QTL MAPPING OF SYMBIOTIC NITROGEN FIXATION IN DRY BEAN; DRY BEAN PERFORMANCE UNDER ORGANIC PRODUCTION SYSTEMS By James A. Heilig A DISSERTATION S ubmitted to Michigan State University in partial fulfillment of the requirements f or th e degree of Plant Breeding, Genetics and Biotechnology , Crop and Soil Sciences - Doctor of Philosophy 2015 ABSTRACT QTL MAPPING OF SYMBIOTIC NITROGEN FIXATION IN DRY BEAN; DRY BEAN PERFORMANCE UNDER ORGANIC PRODUCTION SYSTEMS By James A. Heili g Michigan has been a leader in organic dry bean ( Phaseolus vulgaris L) production. Previous research has found that dry bean yields were substantially lower under organic conditions compared with adjacent conventional production. Since pests are control led with approved methods in each respective system, fertility appears to be an issue where the two systems may differ. Seventy - nine black and navy bean elite breeding lines and commercial checks, and a non - nodulating check were evaluated for yield under organic conditions in 3 MI locations in 2011 through 2013. These same genotypes were also assayed for nodulation characteristics, N fixation, and shoot and root growth in the greenhouse under N free conditions. Several traits measured in the greenhouse w ere significantly correlated to traits measured in the field. In particular, percent N derived from the atmosphere (%Ndfa) in the greenhouse was correlated with seed yield, N yield, and %Ndfa in the field for most site years , suggesting that enhancing sym biotic nitrogen fixation (SNF) traits could improve productivity in organic bean systems. Variability for SNF ability has been reported within P . vulgaris . The black bean landrace ever is poorly adapted to cultivation in northern latitudes due to long season maturity and indeterminate type III growth habit. The recombinant inbred line (RIL) population developed by crossing Puebla 152 with the was used to investigate the inheritance of enhanced SNF ability. The RIL population consisted of 122 lines and was evaluated in the greenhouse under N free conditions, and under low N conditions in the field in East Lansing (EL), MI and in Isabela, Puerto Rico (PR). The %Ndfa averaged between 12.7 % up to 66.6 %, although individual RILs ranged up to 90.5 %Ndfa. Traits measured in the greenhouse such as shoot biomass and biomass difference correlated moderately with yield and %Ndfa traits measured in the field. A quantitative trait loci (QTL) analysis of the phenotypic data from the field and greenhouse was conducted using single - nucleotide polymorphism (SNP) markers developed through the BeanCAP . The phenotypic data included traits for yield, nodule rat ing , biomass growth, agronomic traits, and N fixation. A total of 19 QTL associated with SNF traits were identified on all 11 chromosomes except Pv02 and large clusters of QTL were discovered on Pv01, Pv06, and Pv08. Many of the QTL associated with %Ndfa, N harvest index, and %N in biomass were also associated with candidate genes expressed in the nodules and roots. Candidate genes such as Phvul.006G146400 , which is a chitin elicitor receptor kinase is involved in recognition of rhizobia in the early estab lishment of the symbiotic relationship. Other candidate genes are transcription factors, such as Phvul.006G034400 that is associated with %Ndfa determined by natural abundance 15 N analysis , is a MADS - box family gene and is expressed in young and mature green pods. The majority of QTL associated with genes expressed in the root or nodule are derived from Puebla 152 while QTL associated with genes with enhanced expression in ste ms and pods are associated with Zorro. This follows a pattern where Puebla 152 has superior SNF ability, whereas Zorro is highly efficient in partitioning the fixed - N into the seed. The QTL described serve as potential targets for improvement of SNF char acteristics in adapted commercial dry bean genotypes. iv TABLE OF CONTENTS LIST OF TABLE S v i LIST OF FIGURES ix CHAPTER 1 LITERATURE REVIEW 1 Introduction 1 S ymbiotic Nitrogen Fixation i n Beans 2 Role o f Rhizobia 7 Rhizobia a nd Other Benefits 10 Puebla 152 Genotype 13 Methods t o Measure S NF 14 Q TL Analysis a nd S NF 18 Summary 19 LITERATURE CITED 21 CHAPTER 2 EVALUATION OF SYMBIOTIC N FIXATION OF BLACK AND NAVY BEAN ADVANCED BREEDING LINES UNDER ORGANIC PRODUCTION SYSTEMS 2 8 Abstract 2 8 Introduction 2 9 Materials a nd Methods 3 1 Plant Material 3 1 Inoculation 3 2 Planting a nd Cultivation 32 Data Collection 3 3 Harvest 3 3 Nitrogen Analysis 3 4 Gre enhouse Assay 3 4 Statistical Analysis 3 5 Results a nd Discussion 35 Conclusions 40 APPENDICES 41 APPE NDIX A - CHAPTER 2 TABLES 4 2 APPENDIX B - CHAPTER 2 SUPPLEMENTAL TABLES 50 LITERATURE CITED 6 2 CHAPTER 3 NITROGEN FIXATION ABILITY OF A PUEBLA 152/ZORRO RIL POPULATION EVALUATED UNDER GREENHOUSE AND FIELD CONDITIONS 6 5 Abstract 6 5 Introduction 6 5 Materials a nd Methods 6 9 Plant Material 6 9 Field Trials 70 v Data Collection 7 2 Harvest 7 2 Nitrogen Analysis 7 3 Greenhouse Assay 7 3 Statistical Analysis 7 4 Results a nd Discussion 7 5 Conclusions 8 1 APPENDICES 8 3 APPENDIX A CHAPTER 3 TABLES 8 4 APPENDIX B CHAPTER 3 SUPPLEMENTAL TABLES 9 5 LITERATURE CITED 121 CHAPTER 4 QTL ANALYSIS OF SYMBIOTI C NITROGEN FIXATION IN THE PUEBLA 152/ZORRO DRY BEAN RIL POPULATION 12 4 Abstract 12 4 Introduction 12 5 Materials a nd Methods 12 8 Single Nucleotide Polymorphism ( S NP ) Genotyping 12 8 Linkage Map Construction 12 9 Q TL Analysis 12 9 Results a nd Discussion 1 30 APPENDICES 13 7 APPENDIX A CHAPTER 4 FIGURES AND TABLES 13 8 APPENDIX B CHAPTER 4 SUPPLMENTAL FIGUR E 14 7 LITERATURE CIT ED 1 51 vi LIST OF TABLES Table 2.01. Hectares planted and production of organically produced dry bean in the United States 4 2 Table 2.02. Field location, year , planting date, precipitation, soil type and soil chemistry of fields where black and navy bean genotypes were evaluated under organic conditions in 2011, 2012, and 2013 in Michigan. 4 3 Table 2.03. Seed yield and N yield of 79 black and navy bea n genotypes grown under organic conditions in Frankenmuth, Caro, and Wisner, MI in 2011, 2012, and 2013. 44 Table 2.04. Percent N derived from the atmosphere (%Ndfa) and percent N in seed of 79 black and navy bean genotypes grown under organic conditions i n Frankenmuth, Caro, and Wisner, MI in 2011, 2012, and 2013. 4 5 Table 2.05. Pearson correlations for %Ndfa (lower left) and %N (upper right) for 79 black and navy bean genotypes grown under organic conditions in Frankenmuth, Ca ro, and Wisner, MI in 2011, 2012, and 2013 4 6 Table 2.06. Traits measured on 79 black and navy dry bean genotypes grown under N free conditions in the greenhouse in East Lansing MI in 2011 and 2012 . 4 7 Table 2.07 . Pearson Correlations of tra its measured of 79 black and navy bean genotypes grown under N free conditions in the greenhouse in East Lansing, MI in 2011 and 2012 . 4 8 T able 2.08. Pearson correlations for field traits and greenhouse traits measured for 79 black and navy bean genotyp es grown in the field under organic production systems and in the greenhouse in East Lansing, MI under N free conditions in 2011 and 2012 . 4 9 Table S 2 .0 1 . Seed yield and N yield of 79 black and navy bean genotypes grown under organic condi tions in 2011, 2012, and 2013 in Frankenmuth, Caro, and Wisner, MI . 50 Table S 2. 02 . Percent N derived from the atmosphere (%Ndfa) and percent N in seed of 79 black and navy bean genotypes grown under organic conditions in 2011, 2012, and 2013 in Frankenmuth, Caro, and Wisner, MI. 5 4 Table S 2. 03 . Traits measured on 79 black and navy dry bean genotypes grown under N free conditions in the greenhouse in East Lansing MI. 5 8 Table 3.01. Planting date, precipitation, and soil characteristics of plots in East Lansing, MI, where Puebla 152/Zorro RILs were grown. 8 4 Table 3.02. Yield of commercial checks and Puebla 152/Zorro RIL lines grown in East Lansing in 2011 to 2013 and Puerto Rico in 2012 and 2013. 8 5 vii Table 3.03. Percent N in the biomass and seed of commercial checks and Puebla 152/Zorro RIL lines grown in East Lansing in 2011 to 2013 and Puerto Rico in 2012 and 2013. 8 6 Table 3.04. Percent N derived from the atmosphere (%Ndfa) calculated using th e natural abundance method ( 15 N ) and the difference method for checks and RILs of the Puebla 152/Zorro RIL population grown in East Lansing, MI in 2011 to 2013 and Isabella, Puerto Rico in 2012 and 2013 . 