybean. Figure 5. PCR product obtained using the SEACAMCGG134/137 primer. Lanes: 1. ten bp ladder;2. resistant parent; 3. susceptible parent; 4. resistant bulk; 5. susceptible bulk; 6-11. F2 resistant individuals; 12-17. F2 heterozygote individuals;18-24. F2 susceptible individuals 76 To evaluate the robustness of the SEACA/MCGG134I137 marker, it was tested on a diverse group of bean breeding lines known to carry the bc-3 gene. The bean genotypes represented a collection of genotypes from kidney, pinto, black, cranberry, small red, and great northern commercial classes representing the two gene pools of common bean (Table 16). Table 16. Survey of genotypes carrying bc-3 gene for presence of SEACAMCGG marker Genotypes Gene pool” Class Size of the band (bp) USCR-7 A Cranberry 134 USCR-9 A Cranberry 134 USLK-2 A Kidney 137 USDK-4 A Kidney 137 USW K-6 A Kidney 134 PR0066-6 MA Small red 134 l99532 MA Small red 134 I99530 MA Small red 134 RAVEN MA Black 134 800108 MA Black 134 BDM-RM R-1 1 MA Pinto 134 BDM-RMR-16 MA Pinto 134 BELDAKMI MA Pinto 1 34 699750 MA Great northern 134 “ A: Andean; MA: Middle America; 134bp: Resistance band; 137bp: Susceptibility band 77 Discussion NILS from the Middle American gene pool originating from the Bunsi/Raven population were first used to find a RAPD marker linked to bc—3 gene. The efficiency of NILS for targeting RAPD markers linked to pest resistance genes was suggested by Haley et al. (1994c) as a strategy to reduce the number of false positive polymorphisms, and strengthen the linkage of the identified markers. In this study, putative polymorphic primers obtained from screening NILS were tested on parents and bulks from the Andean gene pool. The use of DNA from Middle America and Andean gene pools facilitated the identification of RAPD markers flanking the Are gene for anthracnose resistance in common bean (Young and Kelly, 1996). Combining these two strategies with BSA, which reduces targeted markers to a smaller locus within the genome, was expected to facilitate and improve the finding of a robust marker linked to bc-3 gene. These strategies reduced the incidence of false positive bands since only a small number of polymorphic primers between both parents and bulks were observed compared to the number of primers tested. None of the primers however, proved to cosegregate with disease resistance in both populations. Thus, the RAPD technique failed to identify a marker linked to bc—3 gene. The inability of RAPD as a tool to find markers linked to recessive genes was also encountered when searching for a RAPD marker linked to bgm-2 gene conferring resistance to been golden mosaic virus (66. Mufloz Perea, personal communication). While the nature of recessive genes is unknown, one of the hypotheses explaining their role in resistance is that the resistant host lacks a host function essential for 78 critical steps in viral pathogenesis. Consequently the dominant allele encodes a host factor, which is required for virus replication and/or movement in the susceptible host (Fraser, 1992; Johansen et al., 2001; Harrison, 2002). The hypothesis that recessive resistance may actually be caused by the lack of a host factor was supported by the evidence that the va locus conditioning resistance to potato virus Y in N. tabacum is caused by the loss of a large chromosomal fragment that encompasses the susceptibility gene Va (Noguchi et al., 1999). However, the generalization that the nature of recessive genes in viral resistance is due to missing or deleted sequences does not seem to be universal given that some researchers were able to find useful markers linked in repulsion phase to bc-12 and bgm-1 recessive genes conditioning resistance to BCMV and BGMV in common bean (Miklas et al., 2000a; Urrea et al., 1996). The negative regulator of resistance encoded by the susceptibility allele seems to be controlling the function of recessive virus resistance genes in common bean. Finding markers linked to a recessive loss of function allele in the host could prove challenging. The genetic variability observed between parents was higher between Raven and Bunsi (16.6%) than between USCR-7 and RWK10 (11.4%). This increased DNA variability in Middle America gene pool compared to Andean gene pool was also observed by Beebe et al. (2000; 2001) who concluded that a narrower genetic base in Andean gene pool existed compared to the Middle American pool. In the same genotypes, AFLP detected greater genetic variability between the Andean parents than RAPD. In only 42 primer pair combinations 20.5% were polymorphic between the Andean parents suggesting that this 79 technique may be more powerful than RAPD in detecting polymorphism among highly related genetic material. The AFLP technique enabled the finding of a codominant marker tightly linked to