8 7 Table 3.05. Amount of N ( kg ha - 1 ) in seed and biomass, and N harv est index (NHI) of Puebla 152/Zorro RILs grown in East Lansing, MI and Puerto Rico in 2011, 2012, and 2013. 8 8 Table 3.06. Flowering, maturity, and plant height of Puebla 152/Zorro RILs grown in East Lansing, MI and Puerto Rico in 2011, 2012, and 2013. 8 9 Table 3.07. SNF traits measured in the greenhouse in East Lansing, MI on Puebla 152/Zorro RILs grown in the greenhouse growing in N free conditions. 90 Table 3.08. Pearson correlations between traits measured in the field in 2011 - 2013 and in the greenhouse on the Puebla 152/Zorro RIL population grown in East Lansing, MI. 9 1 Table 3.09. Pearson Correlations between traits measured in the field on the Puebla 152/Zorro RIL population grown in East Lansing, MI. 9 2 Table 3.10. Pea rson Correlations between traits measured on the Puebla 152/Zorro RIL population grown under N - free conditions in the greenhouse in East Lansing, MI. 9 3 Table 3.11. Pearson Correlations between traits measured in the field on the Puebla 152/Zorro RIL p opulation grown in Isabela, Puerto Rico and in the greenhouse in East Lansing, MI . 9 4 Table S 3. 01 . Yield of commercial checks and Puebla 152/Zorro RIL lines grown in East Lansing in 2011 to 2013 and Puerto Rico in 2012 and 2013. 9 5 Table S 3. 02 . Percent N in the biomass and seed of commercial checks and Puebla 152/Zorro RIL lines grown in East Lansing in 2011 to 2013 and Puerto Rico in 2012 and 2013. 9 9 Table S 3. 03 . Percent N derived from the atmosp here (%Ndfa) calculated using the natural abundance method ( 15 N ) and the difference method for checks and RILs of the Puebla 152/Zorro RIL population grown in East Lansing, MI in 2011 to 2013 and Isabella, Puerto Rico in 2012 and 2013 . 10 3 Table S 3. 04 . Amount of N ( kg ha - 1 ) in seed and biomass, and N harvest index (NHI) of Puebla 152/Zorro RILs grown in East Lansing, MI and Puerto Rico in 2011, 2012, and 2013. 10 9 viii Table S 3. 05 . SNF traits measured in the gr eenhouse in East Lansing, MI on Puebla 152/Zorro RILs grown in the greenhouse growing in N free conditions. 11 5 Table 4.01. Quantitative trait loci (QTL) for biomass , agronomic and S N F traits in the Puebla 152/Zorro RIL population grown in the field in EL and PR in 2011 - 2013 and in greenhouse under N free condi tions in East Lansing, MI . 13 8 ix LIST OF FIGURES Figure 4.01. Dry bean chromosomes Pv01, Pv03, Pv04, Pv05, Pv06, Pv07, Pv08, Pv09, Pv10, and Pv11 showing QTL for Symbiotic N - fixation (SNF) from the N - free greenhouse analysis and in the field in East Lansing, MI and Isabela, Puerto Rico in 2011 to 2013. 14 1 Figure S 4.0 1 . Linkage map of the Puebla 152/Zorro population. 14 7 1 CHAPTER 1 LITERATURE REVIEW Introduction Dry bean ( Phaseolus vu lgaris L.) is an important food crop providing a nutrient dense, high protein, and low calorie staple while delivering up to 35% of global dietary protein (Broughton et al. 2003) . P roduction worldwide has increased from 17.5 M tonnes in 1990 to 23.1 M ton nes in 2013 ( FAOSTAT , 2015 a ) representing a 32% increase over the time period with 35.4% of total production being in the Americas. The United States ranked 6 th in production in 201 3 ( FAO STAT , 2015 b ) behind China, India, Myanmar, Brazil and Mexico ; however, this may be Phaseolus vulgaris . Michigan is one of 18 states with major production of dry beans and ranks 2 nd in total dry bean production, after North Dakota ( USDA - ERS, 2015 ) . The two major market classes grown in Michigan are black beans and navy beans and Michigan is the leading producer of black beans in the country and the second leading producer of navy beans after N orth Dakota ( USDA - ERS, 2015 ). In the United States, black beans , 95,590 ha planted, rank behind the leading market class es of pintos with 246,170 ha planted and navy beans 100,240 ha planted ( USDA - NASS , 2015 ) . From 2008 to 2011 organic dry bean production in the U.S. has increased by 44.3% while area planted in Michigan increased from 1,960 ha to 3,545 ha (USDA - NASS , 2015 ). In both 2008 and 2011 , Michigan was the leading producer of organic dry beans. The leading market class produced organically in Mich igan is black beans, with an increasing interest in other market classes ( Findlay and Sattelberg personal communication). Heilig and Kelly (2012) showed that beans grown under certified organic conditions yielded on average 20% lower than those grown on a djacent conventional fields. Those genotypes belonging to the Andean gene pool performed poorly compared to genotypes from Middle American gene pool and yielded 25% less overall. 2 Heilig and Kelly (2012) noted that those genotypes performing poorly in org anic production systems also yielded poorly in conventional production systems. Organic production systems rely on addition of nutrients to the soil through amendments such as composts and manure (Hill, 2014) . C over crops are used in these systems to both fix nitrogen (legume cover crops) and retain nutrients in the soil (non - leguminous cover crops) which prevent leaching form the soil. Dry bean was domesticated in a region from Central America south to the Andes region of South America ( Belluc c i et al., 2014; Schmutz et al., 2014). Prior to domestication, P . vulgaris had begun to diverge into two distinct populations with partial reproductive barriers (Gepts, 1998) . The Middle American Gene Pool originate d in Mexico and Central America while the Andean Gene Pool originate d in the Andes region of South America ( Schmutz , et al., 2014: Singh, et al., 1991). Each gene pool is further divided into multiple r aces based on morphological , allozyme and molecular differences ( Blair et al., 2013 ) . Singh et al. ( 1991) divide d the Middle American Gene Pool into Races Jalisco, Durango, and Mesoamerica and Beebe et al. (200 0 ) later adding Race Guatemala. The Andean gene pool is divided into three races , Peru, Nueva Granada, and Chile. S ymbiotic Nitrogen Fixation i n Beans Nearly a century ago Sevey (1918) known to agricultur a reference to the ability of dry bean to acquire N from the atmosphere through the association with soil bacteria, Rhizobium . Yet, i n the 21 st Century, why do dry bean producers still need to apply N fertilizers and other soil amendments to achieve competitive yields in dry bean? Nitrogen a pplication recommendations range from 11 kg ha - 1 without irrigation in Michigan ( M SU, 2015 ) up t o 23 kg ha - 1 under irrigated conditions in Nebraska (Hergert and Schild, 2013) . 3 Dry bean is often considered a poor N fixer (Piha and Munns, 1987; Fage ria et al., 2014 ) i n comparison to soybean ( Glycine max (L.) Merr. ) and chickpea ( Cicer arietanum L. ) . Piha and Munns (1987) conducted an acetylene reduction assay to determine the activity of the nitrogenase in the nodules of the species . They noted that dry bean evolved more H 2 during fixation than soybean or chickpea representing a reduction in efficienc y of symbiotic nitrogen fixation ( SNF ) for bean. They also compared the size and number of the nodules and discovered that the nodules of dry bean were smaller but more numerous than soybean or chickpea. Piha and Munns (1987) also noted that the period b etween germination and flowering of dry bean was much shorter, 27 days in their study, compared to chickpea, which averaged 34 days to flowering. In addition, the interval between flowering and physiological maturity was much shorter for chickpea suggesti ng that the chickpea simply had more time in a vegetative state to establish nodules and fix N before the strong sink strength of the seed for photosynthate competes with the nodules for resources (Piha and Munns, 1987). In an effort to better calculate t he contribution of pulse crops to soil N levels, Walley et al. (2007) used published data from the Northern Great Plains area to investigate the contributions of N fixation in pulse crops such as pea ( Pisum sativum L. ), lentil ( Lens culinaris Medik. ) , chic kpea, dry bean and faba bean ( Vicia fabia L. ) . Similar to other findings Walley et al. (2007) determined that dry bean had the lowest average percent nitrogen derived from the atmosphere ( %Ndfa ) (40 %Ndfa ) and also had the highest amount of variability yea r to year and by location among the pulse crops studied. Faba bea n was the highest fixer achieving 84 % Ndfa . Symbiotic nitrogen fixation (S NF) is a complex trait. Not only must the plant be able to form compatible symbioses with the appropriate rhizoba cteria, it must also form sufficient nodule mass and effectively move fixed N through the plant to the seeds. Nodule number has been 4 shown to vary among dry bean genotypes (Pereira et al., 1993). There was a significant correlation (r 2 =0.64, p<0.01) betw een nodule number and N fixed in a population of dry beans bred for enhanced N fixation (Pereira et al., 1993). There is considerable variation in this trait within dry bean germplasm (Wolyn et al., 1991; Pereira et al., 1993; Fageria et al., 2014). Impro vement should be possible as dry bean appears to be responsive to selection for improved SNF by selecting directly or indirectly for fixed N (Wolyn et al., 1991; Elizondo Barron et al., 1999). St. Clair and Bliss (1991) selected four inbred backcross lin es from their Puebla 152/ Sanilac population which showed superior acetylene reduction assay ( ARA ) levels. These plants were intercrossed and the F 3 progeny were tested for their ability to fix N. The majority of the 25 resultant progenies were superior N fixers when compared to Sanilac. S everal of the lines studied fixed N similar to high N - fixing parent Puebla 152 while having agronomic traits similar to Sanilac which would make them more amenable to direct harvest (St. Clair and Bliss, 1991). In addi tion to the ability to fix N, efficien t use of N is important. Fageria et al. (2013) noted variability among the 20 dry bean genotypes for nitrogen use efficiency (NUE) . Values rang ed from 7.3 mg mg - 1 seed in genotype BRS Valente for each mg N applied t o 21.2 mg mg - 1 for line CNFP 7624. However, Fageria et al. (2013) did not mention if the potting mix used was sterilized nor if the plants were nodulated , as t heir N - free treatment yielded nearly as much seed N (43.6 g kg - 1 for the zero N treatment compar ed to 46.9 g kg - 1 in the fertilized treatment) though none was intentionally added. The source of this N was fixation. Thus, traits associated wi th partitioning likely interact with SNF to achieve enhanced yield. Phaseolus vulgaris with many different strains of rhizobacteria from several different species and genera ( Michiels 5 et al., 1998; Herrera - Cervera, et al., 1999 ; Ribeiro et al., 2013 ). Using 100 different rhizobacterial s trains isolated from nodules of a wide range of host plants in Fabacea, Michiels et al. (1998) discovered that the majority of the strains were able to form nodules in either or both (Mesoamerican) Limburgse Vroege (Andean) . The rhizobial genera included Rhizobium, Bradyrhizobium, Azorhizobium, Mesorhizobium, and Sinorh i zobium (Michiels et al., 1998). Not all strains were equally able to form nodules on the two dry bean genotypes studied nor were all strains forming nodules able to fix N (fewer than 70% were able to fix N ) on either or both dry bean genotypes . Michiels et al. (1998) concluded that Carioca had higher level of nodulation compared to Limburgse Vroege . This may indicate variability within P . vulgaris itself to form nodules and fix N and perhaps suggesting that there may be differences in fixation between members of the Andean and Mesoamerican gene pool . Herrera - Cervera et al., (1999) investigated the diversity of rhizobia inhabiting nodules on dry bean genotyp Most strains identified belong ed to Rhizobium etli, R. girardinii, R. gallicum, R. leguminosarum, and Sinorhizobium fredii (Herrera - Cervera et al., 1999) . These strains we re isolated from nodules formed on plants grown in Spain. O ne of the species, R. etli , is an American species nodulating P. vulgaris while Sinorhizobium are generally known to nodulate soy bean . The authors speculate that the R. etli strains must have bee n transported to Spain with bean seeds imported from South or Central America (Herrera - Cervera et al., 1999). Rhizobacteria are typically found in soils where beans are grown, though not all rhizobacteria are able to effect nodulation on dry bean or are able to fix N in symbios i s with dry bean. Mora et al. (2014) investigat ed rhizobial strains that could form nodules, but not fix N in association with dry bean. What at first appeared to be contamination of their non - fixing bacterial cultures with 6 fixing bacteria led to a very interesting finding: several strains of R . phaseoli and R. leguminosarum persisted within the seed tissue protected by the seed coat. These endophytic rhizobial strains were able to multiply and form nodules during development of the bean plant (Mora et al., 2014). While seeds are inoculated with rhizobia following various me dia such as peat, clay, or liquid, the discovery of endophytic rhizobial strains offer an alternate method. ul method in regions of the world where producing or obtaining Rhizobium inoculant is difficult or impossible (Mora et al., 2014). Aside from the variation noted in S NF ability in dry bean and in the efficiency with which different rhizobia strains fix N , dry bean nodule development is variable resulting in genotypes producing nodules of different size. Rodino et al. (2011) identified two main groups of nodules - - the other - (SNO) whi le investigating the S NF ability of a diverse group of 128 dry bean lines . Plants producing nodules over 2 mg nodule - 1 were considered BNO while plants producing nodules less than 1.5 mg nodule - 1 were considered SNO (Rodino et al., 2011). Nodules of BNO plants were concentrated on the crown roots and were lower than average in number while nodules of the SNO plants were spread throughout the root system. BNO was associated with plants that produced a greater above ground biomass which was interpreted to mean that the BNO plants were more efficient fixers (Rodino et al., 2011). Interestingly, the small size and diffuse distribution of nodules on the high N - fixing Puebla 152 would best be described as SNO under the parameters described by Rodino et al. (20 11) (personal observation). Wolyn et al. (1989) similarly associated small nodules distributed diffusely throughout t he root system with improved S NF . 7 Role of Rhizobi a The interaction of N level and inoculation (+ or - ) of 15 Brazilian dry bean genotype s was investigated by Fageria et al. (2014) . They found that there was considerable variation among the different genotypes for yield under different N levels and whether the plants were inoculated or not. Aside from seed yield, traits measured included shoot biomass, root biomass, 100 seed weight, and seeds per pod, all of which may be considered components of yield. They conclu ded that dry beans were poor nitrogen fixers and maximum yields could only be obtained by the addition of high rates of fertili zer N. Apparently this conclusion was based solely on the average seed yield for all 15 genotypes in each treatment taken together . Looking at the individual lines it becomes apparent that the yield response varies considerable among the genotypes as som e, such as CNFC 10408 yielded more withou t additional fertilizer N with r hizobial inoculation than with the addition of 200 mg N kg - 1 soil yielding 10.0 g seed plant - 1 vs 7.7 g seed plant - 1 , respectively. For comparison, the average yield per plant for th e 15 genotypes without N fertilizer but with rhizobium was 9.83 g and 11.66 g for the 200 mg N kg - 1 soil treatment. The genotype CNFC 10408 performed similarly for shoot dry weight, seeds per pod, 100 seed weight, and root dry weight. A more appropriate conclusion would be that there is substantial variability for performance under various N and rhizobia rates. Fageria et al. (2014) also concluded that rhizobia inoculant with the addition of a small amount of N fertilizer was detrimental to yield when co mpared to the high N treatment without rhizobia. Perhaps this suggests that there is a diversion of resources to either the nodules themselves or the rhizobia since root dry weight of the N plus rhizobia inoculant was lower than the control (no N or rhizo bia). Contrary to these findings, Muller et al. (1993) found that mineral N application