bc-3 gene in a small population of 58 F2 plants suggesting that this marker might be even more tightly linked in a larger population. The finding of this marker in the RWK10/USCR-7 population using AFLP analysis demonstrated the power of the AFLP technique over RAPDs suggesting that been breeders should consider the potential use of this type of marker in breeding highly related material despite the past success with RAPD markers in common bean (Kelly and Miklas, 1998). The finding of a codominant AFLP marker was unexpected. The codominant nature of the AFLP and STS markers detected in this study is particularly useful in the breeding of recessive bc—3 gene. Codominant markers will help in the identification of heterozygous individuals at each generation, eliminating the number Of generations required for progeny testing. However it was not possible to visualize the polymorphism generated by the SEACAMCGG134/137 marker on regular agarose gels. This may be explained by the small size (137 bp) of the original AFLP fragment. In fact, the fragment generated was 169 bp including 32 bp for the EcoRl and Msel adapters that were eliminated in designing the primers. With this small fragment size, it becomes difficult to design appropriate primers that could result in PCR products capable of being distinguished by standard agarose gel. In addition the low resolution of the agarose gel cannot detect the small size differences as effectively as polyacrylamide gel electrophoresis. The use of metaphor agarose gel may facilitate the discrimination between the resistant and susceptible 80 fragments since Meksem et al. (2001) demonstrated the ability of metaphor agarose gel to separate fragments with a 4bp difference in soybean. The fragment representing the resistance phenotype generated by the SE ACAMCGG134/137 marker was present in pinto, cranberry, small red, black, and great northern beans known to carry bc-3 gene. This association suggested that the marker might be useful for MAS in those distinct seed Classes. In kidney beans, the resistance-associated band was present in USWK-6, but the resistant lines USLK-2 and USDK—4 genotypes showed the band for susceptibility. These results suggest that this marker needs to be tested in a large number of kidney beans to ascertain its usefulness in this class. Recombination could have occurred between the marker and the bc-3 gene in repeated backcrossing used to develop resistant kidney bean lines USDK-4 and USLK-2 (Miklas et al., 2002a). The SE ACAMCGG134H37 marker should facilitate marker-assisted breeding for bc-3 gene introgression to develop bean varieties resistant to BCMNV. lmportantly this marker may offer the unique opportunity to develop pyramid varieties combining the l and bc-3 genes, which is currently known to confer resistance to all known strains of BCMV and BCMNV in common bean. 81 Conclusion Among the 1,009 RAPD primers tested on Raven, Bunsi, USCR-6 and RWK10, none cosegregated with the disease reaction of the populations, which prevented the detection of a RAPD marker linked to the recessive gene bc-3. When the genetic variability between the parents used in this study was compared, greater DNA variability was found between Middle American parents Bunsi and Raven than between Andean USCR-7 and RWK 10 parents based on RAPD data. The AFLP technique was more powerful in generating polymorphisms between USCR-7 and RWK10 and it proved to be more valuable in finding a marker linked to bc-3 gene. The AFLP marker generated by the primer combination EACAMCGG139/172 was a codominant marker that was linked at 4.4 CM from the bc-3 gene. The AFLP EACAMCGG133172 marker was transformed into a STS SEACAMCGGI34n37 marker for the purpose of facilitating the scoring. The SEACAMCGG1341137 marker cosegregated with the disease resistance in the same ratio and distance as the original AFLP marker. The codominant scoring of the SEACAMCGG134/137 marker should facilitate the selection of homozygous bc-3 lines given that it will allow breeders to differentiate homozygous from heterozygous individuals. The resistant fragment could not be separated from the susceptible fragment generated by the SE ACAMCGG 134/137 marker on regular agarose gel, so metaphor agarose gel may be more useful in this regard. When the SEACAM333134I137 marker was tested on different bean varieties carrying the bc-3 gene, it appeared that it might have limited use in kidney beans because some of the resistant varieties in this class exhibited the susceptibility 82 band. The SEACAMCGG134/137 appeared to have broader application in other bean market classes tested. A supplemental survey in the kidney class may show where this marker can be applicable within kidney beans since white kidney line proved to have the band linked to the resistance while light and dark red kidney bean lines possessed the susceptibility band. The identified SEACAMCGGm/m marker will offer the opportunity to use MAS in developing bean varieties resistant to BCMNV and more importantly to develop pyramid varieties combining the I and bc-3 genes resistant to all know strains of BCMV and BCMNV in common bean. 83 APPENDIX TI'GATGAGTCCTGAGTAACGGTCATACATTI'ATACAAAACCATAAGTGTTTATCATAA'I'I'C ATC'ITI'CTCGTGTCAGACTTCTACTGACTCTACl | I IGAAAGACTCGTGTACTGAAATTAG CATTCAGAGACTTGCACCTGTCAAACTAATCCATGTGAATTGGTACG Figure 6. Sequence of the resistant AFLP fragment 84 References Afanador, L.K., S.D, Haley, and JD. Kelly. 