increased N fixation though some SNF traits were affected differently, such as an increase in 8 nodule size and biomass with the application of N early allowing for enha nced N fixation after flowering. Puebla 152 cultivar was not as responsive to mineral N application as the other high fixing dry bean genotype Negro Argel (Muller et al., 1993). The form of the N supplied in the soil may have an impact on the extent to which nodulation and fixation is reduced. Hine and Sprent (1987) used nitrate, ammonia, and urea to study the impact of the source of N on the growth of bean plants. Both nitrate and ammonia application resulted in a significantly lower number of nodules , whereas u rea did not depress the number of nodules formed. Hine and Sprent (1987) tested urea levels from 0 mol m - 3 to 10 mol m - 3 and found that nodule levels were not affected with application rates up to 4 mol m - 3 . Application of any N source, as well as the increasing levels of urea resulted in increased biomass. Najareddy et al. (2014) found that higher levels of nitrate did not reduce the number of infection sites but did reduce the development of those infection sites into nodules. Higher levels of nitrate resulted in taller shoots and more biomass while roots were shorter with increased nitrate levels regardless if the nitrate was provided for the first 5 d after germination or continuously until flowering (Najareddy et al., 2014). Application of a starter fertilizer at planting might actually help improve N fixation especially if levels are low enough to not hinder the development of the nodule after initial infection. It seems that some amount of soil N early in development is actually benef icial to the establishment and growth of nodules. The source of N may vary throughout the growth cycle of dry bean with vegetative N early in development being primarily soil N which is depleted during growth of the plant and establishment of nodules and S NF when fixed N becomes the dominant source (Thomas et al., 1984 ; St Clair et al., 1988 ; Lynch and White , 1992 ). The dependence on fixed N during the reproductive cycle may explain why Hungria and Neves (1987) found that 60 to 64% of N in seed was fixed N . Soil N is used to 9 grow the early vegetative portions of the plant. As soil N levels decline and the nodules are fully developed and functioning the seed is beginning to be the sink for any N fixed. Only as the plant approa ches physiological maturity i s N in leaves, stems, roots, and pods remobilized to the seed. Nitrogen fixing plants invest a considerable amount of resources into establishing symbioses and subsequent N fixation which consumes up to 30% of the photosynthate produced by the plant (Schu bert, 1986). O ther crops belonging to the Fabacea are able to fix N similar to dry bean . Kim et al. (2013) suggest that SNF is a basic and integral characteristic of legume s pecies. Utilizing a set of 20 S NF related genes in soybean, chickpea, Lotus ja ponicas , Medicago truncatula , pigeon pea ( Cajanus cajan ), and dry bean , Kim et al. (2013) found that there was a high level of conservation among these six species. S oybean and chickpea were the most closely related based on sequence of the 20 genes inves tigated, followed by soybean and dry bean whereas d ry bean and M . truncatula were the most distantly related pair (Kim et al., 2013). In field pea early SNF is linked to the developing the pho tosynthate avail ability drives S NF (Liu et al., 2013). T he C supplied to nodules is relatively constant during early development, but is reduced during the transition to the reproductive stage, and then rises during pod fill through maturity. This pattern seems to follo w the pattern observed in dry bean, with highest demand for N occurring during pod fill. A QTL analysis of 207 genome associated with SNF characteristics (Bourion et a l. , 2010). Many of the QTL that were related to traits such as shoot and root biomass production, %Ndfa , accumulated C, and nodule number colocalized within the genome. Soybean is often cited as superior to dry bean in N fixation ability. As a result of this enhanced ability supplemental N is not typically provided to 10 soybean which relies completely on SNF for N requirements . Although SNF is usually sufficient to achieve maximum yields in soybean the high - yielding modern genotypes may be reaching their maximum SNF ability according to Nicolas et al. (2006) . This group found several QTL associated with nodule number, nodule dry weight, and shoot dry weight which are all traits often associated with SNF. Nicolas et al. (2006) found several interactions b etween unlinked loci for shoot dry weight, nodule number, and nodule dry weight and an epistatic interaction between loci for nodule number and nodule dry weight. These results demonstrate the complexity of SNF while offering several QTL which may be util ized in marker assisted selection (MAS) to further improve SNF ability in soybean to meet the N demands of higher performing genotypes . Rhizobia and O ther B enefits Benefi cial effec ts of the symbiosis of rhizobia and dry bean go beyond S NF resulting in in direct benefits such as nutrient acqu isition and control of disease (Yadegari et al., 2010; Abbaszadeh - dehaji et al., 2012; Neila et al., 2014; Ahemad and Kibret, 2014 ). The interaction between rhizobia and other plant growth promoting rhizobacteria (PGPR ) such as coinoculation of dry bean plants with other rhizobacteria , such as Pseudomonas fluorescens and Azospirillum lipoferum further increas ed N fixation, biomass accumulation and protein con tent (Yadegari et al., 2010). These benefits are not limited to legume crops but extend to other crops such as corn ( Zea mays ) and wheat ( Triticum aestivum ) (Ahemad and Kibret, 2014). The diversity of the soil microbial community may be affected by inoculation with rhizobia. Trabelsi et al. (2011) noted an increas e in the microbial Coco inoculated with different strains of rhizobium: Sinorhizobium ( Ensifer ) meliloti strain 4H41 and Rhizobium gallicum strain 8a3. T here was little effect on the soil content o f nitrate, phosphate, 11 or ammonium; however, there was a significant increase in the number of bacterial species present in the soil whether the inoculants were applied individually or combined especially as the season progressed with greatest differences b eing seen at harvest (Trabelsi et al., 2011). N ot all of the bacteria were identified, as many belong to groups known for their benefit to plants, including many rhizobia and actinorhyzal species. The control, which had no inoculation but fertilizer show ed only a modest growth in the diversity of soil bacteria (Trabelsi et al., 2011). While an increase in soil bacterial diversity would be considered an indirect benefit it could clearly improve productivity as many of the bacteria may prove beneficial to t he growth of plants. Soil nutrient availability may also be enhanced by some strains of rhizobium. Abbaszadeh - dehaji et al. (2012) found that 14 rhizobium strains selected for their high symbiotic effectiveness were also able to produce growth enhancing phytohormones such as auxin , solubilize P and Zn, and produce siderophores which are involved in chelating soil Fe and mobiliz ing the Fe into plant roots. Improved availability of soil nutrients is not only benefici al to plant growth but also to S NF acti vity of the plant. Under P deficient conditions, S NF may be reduced (Lazali et al., 2014 ). W hen P was sufficient the nodules were effective at excluding O 2 , which is necessary for the proper function of nitrogenase responsible for reducing N 2 in the nodul e . U nder P deficiency the nodules were more permeable to O 2 thus reducing S NF (Lazali et al. , 2014) . Studies by Neila et al. (2014) using both soluble and insoluble P showed that nodule number on dry bean variety Coco - blanc was reduced in plants with in sufficient P and that different rhizobial strains from Tunisia were able to solubilize P at different rates. Under soluble P conditions, Rhizobium sp. strain P.Bj.09 was inferior to Rhizobium tropici strain CIAT899, whereas R. sp. strain P.Bj.09 formed mo re nodules on dry bean variety Coco - blanc when only 12 insoluble P was applied (Neila et al., 2014). Thus P supply can impact nodule formation, development , and function . Piha and Munns (1987) suggested that the release of H + represented a loss of energy fro m the symbiotic system in dry bean and might explain the reduced S NF ability of dry bean compared to other legume crops . H owever, Alkama et al. (2012) noted that the release of H + was effective at acidifying the rhizosphere and thus solubilize P from the soil especially when P was limiting . While the efflux of H + may represent a loss of energy it may serve a greater purpose in making available P thus the energy cost may be offset by the P obtained . The presence of rhizobia and other rhizobacteria have re al benefit to not only N fixing legumes but also to a diverse range of crop plants, w hether through direct action such as forming a symbiosis with the plant to fix N, production of siderophores which chelate metals like Fe, soil acidification for P availab ility, providing competition to pathogenic organisms, or producing plant hormones such as IAA , (as reviewed in Ahemed and Kibret, 2014) . The benefits of rhizobial inoculation go beyond nutrient availability and into defense of pests such as Mexican bean beetles ( Epilachna varivestis Mulsant). Comparing lima bean ( Phaseolus lunatus L.) which had been inoculated with Bradyrhizobium sp. to uninoculated controls, Ballhorn et al. (2013) found that the volatile organic compounds (VOCs) released by colonized li ma bean plants were more repellent to the Mexican bean beetle than the VOCs released by non - colonized plants. Once the jasmonic acid (JA) pathway was induced by mechanical damage, insect damage, or application of JA those plants which were colonized by th e rhizobial strain caused the bean beetles to avoid them , whereas the same repellent bouquet was not produced by the non - colonized plants (Ballhorn et al., 2013). Nod factors, signaling molecules produced by rhizo bia living free in the soil, when extracted from rhizobacteria have plant growth enhancing characteristics. Pea seed treated with an extract 13 of liquid culture of R . leguminosarum bv. viciae GR09 germinated in 50% less time that the water control (Podlesny et al., 2014) . In addition, leaf area and green pods were increased significantly in plants sprayed with the extract over plants not sprayed ( Podlesny et al., 2014). Other benefits of the nod factor extract included an increase in chlorophyll levels, wh ich is correlated with the N status of the plant implying that though not inoculated with rhi zobia, plants sprayed with the N od factor extract had higher levels of N in their tissue than plants not sprayed (Podlesny et al., 2014). Puebla 152 Genotype Dry bean genotype Puebla 152 is a mid - sized type III black bean belonging to the Middle American gene pool (Sing h et al. , 1991) that originat ed as a landrace selection from Puebla, Mexico. Several studies have found the dry bean genotype Puebla 152 to be sup erior in nodule development and subsequent N fixation ( St. Clair et al., 1988; Bliss et al., 1989; Park and Butter y , 1989; Pereira et al., 1989; Chaverra and Graham, 1992; Thomas et al., 1984; Wolyn et al., 1991 ; Tsai et al. , 1998 ). Puebla 152 was the don or parent of f ive high N fixing dry bean germplasm lines (WBR22 - 3, WBR22 - 8, WBR22 - 34, WBR22 - 50, and WBR22 - 55) released by Bliss et al. (1989) . ICA Pijao w as the recurrent parent. An estimated 44% of the shoot N dfa , compared to 35% N dfa for their high fi xing check, Rio Tibagi (Bliss et al., 1989). Chaverra and Graham (1992) studied early nodule formation and different inoculation rates on 40 dry bean genotypes, including Puebla 152. They found that Puebla 152 formed a high number of nod u le initials by day 8 after inoculation and that it was responsive to increase inoculation rates while other genotypes seemed to have reduced nodule formation at higher rates of inoculation. For plants harvested 51 days after planting, Puebla 152 had a high nodule dry we ight, high shoot weight, and accumulated a moderate level of N per plant. Puebla 152 w as also shown to have 14 superior S NF characteristics such as nodule number and dry weight with the application of varying rates of fertilizer N (Park and Buttery, 1989). Since agricultural soils have different levels of soil N it is beneficial to grow dry bean genotype s that are not inhibited or reduced in their ability to fix N when supplemental N is available. Puebla 152 has been useful in the study of many dry bean ch aracteristics. It has been used as a parent in several QTL studies including selection for sugar levels in snap bean pods ( Vanden l angenberg et al., 2012) and for root rot resistance in snap bean ( Navarro et al. , 2009; Ronquillo - Lopez et al., 2010 ). Metho ds to Measure S NF Several methods exist to measure the amount of nitrogen derived from the atmosphere ( %Ndfa ) . These includ e : the difference method ( Pereira et al., 1993 ; Muller et al., 1993 ), ureide levels in stem sap (Thomas et al., 1984 ; Hungria and Ne ves, 1987; Diatloff et al., 1991 ), acetylene reduction assay ( Hungria and Neves, 1987; Piha and Munns, 1987 ; Boddey et al., 1996 ), 15 N enrichment ( Hungria and Neves, 1987; St.Clair et al., 1988 ; Boddey et al., 1996 ) , 15 N depletion ( St. Clair et al., 1988; Pereira et al., 1989 ; Wolyn et al., 1991 ) , and 15 N natural abundance method ( Pereira et al., 1989 ). The simplest method to measure SNF is the difference method, which relies on calculating N fixed with the use of a non - nodulating reference plant . Pereira et al. (1993) used the non - The following equation is used to calculate %Ndfa : %Ndfa =Nfixer - Nno - nod/Nfixer Where Nfixer is the total N in the seed of the fixing crop and Nno - nod is the total N in the seed of the non - nodulating reference. It must be assumed that both the reference plant and fixing crop plant access similar layers of the soil. 15 Use of a fertilizer enriched, or depleted, in 15 N from the standard atmospheric content of 0. 368 atom %, meaning 0.368 % of all N atoms in the atmosphere are 15 N vs. 99.632 atom % 14 N , has been utilized to determine which fraction of the N in a fixing plant was derived from the atmosphere or the soil . Pereira et al. (1989) utilized ammonium sulfate (NH 4 ) 2 SO 4 containing 0.01 atom% 15 N applied throughout the growth cycle to both fixing dry bean lines and non - fixing soy bean lines. At different growth stages and seed 15 N atom% were then measured, Pereira et al. (1989) used the following equation to det ermine %Ndfa . %Ndfa = ((atom% 15 N(nfs) - atom% 15 N(fs)) X 100)/(0.368 atom% 15 N atom% 15 N(nfs) ) Where nfs=non fixing system (non - nodulating soybean) and fs=the fixing system (the dry bean studied). Using this method Pereira et al. (1989) determined that Puebla 152 obtained 31.6% of seed N dfa whereas the navy bean Sanilac obtained only 5.7% of its N from the atmosphere. These two genotypes were the high and low fixers, respectively , for this study. The 5.5 fold increase in %Ndfa for Puebla 152 translate d into an increase in seed yield (4,457 kg ha - 1 ) by a factor of 5.4 times the seed yield of Sanilac ( 818 kg ha - 1 ) . One drawback of the 15 N depleted or enric hed fertilizers is availability and cost. In addition, it is important that the fertilizer be ade quately mixed throughout the root profile and is equally available to all test plants. The non - fixing reference plant must be a similar root type as the species being evaluated, and p referably of the same species (Boddey et al., 1996) . Wheat has tradition ally been used as a reference crop in such studies since it is assumed to access the same soil profile as dry beans . H owever , Boddey et al. (1996) determined that was not the case when comparing the N acquisition from soil of the non 16 Rama e kers et al. (2013) utilized a modification of the 15 N depletion method to determine %Ndfa . Relying on the natural abundance of 15 N in the soil, a no n - fixing reference crop is used to determine the level of 15 N in the soil, wh ich would be high in a non - fixing plant and low in a plant fixing N from the air. They used the following equation to determine %Ndfa : %Ndfa =( 15 N non - fixing reference - 15 N fixing line)/( 15 N non - fixing reference - B) Where B is the 15 N of the fix ing line when it is relying completely on SNF. Rama