1993. Adoption of mini-prep DNA extraction method for RAPD marker analysis in common bean (Phaseolus vulgaris). Ann. Rep. Bean Improv. Coop. 36: 10-11. Alconero, R., J.P. Meiner, and A. Santiago, 1972. A new strain of common bean mosaic in Puerto Rico. Phytopathology 62: 667(Abstr.). Allen, D.J., M. Dessert, P. Trutmann, and J. Voss. Common beans in Africa and their constraints. In : CIAT (Centro Intemational de Agriculture Tropical). 1989. Bean production problems in the tropics. 2"d ed. Schwartz, HF. and Pastor-Corrales, M.A.(eds.). Cali, Colombia. 9-31. All, MA .1950. Genetics of resistance to the common bean mosaic virus in the bean (Phaseolus vulgaris L.). Phytopathology 40: 69-79. Alzate-Marin, A.L, H. Menarim, 6. Scares Baia, T.J. Paula JR, K. Almeida De Souza, MR. 03 Casta, E. Goncalveands De Barros, and MA. Moreira .2001. Inheritance of anthracnose resistance in the common bean differencial cultivar 62333 and identification of a new molecular marker linked to the C042 gene. J. Phytopathology 149: 259-264. Alzate-Marin, A.L, K.S. Almeida, E. G. Barros, and MA. Moreira. 2002. Validation of RAPD markers linked to 00-4 anthracnose resistance alleles in common bean cultivar Pl 207.262. Ann. Rep. Bean Improv. Coop. 45: 114-115. ASARECA. 1998. Bean common mosaic virus: Resistant varieties developed. Agriforum 4: 6- 7. ASARECA. 2002. East Africa: Viral disease hampers bean production. Agriforum 19: 3 Ariyarathne, H. M., D.P. Coyne, 6. Jung 6., JR. Steadman, P.N. Miklas, and M.J.Bassett. 1999. Molecular mapping of disease resistance genes for halo blight, common bacterial blight, and bean common mosaic virus in a segregating population of common bean. J. Amer. Soc. Hort. Sci.124 (6): 654-662. Awale, HE. and JD. Kelly. 2001. Development of SCAR markers linked to 00-42 gene in common bean. Ann. Rep. Bean Improv. Coop. 44: 119-120. Balardin, R.S, A.M. Jarosz, and JD Kelly. 1997. Virulence and molecular diversity in Colletotn'chum lindemuthianum from South, Central, and North America. Phytopathology 87: 1184-1191. 85 Balardin, RS, and JD. Kelly. 1998. Interaction between Colletotn'chum lindemuthianum races and gene pool diversity in Phaseo/us vulgaris. J. Amer. Soc. Hort. Sci. 123 (6): 1038-1047. Beaver, J.S, P.N. Miklas, J.D. Kelly, J.R. Steadman, and J.C.Rosas. 1998. Registration of PR9357-107 small red dry bean germplasm resistant to BCMV, BCMNV, and rust. Crop Sci. 38: 1408-1409. Beebe, S. and M. Pastor Corrales. 1991. Breeding for disease resistance. In: Van Schoonhoven , A. and O. Voysest (Ed.), Common Beans ; Research for crop improvement, pp 561 -61 7. Beebe, 8., PW. Skroch, J. Tohme, M. C. Duque, F. Pedroza, and J. Nienhuis. 2000. Structure of genetic diversity among common bean landraces of Middle American origin based on correspondence Analysis of RAPD. Crop Sci. 40: 264-273. Beebe, S., J. Rengifo, E. Gaitan, M. C. Duque, and J. Tohme. 2001. Diversity and origin of Andean landraces of common bean. Crop Sci. 41: 854-862. Berti, F. 1985. Synthese de travaux realises sur phaseolus aux Zaire, Rwanda et Burundi de 1945 a nos jours.ClAT 53p. Berger, PH, 8.0. Wyatt, P.J. Shiel, M.J. Silbemagel, K. Druffel, and GI. Mink. 1997. Phylogenetic analysis of the Potyvin'dae with emphasis on legume- infecting potyviruses. Arch. Virol. 142: 1979-1999. Bradeen, J.M.and P.W. Simon. 1998. Conversion of an AFLP fragment linked to the carrot Y2 locus to a simple, codominant, PCR-Based marker form. Theor. Appl. Genet. 97: 960-967. Broughton, W.J., 6. Hernandez, M. Blair, S. Beebe, P. Gepts, and J. Vanderleyden. 2003. Beans (Phaseolus spp)- model food legumes. Plant and Soil 252: 55-128. Buruchara, R.A., M.A.Pastor- Corrales, L. Afanador, C.Jara, B. Bosco, and R.Mijana. 1998. Pathogen diversity of Phaeoisariopsis griseola in Africa. Abstract 2.2.59 ICP P98. Busogoro, J.P, M.H. Jijakli, and P. Lepoivre. 1999a. Virulence variation and RAPD polymorphism in African isolates of Phaeoisariopsis griseola (Sacc.) Ferr., the causal agent of angular leaf spot of common bean. European Journal of Plant Pathology 105: 559-569. 86 FELQJ” Busogoro, J.P, M.H. Jijakli, and P. Lepoivre. 