e kers et al. (2013) averaged the 15 N of several fixing genotypes grown under greenhouse conditions in N free media and N free nutrient solution. They do not specify which lines were included to calcul ate the B value nor do they provide the B value . Evaluation of S NF levels earlier in the growth cycle prior t o harvest would be beneficial to more rapidly screen dry bean lines for SNF ability. St. Clair et al. (1988) discovered tha t there was little ag reement among the rank of the genotypes studied at R3 and R9, except Puebla 152 which fixed the most nitrogen at both stages. Thus, determining S NF levels at harvest would be advantageous compared to earlier time points. Lynch and White (1992) found that different organs are sinks for N at different developmental stages. Initially, N was partitioned in vegetative portions of the plant while later in the season pods and seeds were the destination of plant N, which was likely being relocated from vegetativ e tissues to reproductive tissues (seeds) (Lynch and White, 1992). Similarly, Boddey et al. (1996) and Wolyn et al. (1991) determined that early measurements of SNF levels were not necessarily related to SNF levels in seed at maturity. Differences in a 15 N can result in over or underestimation of %Ndfa . Lazali et al. (2014) looked at six RILs selected from a population 17 developed by crossing BAT477 and DOR364. The individual lines were selected for their tole rance or sensitivity to P deficiency. They found that there were differences across P levels and among the genotypes in their discrimination against 15 N (Lazali et al., 2014). Looking at different portions of the plant - roots, shoots, and nodules , they fo und that a higher proportion of 15 N remained in the roots , specifically the nodules, while the proportion of 15 N/ 14 N was lower in the shoots. Thus, measuring 15 N in the shoot might cause an overestimation of %Ndfa . Lazali et al. (2014) did find a signifi cant correlation between 15 N in nodules and P sensitivity and N fixed by the plant. In addition to the genotypes discriminating against 15 N, different strains of Rhizobia also discriminate differently for 15 N. Yoneyama et al. (1986) found that not only di d the 10 Rhizobium strains studied vary in their SNF ability, the amount of 15 N in the shoot of the three inoculated by different strains. In the laboratory, it is possible to control the specific strain of Rhizobium, however, in the field nodule occupancy is likely to vary even within the same plant. Yoneyama et al. (1984) noted that the amount of 15 N varied by the plant organ with stems and petioles having considerably le ss 15 N than leaves. Studying the kidney bean genotypes 15 N was +9.3 and +8.5 in the nodules, +1.5 and - 1.0 for pods, and - 0.6 and +2.8 for stems, respectively (Yoneyama et al., 1984). The selection of the plant part in calcu lations could cause a very different estimation of %Ndfa . Dry bean is a ureide transporting legume , meaning that N dfa is often translocated through the plant as alontoin and alontoic acid (Thomas et al., 1984) . Thus, S NF levels could be inferred from th e composition of the bleeding stem sap. During early developmental stages the predominant form of N is nitrate, which is derived from the soil. As plants mature and advance into pod filling phase ureides become a much larger portion of the N in plant sap (Thomas et al., 18 1984 ; Diatloff et al., 1991 ) . Surprisingly, Thomas et al. (1984) found that among genotypes Puebla 152 and Sanilac, and seven lines derived from an inbred backcross of Sanilac (the recurrent parent) and Puebla 152 , little variation in th e sap composition, especially early in development was observed . As pods of Puebla 152 began to fill the ure i de content in the N - treatment also increased relative to other forms of N, including nitrate. When sap flow rate is considered, however, the dif ferences become more dramatic with Puebla 152 clearly fixing more N than Sanilac and most of the inbred lines. Diatloff et al. (1991) found the same pattern with other navy bean genotypes. QTL Analysis and SNF Several studies have been co nducted to map QT L for various S NF traits ( Nodari et al., 1993; Tsai et al., 1998). In an attempt to look for QTL involved in the interaction between host and bacteria, Nodari et al. (1993) used an F 3 population from BAT93 (Mesoamerican derived genotype with fewer nodul es a nd resistance to common bean blight (CBB) ) and Jalo EEP558 (Andean selection with high nodule number, susceptib le to CBB.) Four QTL which explained a total of 52% of the phenotypic variation for nodule number were discovered. One locus appeared to have an effect on both nodule number and CBB resistance which is not surprising since many stages in the interaction with pathogenic bacteria are similar to interactions with beneficial bacteria . This region, on Pv07 contributed by the BAT93 parent , was a ssociated with CBB resistance but with low nodule number (Nodari et al., 1993). Tsai et al . (1998) used a similar population by crossing the high nodulating dry bean, Jalo EEP558 with the low nodulating BAT93 to investigate nodule number and CBB resistanc e inheritance under contrasting N conditions. Both parents contributed positive alleles to nodule number and CBB resistance in the F 2 derived F 3 RILs. Given that the low nodulating parent (BAT93) contributed 19 alleles with a positive effect on nodule number and the CBB susceptible parent similarly contributed posi tive alleles for CBB resistance. Rama e kers et al. (2013) us ed 85 RILs developed from G2333 x G19839 to investigate characteristics associated with SNF in both the greenhouse under N free conditions and in the field. They measured traits such as leaf chlorophyll content, shoot dry weight, total biomass N, seed yield and total N in seed. Many QTL were discovered for SNF traits, such as SPAD (a measure of chlorophyll, and hence N level in leaves) at different growth stages ( R 2 = 11.49 % to 35.53 % ) , %N in the shoot, root, and plant ( R 2 = 16.3 % to 21.01 % ) , and total N in the shoot, root, and plant ( R 2 = 14.69 % to 20.87 % ) , in the g reenhouse along with to nodule number QTL ( R 2 = 17.25 % and 16.72 % ) , two nodule dry weight QTL ( R 2 = 12.97 % and 19.07 % ) , and one %Ndfa at harvest QTL ( R 2 =18.79 % ) , in the field. They found different QTL between the field and greenhouse experiment but there were QTL that overlapped between both experiments such as a SPAD QTL on Pv01 an d two QTL on Pv07 for SPAD at pod filling in the greenhouse and the field. The QTL reported have low to moderate effect on the phenotype but could prove useful in developing markers for MAS . Summary D ry bean is an important crop and an important source of protein for low income people , worldwide . It is grown over many regions, including the northern tier and intermountain states in the U.S. In Michigan, dry bean is an important component of crop rotations especially in the main growing region, often kn , an area consisting of 4 major bean growing counties including Tuscola, Huron, Sanilac, and Bay. While a member of a plant family (Fabacea) known for SNF, dry bean is not the most efficient at fixing N especially when compared to relate d species such as soybean . The variability found within the species for improved N fixation , however, serves as important genetic material to improve SNF ability in 20 commercially acceptable dry bean lines. This improvement could help to reduce dependence o n N inputs which will also help to reduce production costs and damage to the environment caused by runoff of N from fields into ground water and adjacent waterways . Use of dry bean genotypes, such as Puebla 152, has been an important part of research cond ucted on SNF in dry bean with the focus on improving the SNF ability of dry bean. While not well adapted to commercial production at northern latitudes in the U.S. when used as a parent with a commercially adapted line s, Puebla 152 is a dependable source o f traits relating to SNF. Using genomic tools, such as SNP markers and QTL analysis, traits associated with improved SNF may be moved from the poorly adapted Puebla 152 to commercially acceptable dry bean lines. This information may also be useful in dev eloping genotypes with improved SNF characteristics in other market classes . The objective of this study was to integrate the superior SNF traits of Puebla 152 into genotype s better adapted to current productions methods in northern latitudes as well as t o elucidate genomic regions associated with SNF with the goal of identifying markers useful in selecting for enhanced SNF in future dry bean breeding programs. 