1999b. Identification of a novel source of resistance to angular leaf spot disease of common bean within the secondary gene pool. Plant Breeding 118 (5); 417-423. Coyne, D.P., J.R. Steadman, G. Godoy-Lutz, R. Gilbertson, E. Amaud- Santana, J.S. Beaver, and J. R. Myers. 2003. Contribution of the Bean lCowpea CRSP to management of been disease. Field Crops Research 4164: 1-14. Dean, L. L. and CW. Hungerford, 1946 A new bean mosaic in Idaho. Phytopathology 36: 324-326. Dean, LL. and V.E. Wilson, 1959. A new strain of common bean mosaic in Idaho. Plant Disease Reporter 43: 1 108-1110. Drijfhout, E., M. J., Silbemagel, and D. W. Bulke. 1978. Differentiation of strains of been common mosaic virus.Neth.J. Pl. Path. 84:13-26. Drijfhout, E. and L. 803, 1977. The identification of two new strains of bean common mosaic virus. Netherlands Journal of Plant Pathology 83: 13-25. Drijfhout, E. 1978. Genetic interaction between Phaseolus vulgaris and bean common mosaic virus with implications for strain identification and breeding resistance. Agricultural Research Reports 872: 1-98. Centre for Agriculture Publishing and documentation, Wageningen, Netherlands. Duvick, D.N. 1996. Plant breeding, an evolutionary concept. Crop Sci. 36: 539- 548. Ender, M. 2003. QTL analysis of genetic resistance to white mold (Sclerotinia sclerotiorum) in common bean (Phaseolus vulgaris). PhD Thesis, Michigan State University. 166 pp Fang, G.W., R.F. Allison, E.M. Zambolim, D.P. Maxwell, and R.L. Gilbertson. 1995. The complete nucleotide sequence and genome organization of bean common mosaic virus (NL3 strain). Virus Research 39: 13-23. FAO STAT Database collections 2003: http:/l apps.fao.org/page/collections. Ferreira, C.F., A. Borém, 6.A. Carvalho, S. Nietsche, T.J. Paula Jr, E.6. Barros, and MA. Moreira. 2000. Inheritance of angular leaf spot resistance in common bean and identification of RAPD marker linked to a resistance gene. Crop Sci. 40:1130-1133. Frandsen, NO. 1952. Untersuchungen zur virusresistenzztichtung bei Phaseolus vulgaris L. 1. Phytopathologische untersuchungen. Zeitschrift ftlr pflanzenzflchtung 31: 381-420. 87 Fraser, R.S.1992. The genetics of plant- virus interactions: implications for plant breeding. Euphytica 63:175—1992. Freyre, R., P.W. Skroch, V. Geffroy, A.F. Adam-Blondon, A. Shirrnohamadali, W.C. Johnson, V. Llaca, R.O. Nodari, P.A.Pereira, S.M.Tsai, J.Tohme, M. Dron, J. Nienhuis. C.E. Vallejos, and P. Gepts. 1998. Towards an integrated linkage map of common bean. 4. Development of a core linkage map and alignment of RFLP maps. Theor. Appl. Genet. 97: 847- 856. Gamez, R., A. Osores, and E. Echandi, 1970. Una raza nueva del virus del mosaico comt'In del frijol. Turrialba 20: 397-400. Galvez, 6E. and F.J. Morales. Aphid- transmitted viruses. In: CIAT (Centro lntemacional de Agricultura Tropical). 1989. Bean production problems in the tropics.2"d ed. Schwartz, HF and Pastor-Corrales, M.A.(eds.). Cali, Colombia. 333-361. Gepts, P. 1988. A Middle American and an Andean common bean gene pool. In: P. Gepts (Ed.), Genetic Resources of Phaseolus Beans; Their maintenance, Domestication, and Utilization, pp 375-390. Kluwer, London. Gepts, P. 1988. Phaseolin as an evolutionary marker. In: P. Gepts (Ed.), Genetic Resources of Phaseolus Beans; Their maintenance, Domestication, and Utilization, pp 215-241. Kluwer, London. Gepts, P. and D. Debouck. 1991. Origin, domestication, and evolution of the common bean (Phaseolus vulgaris L.). In: Van Schoonhoven , A. and O. Voysest (Ed.), Common Beans ; Research for crop improvement pp 7-53. Gepts, P. 1999. Development of an integrated linkage map. In: Singh, S.P. (Ed.), Development in plant breeding. Common bean improvement in the twenty- first century. Kluwer Academic Publisher, Dordrecht, The Netherlands, pp 53-91. Graham, RH. and P. Ranalli. 1997. Common bean. Field Crops Research 53: 131-146. Guo, P.6., G.H. Bai, and 6E. Shaner. 2003. AFLP and STS tagging of a major QTL for Fusarium head blight resistance in wheat. Theor. Appl. Genet. 106: 1011-1017. Guzman-Maldonado, SH, C. Martinez, J.A. Acosta-Gallegos, F. Guevara-Lara, and O. Paredes- LOpez. 2003. Putative quantitative trait loci for physical and chemical components of common bean. Crop Sci. 43: 1029-1035. 88 ‘cohql'\‘_‘° . Hampton, R.O., M.J. Silbemagel, and D.W. Burke. 1983. Bean common mosaic virus strains associated with bean mosaic epidemics in the northwestern United States. Plant Disease 67: 658-661. Haley, S.D., P.N. Miklas, J.R. Stavely, J. Byrum, and JD Kelly. 1993. Identification of RAPD markers linked to a major rust resistance gene block in common bean. Theor. Appl. Genet. 86: 505-512. Haley, 8.0., L. Afanador, and JD. Kelly. 1994a. Selection for monogenic pest traits with coupling and repulsion phase RAPD markers. Crop Sci. 34: 1061-1066. Haley, S.D., L. Afanador, and JD. Kelly. 