21 LITERATURE CITED 22 LITERATURE CITED Abbaszadeh - dahaji, P., G .R. Savaghebi, H. Asadi - rahmani, F. Rejali, M. Farahbakhsh, B. Moteshareh - zadeh, M. Omidvari, and K . Lindstrom. 2012. Symbiotic effectiveness and plant g rowth p romoting traits in some Rhizobium strains i solated from Phaseolus vulgaris L. Plant G rowth Regulation. 68:361 - 370. Ahe mad, M. and M. Kibret. 2014. Mechanisms and a pplications of p lant g rowth p romoting Rhizobacteria: Current p erspective. Journal of King Saud University 26:1 - 20. Alkama, N., G. Ounane, and J.J. Drevon. 2012. Is genotypic variation of H + effl ux under P d eficiency linked with n odulated - r oot r espiration of N 2 - f ixing common bean ( Phaseolus vulgaris L.)? Journal of Plant Physiology. 169:1084 - 1085. Ballhorn, D.J., S. Kautz, and M. Schadler. 2013. Induced p lant defense via volatile production i s d ependent on r hizobial symbiosis . Oecologia. 172:833 - 846. Beebe, S., P.W. Skroch, J. Tohme, M.C. Duque, F. Pedraza, and J. Nienhius. 2000. Structure of g enetic diversity a mong common bean landraces of Middle American o rigin based on c orrespondence an alysis of RAPD. Crop Science. 40:264 - 173. Bellucci, E., E. Bitocchi, D. Rau, M. Rodriguez, E. Biagetti, A. Giardini, G. Attene, L. Nanni, and R. Papa. 2014. Genomics of origin , domestication and e volution of Phaseolus vulgaris . In Genomics of Plant Ge netic Resources edited by R. Tuberosa et al. DOI 10.1007/978 - 94 - 007 - 7572 - 5_20. Pp 483 - 507. Blair, M.W., L.M. Diaz, and J.A. Acosto - Gallegos. 2013. Race structure in the Mexican collection of common bean landraces. Crop Science 53(4):1517 - 1528. Bliss, F. A., P.A.A. Pereira, R.S. Araujo, R.A. Henson, K.A. Kmiecik, J.R. McFerson, M.G. Teixeira, and C.C. Da Silva. 1989. Registration of f ive h igh n itrogen f ixing common bean g ermplasm lines . Crop Science. 29:240 - 241. Bodd e y, R.M., S.H. Muller, and B.J.R. Al ves. 1996. Estimation of the contribution of biological N 2 fixation to two Phaseolus vulgaris g enotypes u sing simulation of p lant nitrogen u ptake from 15 N - labelled soil . Fertilizer Research. 45:169 - 185. Bourion, V., S.M.H. Rizvi, S. Fournier, H. de Lar ambergue, F. Galmiche, P. Marget, G. Duc, and J. Burstin. 2010. Genetic dissection of nitrogen n utrition in p ea t hrough a QTL approach of r oot, nodule , and s hoot variability . Theoretical Applied Genetics. 121:71 - 86. Broughton, W.J., G. Hernandez, M. Blair, S. Beebe, P. Gepts, and J. Vanderleyden. 2003. Beans ( Phaseolus spp ) -- Model food l egumes. Plant and Soil. 252:55 - 128. Chacon, MI, B. Pickersgil, and D.G. Debouck. 2005. Domestication p atterns in common bean (Phaseolus vulgaris L.) and the ori gin of the Mesoamerican and Andean c ultivated races . Theoretical Applied Genetics. 110:432 - 444. 23 Chaverra, M.H. and P.H. Graham. 1992. Cultivar v ariation in tr aits affecting e arly nodulation of c ommon b ean. Crop Science. 32:1432 - 1436. Diatloff, A., R.J. Redden, and D.F. Herridge. 1991. Correlation between xylem ureide levels and n odulation in f ield - grown Phaseolus vulgaris . Australian Journal of Experimental Agriculture. 31:679 - 682. Eckert, F.R., H.J. Kandel, B.L. Johnson, G.A. Rojas - Cifuentes, C . Deplazes, A.J. Vander Wal, and J.M. Osorno. 2011. Row s pacing and n itrogen effects on upright pinto bean cultivars under d irect harvest conditions . Agronomy Journal. 103(5):1314 - 1320. Elizondo Barron, J., R.J. Pasini, D.W. Davis, D.D. Struthman, an d P.H. Graham. 1999. Response to s election for seed y ield and nitrogen (N2) fixation in common bean (Phaseolus vulgaris L.). Field Crops Research. 62:119 - 128. Fageria, N.K., L.C. Melo, and J. de Oliveira. 2013. Nitrogen u se e fficiency in dry b ean g en otypes. Journal of Plant Nutrition. 36:2179 - 2190. Fageria, N.K., L.C. Melo, E.P.B Ferreira, J.P. Oliveira, and A.M. Knupp. 2014. Dry matter , g rain yield , and yield components of dry bean as influenced by nitrogen fertilization and r hizobia. Communica tions in Soil Science and Plant Analysis. 45:111 - 125. FAOSTAT, 2015a. Food and Agriculture Org anization of the United Nations . Found at http://faostat3.fao.org/browse/Q/QC/E accessed June 2015. FAOSTAT , 2015b. Food and Agriculture Org anization of the United Nations . Found at http://faostat.fao.org/ Accessed August 2015. Gepts, P. 1998. Origin and evolution of common bean : Past events and recent trends . HortSci ence. 33(7):1124 - 1130. Heilig, J.A. and J.D. Kelly. 2012. Performance of dry bean g enotypes grown under organic and conventional production s ystems in Michigan. Agronomy Journal. 104(5):1485 - 1492. Hergert , G.H. and J.A. Schild. 2013. Fertilizer man agement for dry edible beans . University of Nebraska Extension Service. ( http://ianrpubs.unl.edu/epublic/live/g1713/build/ ) Herrera - Cervera, J., J. Caballero - Mellado, G. Laguerre, H. Tichy, N. Requena, N. Amarger , E. Martinez - Romero, J. Olivares, and J. S anjuan. 1999. At least f ive rhizobial species nodulate Phaseolus vulgaris in a Spanish s oil. FEMS Microbiology Ecology. 30:87 - 97. Hill, E.C. 2014. Cover crop influence on nitrogen availability, weed dynamics, and dry bean (Phaseolus vulgaris L.) cha racteristics in an organic system. PhD Dissertation. East Lansing: Michigan State University. pp 8 - 25 Hine, J.C. and J.I. Sprent. 1988. Growth of Phaseolus vulgaris on v arious nitrogen s ources: The importance of urease . Journal of Experimental Bota ny. 39(208):1505 - 1512. 24 Hungria, M. and M.C.P. Neves. 1987. Partitioning of n itrogen from b iological fixation and f ertilizer in Phaseolus vulgaris . Physiology Plantarum. 69:55 - 63. Kim, D.H., S. Parupalli, S. Azam, S. Lee, and R.K. Varshney. 2013. Com parative sequence analysis of nitrogen f ixation - related genes in s ix legumes . Frontiers in Plant Science. 4:1 - 8. Lazali, M. A. Bargaz, G. Carlsson, S.M. Ounane, and J.J. Drevon. 2014. Discrimination against 15 N a mong recombinant inbred lines of Phase olus vulgaris L. contrasting in phosphorus use efficiency for n itrogen fixation . Journal of Plant Physiology. 171:199 - 204. Liu, Y. L. Wu, C.A. Watson, J.A. Baddeley, X. Pan and L. Zhang. 2013. Modeling biological dinitrogen fixation of field pea with a process - based simulation m odel. Agronomy Journal. 105(3):670 - 678. Lynch, J. and J.W. White. 1992. Shoot nitrogen d ynamics in tropical common bean . Crop Science. 32:392 - 397. Michiels, J., B. Dombrecht, N. Vermeiren, C. Xi, E. Luyten, and J. Vande rleyden. 1998. Phaseolus vulgaris is a non - s elective host for nodulation . FEMS Microbiology Ecology. 26:193 - 205. MSU, 2015. Michigan State University Extension. 2015. Found at http://www.msue.msu .edu/ Accessed June 2015. Mora, Y., R. Diaz, C. Vargas - Lagunas, H. Peralta, G. Guerrero, A. Aguilar, S. Encarnacion, L. Girard, and J. Mora. 2014. Nitrogen - fixing rhizobial strains isolated from common bean seeds : Phylogeny, physiology , and genome analysis . Applied an d Environmental Microbiology. 80(18):5644 - 5654. Muller, S., P.A.A. Pereira, and P. Martin. 1993. Effect of different levels of mineral nitrogen on nodulation and N 2 fixation of two cultivars of common bean ( Phaseolus vulgaris L.). Plant and Soil. 15 2:139 - 143. Nanjareddy, K., L. Blanco, M.K. Arthikala, X.A. Affantrange, F. Sanchez and M. Lara. 2014. Nitrate regulates rhizobial and mycorrhizal symbiosis in common b ean (Phaseolus vulgaris L.). Journal of Integrative Plant Biology. 56(3):281 - 298. Navarro, F.M., M.E. Sass, and J. Nienhuis. 2009. Marker - facilitated selection for major QTL associated with root r ot resistance in s nap bean ( Phaseolus vulgaris L.). Crop Science . 49:580 - 856. Neila, A., B. Adnane, F. Mustapha, B. Manel, H. Imen, L. Boul baba, G. Cherki, and S. Bouaziz. 2014. Phaseolus vulgaris - rhizobia s ymbiosis increases the phosphorus u ptake and s ymbiotic N 2 fixation u nder insoluble phosphorus . Journal of Plant Nutrition. 37(5):643 - 657. Nicolas, M.F., M. Hungria, and C.A.A. Arias. 