1994b. Identification and application of a random amplified polymorphic DNA marker for the I gene (potyvirus resistance) in common bean. Phytopathology 84: 157-160. Haley, 8.0., L. Afanador, and JD. Kelly. 1994c. Heterogeneous inbred populations are useful as sources of near-isogenic lines for RAPD marker localization. Theor. Appl. Genet. 88: 337-342. Haley, S.D., P.N. Miklas, L. Afanador, and JD. Kelly.1994d. Random Amplified polymorphic DNA (RAPD) markers variability between and within gene pools of common bean. J. Amer. Soc. Hort. Sci. 119: 122-125. Harrison, 8.0. 2002. Virus variation in relation to resiStance breaking in plants. Euphytica 124: 181-192. Hart, L.P, AND A.W. Saettler. 1981 Bean common mosaic virus. Extension Bulletin E-1561. Hazen, S.P., P. Leroy, and R.W. Ward. 2002. AFLP in Triticum aestivum L.: pattern of genetic diversity and genome distribution. Euphytica 125: 89- 102 He, 0., K.V. Matti, and J. Mertsola. 1994. Effects of thermocyclers and primers on the reproducibility of banding patterns in randomly amplified polymorphic DNA analysis. Mol. Cell. Probes 8: 155-160. Hospital, F., C. Chevalet, and P. Mulsant. 1992. Using markers in gene introgression breeding programs. Genetics 132: 1 199-1210. Hubbeling, N. 1963. Complicaties bij de toetsing van bonerassen op resistentie tegen Phaseolus vulgaris 1 ten gevolge van het voorkomen van afwiijkende virusstammen. Mededelingen van de landbouwhogeschool en de opzoekingstations van de staat te gent 28: 1025-1033. 89 ISAR, Bean program 2001. Activity review, August 2001. p 1-8. Jacobsen, H.J. 1999. Genetic transformation In: Singh, S.P. (Ed.), Development in plant breeding. Common bean improvement in the Twenty-First Century. Kluwer Academic Publisher, Dordrecht, The Netherlands, pp 125-1 32. Jones, C.J, K.J.Edwards, S. Castaglione, M.O. Winfield, F. Sala, C. van de Wiel, G. Bredemeijer, B. Vosman, M. Matthes, A. Daly, R. Brettschneider, P.Bettini, M. Buiatti, E. Maestri, A. Malcevschi, N. Marrniroli, R. Aert, G.Volckaert, J. Rueda, R. Linacero, A.Vazquez, and A. Karp. 1997. Reproducibility testing of RAPD, AFLP and SSR markers in plants by a network of European laboratories. Molecular Breeding 3: 381-390. Johansen, IE, 0.8. Lund, C.K. Hjulsager, and J. Laursen. 2001. Recessive resistance in pisum sativum and potyvirus pathotype resolved in gene for cistron correspondence between host and virus. Journal of Virology 75: 6609-6614. Johnson, E., P.N. Miklas, J.R. Stavely, and J.C. Martinez-Cruzado. 1995 Coupling and repulsion RAPD for marker assisted selection of P|181996 rust resistance in common bean. Theor. Appl. Genet. 90: 659-664 Johnson, W. C and P. Gepts. 1994. Two new molecular markers linked to bc-3 gene Ann. Rep. Bean Improv. Coop. 37: 206-207. Johnson, W. C., P. Guzman, D. Mandala, A. BC. Mkandawire, S Temple, R. L. Gilbertson, and P. Gepts. 1997. Molecular tagging of bc-3 gene for introgression into Andean common bean. Crop Sci. 37:248-254. Khan, J.A., D. Lohuis, R. Goldbach, and J. Dijkstra. 1993. Sequence data to settle the taxonomic position of bean common mosaic virus and blackeye cowpea mosaic virus isolates. J. Gen. Virol. 74: 2243-2249. Kelly, J.D., A.W. Saettler, and MW. Adams. 1983. New necrotic strain of bean common mosaic virus in Michigan. Ann. Rep. Bean Improv. Coop. 26: 49- 50. Kelly, J.D., 6.L. Hosfield, G.V. Vamer, M.A. Uebersax, S.D. Haley, and J. Taylor. 1994. Registration of ‘Raven‘ black bean. Crop Sci. 34: 1406- 1407. Kelly, J.D. 1995. Use of Random Amplified Polymorphic DNA markers in breeding for major gene resistance to plant pathogens. HortScience 30: 461-465. 90 Kelly, J.D., L., Afanador, and SD. Haley. 1995. Pyramiding genes for resistance to been common mosaic virus. Euphytica 82: 207-212. Kelly, JD. 1997. A review of varietal response to bean common potyvirus in Phaseolus vulgaris. Plant Varieties and Seeds 10: 1-6. Kelly, JD. and P.N. Miklas. 1998. The role Of RAPD markers in breeding for disease resistance in common bean. Molecular Breeding 4 (1): 1-11. Kelly, J.D, P. Gepts, P.N. Miklas, and DP. Coyne. 2003.Tagging and mapping of genes and QTL and molecular-marker assisted selection for traits of economic importance in bean and cowpea. Field Crops Research 4173: 1- 20. Kesseli, R.V., l. Paran, and R.W. Michelmore. 1992. Efficient mapping of specifically targeted genomic regions and the tagging of these regions with reliable PCR-based genetic markers. CSSA-ASHS-AGA Joint plant breeding symposia series, Minneapolis. Crop Sci. Soc. Amer., Madison, Wis. P.31-36. Kolkman, J.M. and JD. Kelly. 2003. QTL conferring resistance and avoidance to white mold in common bean. Crop Sci. 43: 539-548. Kriegner, A., J.K. Cervantes, K. Burg, R.O.M. Mwanga, and D. Zhang 2003. A genetic linkage map of sweet potato (lpomoea batatas) based on AFLP markers. Molecular Breeding 1 1: 169-185. Lander, E.S., P.6reen, J. Abrahamson, A.Barlow, M.J. Daly, S.E. Lincoln, and L. Newburg. 1987. MAPMAKER: An interactive computer package for constructing primary genetic linkage maps with experimental and natural populations. Genomics1 : 174-181 . Landergren, ULF, M. Nilsson, and P-Y Kwok 1998. Reading bits of genetic information: Methods for single nucleotide polymorphism analysis. Genome Research 8: 769-776. Lombard, V., B.Tireau, F. Blouet, D. Zhang, C.P. Baril. 2002. Usefulness of AFLP markers to estimate varietal homogeneity of rapeseed inbred line varieties in the context of plant registration and protection. Euphytica 125: 121-127 MacPherson, J. M., P. E. Eckstein, 6. J. Scoles, and A. A. Gajadhar. 1993. Variability of the random amplified polymorphic DNA assay among thermal cyclers, and effects of primer and DNA concentration. Mol. Cell. Probes 7: 293-299 91 Masojc’, P. 2002. The application of molecular markers in the process of selection. Cellular & Molecular Biology Letters 7:499-509. McKem, N.M., G.l. Mink, O.W. Barnett, A. Mishra, L.A Whittaker, and M.J. Silbemagel, C.W. Ward, D.D. Shukla, 1992. Isolates of bean common mosaic Virus comprising two distinct potyviruses. Phytopathology 82: 923- 928. McMillin, DE. 1983. Plant lsozymes: A historical perpective. In: Tanksley, SD. and T. J. Orton (Ed.), development in plant genetics and breeding, 1A. lsozymes in plant genetics and breeding, partA. Elsevier Science Publisher, Amsterdam, The Netherlands, pp 3-13. Meksem, K., E. Ruben, D. Hyten, K.Triwitayakom, and DA. Lightfoot 2001. Conversion of AFLP bands into high throughput DNA markers. Mol. Genet. Genomics 265: 207-214. Melotto, M., L. Afanador, and JD. Kelly. 1996. Development of a SCAR marker linked to the lgene in common bean. Genome 39: 1216-1219. Melotto, M. and JD. Kelly 1998. SCAR markers linked to major disease resistance genes in common bean. Ann. Rep. Bean Improv. Coop. 41: 33- 34. Melotto, M and JD. Kelly. 2000. An allelic series at the Co-1 locus conditioning resistance to anthracnose in common bean of Andean origin. Euphytica 116: 143-149. Melotto, M. And J.D. Kelly 2001. Fine mapping of the 00-4 locus of common bean reveals a resistance gene candidate, COK-4, that encodes for a protein kinase. Theor. Appl. Genet. 103: 508-517. Méndez de Vigo, B., C. Rodriguez, A.Paneda, R. 6 irardez, and J.J. Ferreira 2002. Development of a SCAR marker linked to Co-9 in Common Bean. Ann. Bean lmpov. Coop. 45: 116-117. Meunier, JR, and P.A.D.Grimont. 1993. Facteurs affectant la reproductibilité du typage moléculaire par amplification arbitraire de l'ADN. Res. Microbiol. 144: 373-379. Michelmore, R.W., l. Paran, and RV. Kesseli. 1991. Identification of marker linked to disease resistance genes by bulked segregant analysis: a rapid method to detect markers in specific genomic regions by using segregating populations. Proc. Natl. Acad. Sci. 88: 9828-9832. 92 Mienie, C.M.S., H.Fourie, M.A. Smit, J. Van Staden, and PC. Botha. 2002. Identification of AFLP markers in soybean linked to resistance to Meloidogyne javanica and conversion to sequence characterized amplified regions (SCARs). Plant Growth Regulation 37: 157-166. Mink, GI. and M.J. Silbemagel. 1992. Serological and biological relationships among viruses in the bean common mosaic virus subgroup. Arch. Virol. Suppl. 5: 397-406. Miklas, P.N., J.R. Stavely, and JD. Kelly. 1993. Identification and potential use of a molecular marker for rust resistance in common bean. Theor. Appl. Genet. 85: 745-749. Miklas, P.N., L. Afanador, and JD. Kelly. 1996. Recombination-facilitated RAPD marker-assisted selection for disease resistance in common bean. Crop Sci. 36:86-90. Miklas, P.N., S., Lamber, G., Mink, and M. Silbemagel. 1998. Many beans with bc-3 resistance to BCMNV are susceptible to BCMV. Ann.Rep. Bean Improv. Coop. 41:33-34. Miklas, P.N., JR. Smith, R. Riley, K.F. Grafton, S.P. Singh, G.Jung, and DP. Coyne. 2000b. Marker assisted breeding for pyramided resistance to Common bacterial blight In common bean. Ann. Rep. Bean Improv. Coop. 43: 39-40. Miklas, P. N, R. C. Larsen, R. Riley, and J. D. Kelly. 2000a. Potential marker assisted selection for bc-12 resistance to bean common mosaic potyvirus in common bean. Euphytica 116: 211-219 Miklas, P.N., A.N. Hang, J.D. Kelly, C. A. Strausbaugh, and R.L. Forster. 20023. Registration of three kidney bean germplasm lines resistant to bean common mosaic virus and necrosis potyviruses: USLK-2 light red kidney, USDK-4 dark red kidney, and USWK-6 white kidney. Crop Sci. 42: 674-675. Miklas, P.N. and JD. Kelly. 2002b. Registration of Two Cranberry Bean Gennplasm Lines Resistant to Bean Common Mosaic and Necrosis Potyviruses: USCR-7 and USCR-9. Crop Sci. 42: 673-674 Morales, L.M.A. 1989. Identification of necrosis inducing strains of bean common mosaic virus and breeding for resistance. MS Thesis, Michigan State University, pp 81. 93 ‘1'. 3.. —‘__ : m_r.f".“’"I-'Ij' Moreno, R. R. Gamez, and LC. Gonzales. 1968. El virus del mosaico comI’Jn del frijol (Phaseolus vulgaris) en Costa Rica. Turrialba 18: 257-263. Nemchinova, Y.P., and JR. Stavely. 1998. Development of SCAR primers for the Ur-3 rust resistance gene in common bean. Phytopathology 88:867. Nietsche, S., A.Borém, G.A. Carvalho, R. C. Rocha, T.J. Paula JR, E.G. de Barros, M.A.Moreira. 2000. RAPD and SCAR markers linked to a gene conferring resistance to angular leaf spot in common bean. J. Phytopathology 148: 1 17-121 . Njau, P.J.R and H.F.J. Lyimo. 2000. Incidence of been common mosaic vian and bean common mosaic necrosis virus in bean (Phaseolus vulgaris L.) in the wild legumes seedlots in Tanzania. Seed Sci. &Technol. 28: 85-92. Noguchi,S., T.Tajima, Y.Yamamoto, T. Ohno, T. Kubo. 1999. Deletion of a large genomic segment in tobacco varieties that are resistant to potato virus Y (PW). Mol. Gen. Genet. 262: 822-829. Nyabyenda, P., C. Sekanabanga, L. Nyangumndi. Bean production in Rwanda. ln: CIAT (Centro lntemacional de Agricultura Tropical). 1980. Potential for field beans in Eastern Africa. Regional workshop on potential for field bean in eastern Africa, Lilongwe, Malawi, 1980. 99-121 pp. Ouédraogo, J.T., V. Maheshwari, and OK. Bemer. 2001 Identification of AFLP markers linked to resistance of cowpea (Vigna unguilata) to parasitism by striga gesnerioides. Theor. Appl. Genet. 102: 1029-1036. Pachico, D. Trends in the world Common Bean production. In : CIAT (Centro lntemacional de Agricultura Tropical). 1989 Bean production problems in the tropics.2"d ed. Schwartz, H.F. and Pastor-Corrales, M.A. (eds). Cali, Colombia.1-8 Paran, l. and R.W. Michelmore. 1993. Development of a reliable PCR- based markers linked to downy mildew resistance genes in lettuce. Theor. Appl. Genet. 85: 985-993. Pastor- Corrales, M. A., C. Jara, and SP. Singh. 1998. Pathogenic variation in, source of, and breeding for resistance to Phaeoisariopsis gn‘seola causing angular leaf spot in common bean. Euphytica 103: 161-171. Pastor- Corrales, M. A. and J.C.Tu. Anthracnose. In : CIAT (Centro lntemacional de Agricultura Tropical). 1989. Bean production problems in the tropics. 2"d ed. Schwartz, H.F. and Pastor-Corrales, M.A.(eds.). Cali, Colombia. 77-104. 94 Pedrosa, A., C.E.Vallejos, A. Bachmair, and Schweizer. 2003. Integration of common bean (Phaseolus vulgaris L.) linkage and chromosomal maps. Theor. Appl. Genet. 106: 205-212. Pejic, l., P. Ajmone-Marsan, M. Morgante, V. Kozumplick, P. Castiinoni, G. Taramino, and M. Motto, 1998.Comparative analysis of genetic similarity among maize inbred lines detected by RFLPs, RAPDs, SSRs, and AFLPs. Theor. Appl. Genet. 97: 1248-1255. Pirone, T.P. and S. Blanc. 1996. Helper dependent vector transmission of plant viruses. Ann. Rev. Phytopathol. 34: 227-247. Prins, R., J. Z. Groenewald, G. F. Marais, J.W. Snape, and R.M.D. Koebner. 2001. AF LP and STS tagging of U19, 3 gene conferring resistance to leaf rust in wheat. Theor. Appl. Genet. 103: 618- 624. Prowidenti, R., M.J., Silbemagel, and W.Y. Wang. 1984. Local epidemic of NL-8 strain of bean common mosaic virus in bean fields of western New York. Plant Disease 68: 1092-1094. Reamon-Biittner, S.M., J.Schndelmaier, and C. Jung. 1998. AFLP markers tightly linked to the sex locus in Asparagus officinalis L. Molecular Breeding 4: 91-98. Reamon-Bflttner, SM and C.Jung. 2000. AFLP-derived STS markers for the identification of sex in Asparagus officinalis L. Theor. Appl. Genet. 100: 432-438. Revers, F., 0. Le Gall, T. Candresse, and A. J. Maule. 1999. New advances in understanding the molecular biology of plant/ potyvirus interactions. Mol. Plant Micro. lnten. 12: 367-376. Sartorato, A., S. Nietsche, E.G. Barros, and MA. Moreira. 1999. SCAR marker linked to angular leaf spot resistance gene in common bean. Ann. Rep. Bean Improv. Coop. 42: 23-24 Schafer, PR and AP. Roelfs. 1985. Estimated relation between numbers of urediniospores of Puccinia graminis f.sp. tritici and rates of occurrence of virulence. Phytopathology 75: 749-750 Schneider, K.A., M.E. Brothers, and J.D Kelly. 1997. Marker assisted selection to improve drought resistance in common bean. Crop Sci. 37:51-60 Schwartz, H. F. and F. B. Peairs, 1999. Integrated pest management In: Common bean improvement in the twenty-first century. Singh, S.P.(ed) Kluwer Academic Publishers. 371-388. 95 Sengooba, T.N., N.J.Spence, D.G.A. Walkey, D.J. Allen, and A. Femi Lana. 1997. The occurrence of bean common mosaic necrosis virus in wild and forage legumes in Uganda. Plant pathology 46: 95-103. Silbemagel, M. J., 1969. Mexican strain of bean common mosaic virus. Phytopathology 59: 1809-1812. Silbemagel, M.J., L.J. Mills, and W.Y. Wang. 1986.Tanzanian strain of been common mosaic virus. Plant dis. 70:839-841. Silbemagel, M.J., G.|. Mink, R-L. Zhao, and G-Y. Zheng. 2001. Phenotypic recombination between bean common mosaic and bean common mosaic necrosis potyviruses in vivo. Arch. Virol. 146: 1007-1020. Singh, S.P., J. A. Gutierrez, A. Molina, C. Urrea, and P. Gepts. 1991. Genetic diversity in cultivated common bean. ll. Marker-based analysis of morphological and agronomic traits. Crop Sci. 31: 23-29. Singh, SP. 1999. Common bean. In: Common bean improvement in the twenty- first century. Singh, S.P. (ed) Kluwer Academic Publishers 2-24. Spence, NJ and D.G.A. Walkey. 1995. Variation for pathogenicity among isolates of bean common mosaic virus in Africa and reinterpretation of the genetic relationship between cultivars of Phaseolus vulgaris and pathotypes of BCMV. Plant Pathology 44: 527-546. Stavely, JR. 1984. Genetics of resistance to Uromyces phaseoli in a Phaseolus vulgaris line resistant to most races of the pathogen. Phytopathology 74: 339-344. Stavely, JR. 1998. Bean rust in the United States in 1997. Ann. Rep. Bean Improv. Coop 41 : 177-178. Strausbaugh, C.A., J.R., Myers, R.L. Forster, and PE. McClean. 1999. Bc-1 and Bc-u two loci controlling bean common mosaic virus resistance in common bean are linked. J. Amer. Soc. Hort. Sci. 124: 644-648. Staub, J.E., F.C. Serquen, and M. Gupta. 1996. Genetic markers, map construction, and their application in plant breeding. HortScience 31 : 729- 741. Temple, S. R and F.J. Morales. 1986. Linkage of dominant hypersensitive resistance to bean common mosaic virus to seed color in Phaseolus vulgaris L. Euphytica 35: 331-333. 96 Tanksley, SD. and RA. Jones. 1981. Application of alcohol dehydrogenase allozymes in testing the genetic purity of F1 hybrids in tomato. HortScience 16: 179-181. Tanksley, SD. 1983. Molecular markers in plant breeding. Plant. Mol. Biol. Rpt.1: 3-8. Tar’an, 8., TE. Michaels, and KP. Pauls. 2002. Genetic mapping of agronomic traits in common bean. Crop Sci. 42:544-556. Tullu, A., L. Buchwaldt, T. Warkentin, and B. Taran. 2003. Genetic of resistance to anthracnose and identification of AFLP and RAPD markers linked to the resistance gene in Pl320937 germplasm of lentil (Lens culinaris Medikus). Theor. Appl. Genet. 106: 428-434. Urrea, C. A., P.N. Miklas, J.S. Beaver, and RH. Riley. 1996. A codominant randomly amplified polymorphic DNA (RAPD) marker useful for indirect selection of bean golden mosaic virus resistance in common bean. J. Amer. Soc. Hort. Sci. 121: 1035-1039. Vallejo, V. and J.D. Kelly. 2001. Development of a SCAR marker linked to 00-5 locus in common bean. Ann. Rep. Bean Improv. Coop. 44: 121-122. Vallejo, V. and J.D. Kelly. 2002. The AFLP analysis to tag the Co-12 gene conditioning resistance to bean anthracnose. Plant and animal genome X conference 2002. http://www.int-pagorg/pagn 0/abstracts/ PAGX P233.html Vetten, H.J., D.E, Lesemann, and E. Maiss. 1992. Serotype A and B strain of bean common mosaic virus are two distinct potyviruses. Arch. Virol. 5: 415-431. Vos, R, R. Hogers, M. Bleeker, lVl. Reijans, T. van de Lee, M. Homes, A. Frijters, J. Pot, J. Peleman, M. Kuiper, and M. Zabeau. 1995. AFLP: a new technique for DNA fingerprinting. Nucleic Acid Research 23: 4407-4414. Vuylsteke, M., R.Mank, B.Brugmans, P. Stam, and M. Kuiper. 2000. Further characterization of AFLP data as tools in genetic diversity assessments among maize (Zea mays L.) inbred lines. Molecular Breeding 6: 265-276 Wang, W-Y 1983. Serology of BCMV strains. MSc Thesis, Washington State University 50 pp. Wayne, R, M. Morgante, A. Chaz, M. Hanafey, J. Vogel, T. Scott, and A. Rafalski. 1996. The comparison of RFLP, RAPD, AFLP and SSR 97 (microsatellite) markers for germplasm analysis. Molecular Breeding 2 (3): 225-238. Weeden, N.F., G.M. Timmerman, M. Hemmat, B.E. Kneen, and M.A.Lodhi. 1992. Inheritance and reliability of RAPD markers. CSSA-ASHS-AGA Joint plant breeding symposia series, Minneapolis. Crop Sci. Soc. Amer., Madison, Wis. P.12-17. Young, RA. and J.D. Kelly. 1996. RAPD markers flanking the Are gene for the Anthracnose resistance in common bean. J. Amer. Soc. Hort. Sci.121: 37- 41. Young, R.A. and J.D. Kelly. 1997. RAPD markers linked to three major anthracnose resistance genes in common bean. Crop Sci. 37: 940-946. Young, R.A., M. Melotto, R.O. Nodari, and J.D. Kelly. 1998. Marker assisted dissection of the oligogenic anthracnose resistance in the common bean cultivar, G2333. Theor. Appl. Genet. 96: 87-94. Yu, K, S.J. Park, V. Poysa, and P. Gepts. 2000. Integration of simple sequence repeat (SSR) markers into a molecular linkage map of common bean (Phaseolus vulgaris L.). Journal of Heredity 91: 429-434. Zaumeyer, W.J. and R.W. Goth, 1964. A new severe symptom-inducing strain of common bean mosaic virus. Phytopathology 54: 1378-1385. 98 “mm .‘rfl Illlllllljlllllll]llljljllllllllljllI