2006. Identification of quantitative trait loci c ontrolling nodulation and shoot mass in progenies from two B razilian soybean cultivars . Field Crops Research. 95:355 - 366. 25 Nodari, R.O., S.M. Tsai, P. Guzman, R.L. Gilbertson, and P. Gepts. 1993. Towar d an integrated linkage map of common bean . III. Mapping genetic f actors controlling host - bacteria interactions . Genetics. 134:341 - 350. Park, S.J., and B.R. Buttery. 1989. Identification and characterization of common bean (Phaseolus vulgaris L.) line s well nodulated in the presence of high nitrate . Plant and Soil. 119: 237 - 244. Pacovsky, R.S., P. Da Silva, M.T. Carvalho, and S.M. Tsai. 1991. Growth and nutrient a llocation in Phaseolus vulgaris L. colonized with Endomycorrhizae or Rhizobium. Plant and Soil. 132:127 - 137. Pereira, P.A.A., R.H. Burris and F.A. Bliss. 1989. 15 N - Determined d initrogen f ixation potential of genetically diverse bean lines ( Phaseolus vulgaris L.). Plant and Soil. 120:171 - 179. Pereira, P.A.A., B.D. Miranda, J.R. Atte well, K.A. Kmiecik, and F.A. Bliss. 1993. Selection for increased nodule number in common bean ( Phaseolus vulgaris L.). Plant and Soil. 148:203 - 209. P iha, M.I. and D.N. M unns. 1987. Nitrogen fixation p otential of b eans ( Phaseolus vulgaris L.) compare d with other grain legumes under controlled c onditions . Plant and Soil. 98:169 - 82. Podlesny, J. J. Wielbo, A. Podlesna, and D. Kidaj. 2014. The pleiotropic effects of extract c ontaining nod factors on p ea g rowth and yield . Central European Journal o f Biology. 9(4): 396 - 409. Rama e kers, L., C.H. Galeano, N. Garzon, J. Vanderleyden, and M.W. Blair. 2013. Identification of quantitative trait loci for symbiotic nitrogen fixation capacity and r elated traits in c ommon bean . Molecular Breeding. 31:163 - 18 0. Ribeiro, R.A., E. Ormeno - - Romero, M. Hungria. 2013. Novel Rhizobium i solated from r oot nodules of the common bean ( Phaseolus vulgaris L.) in Andean and Mesoamerican areas. Research in Micro biology. 164:740 - 748. Rodino, A.P., M. De La Fuente, A.M. De Ron, M.J. Lema, J.J. Drevon, and M. Santalla. 2011. Variation for n odulation and plant yield of common bean genotypes and environmental effects on the genotype expression . Plant and Soil. 34 6:349 - 361. Ronquillo - Lopez, M.G., C.R. Grau, and J. Nienhuis. 2010. Variation in r eaction to f usarium spp. identified in a common bean (Phaseolus vulgaris L.) population developed for field - b ased r esistance to root rot and wilt . Crop Science. 50:2303 - 2309. Schmutz, J., P.E. McClean, S. Mamidi, G.A. Wu, S.B. Cannon, J. Grimwood, J. Jenkins, S. Shu, Q. Song, C. Chavarro, M. Torres - Torres, V. Geffroy, S.M. Moghaddam, D. Gao, B. Abernathy, K. Barry, M. Blair, M.A. Brick, B. Chovatia, P. Gepts, D.M. Good stein, M. Gonzales, U. Hellsten, D.L. Hyten, G. Jia, J.D. Kelly, D. Kudrna, R. Lee, M.M.S. Richard, P.N. Miklas, J.M. Osorno, J. Rodrigues, V. Thareau, C.A. Urrea, M. Wang, Y. Yu, M. Zhang, R.A. Wing, P.B. 26 Cregan, D.S. Rokhsar, and S.A. Jackson. 2014. A r eference genome for common bean and g enome - w ide analysis of dual domestications . Nature Genetics. 46(7):707 - 713. Schubert, K.R. 1982. Products of biological nitrogen fixation in higher plants : Synthesis, transport , and metabolism . Annual Reviews in P lant Physiology. 37:539 - 574. Sevey, G.C. 1918. Bean culture . Orange Judd Company, New York, New Yor k . Singh, S.P., P. Gepts, and D.G. Debouck. 1991. Races of common bean ( Phaseolus vulgaris , Fabaceae). Economic Botany 45(3):379 - 396. St. Clair, D. A., D.J. Wolyn, J. DuBois, R.H. Burris, and F.A. Bliss. 1988. Field comparison of d initrogen fixation determined with nitrogen - 15 - depleted and n itrogen - 15 - enrichded ammonium sulfate in s elected inbred b ackcross lines of c ommon bean . Crop Science. 28:773 - 778. St. Clair, D.A. and F.A. Bliss. 1991. Intrapopulation d ecombination for 15 N - determined dinitrogen fixation a bility in common bean . Plant Breeding. 106:215 - 225. Thomas, R.J., J.R. McFerson, L.E. Schrader, and F.A. Bliss. 1984. Composition of ble eding sap nitrogen from lines of field - grown Phaseolus vulgaris L. Plant and Soil. 79:77 - 88. Trabelsi, D. A. Mengoni, H.B. Ammar, and R. Mhadi. 2011. Effect of on - field inoculation of Phaseolus vulgaris with rhizobia on soil bacterial communities . FEMS Microbial Ecology. 77:211 - 222. Tsai, S.M. R.O. Nodari, D.H. Moon, L.E.A. Camargo, R. Vencovsky, and P.Gepts. 1998. QTL mapping for nodule number and common bacterial blight in Phaseolus vulgaris L. Plant and Soil. 204:135 - 145. USDA - ERS, 2015. Unite d States Department of Agriculture Economic Research Service found at http://www.ers.usda.gov/ Accessed June 2015. USDA - NASS, 2015. United States Department of Agriculture National Agriculture Statistics Service Qui ck Stats found at http://quickstats.nass.usda.gov/ Accessed July 2015. Vandenlangenberg, K.M., P.C. Bethke, and J. Nienhuis. 2012. Identification of quantitative t rait l oci associated with fructose , glucose , and sucrose concentration in snap bean p ods. Crop Sciences. 52:1593 - 1599. Walley, F.L., G.W. Clayton, P.R. Miller, P.M. Carr, and G.P. Lafond. 2007. Nitrogen economy of pulse crop production in the Northern Great Plains. Agronomy Journal. 99:1710 - 1718. Wolyn, D.J., J. Attewell, P.W. Ludden, and F.A. Bliss. 1989. Indirect measure of N 2 fixation in c ommon bean ( Phaseolus vulgaris L.) under field conditions : The role of lateral root nodules . Plant and Soil. 113:181 - 187. Wolyn, D.J. , D.A. St Cl air, J. DuBois, J.C. Rosas, R.H. Burris, and F.A. Bliss. 1991. Distribution of n itrogen in common bean (Phaseolus vulgaris L.) genotypes selected for d ifferences in nitrogen f ixation ability . Plant and Soil. 138:303 - 311. 27 Yadegari, M. H.Asadie Rahmani, G . Noormohamadi, and A. Ayneband. 2010. Plant growth p romoting rhizobacteria increase growth , yield , and nitrogen f ixation in Phaseolus vulgaris . Journal of Plant Nutrition. 33(12):1733 - 1743. Yoneyama, T., N. Yamada, H. Kojima, and J. Yazaki. 1984. Variations in 15 N abundances in leguminous plants and nodule fractions. Plant and Cell Physiology. 25(8): 1561 - 1565. Yoneyama, T., K. Fujita, T. Yoshida, T. Matsumoto, I. Kambayashi, and J. Yazaki. 1986. Variation in natural abundances of 15 N among plan t parts and 15 N/ 14 N during N 2 fixation in the legume - rhizobia symbiotic system. Plant Cell Physiology. 27(5): 791 - 799. 28 CHAPTER 2 EVALUATION OF SYMBIOTIC N FIXATION OF BLACK AND NAVY BEAN ADVANCED BREEDING LINES UNDER ORGANIC PRODUCTION SYSTEMS Abstra ct Michigan has been a leader in organic dry bean production. Organic production involves the use of certified inputs which are derived from natural sources. Fertility is managed through crop rotation, cover crops, and addition of composts and manures in stead of the application of synthetically produced fertilizers. Previous research has found that dry bean yields were substantially lower under organic conditions compared to adjacent conventional production. Since pests are controlled with approved meth ods in each respective system, fertility appears to be an issue where the two systems may differ. Seventy - nine black and navy bean elite breeding lines , commercial checks, and a non - nodulating check were evaluated for yield under organic conditions in Fra nkenmuth, Caro, and Wisner, MI in 2011 through 2013. These same genotypes were also assayed for nodulation characteristics, N fixation, and shoot and root growth in the greenhouse under N free conditions. Several traits measured in the greenhouse were si gnificantly correlated to traits measured in the field. In particular, percent N derived from the atmosphere (%Ndfa) in the greenhouse was correlated with seed yield, N yield, and %Ndfa in the field in most site years. Measuring N in seed or plant tissue may not be necessary to estimate symbiotic nitrogen fixation (SNF) characteristics when plants are grown under limited N fertility yield in the field. 29 Intro duction Michigan is one of 18 states with major production of dry beans ( Phaseolus vulgaris L.) and ranks 2 nd in total dry bean production, after North Dakota (USDA - ERS, 2014) . The two major market classes in Michigan are black beans and navy beans with Mi chigan being the leading producer of black beans in the country and the second leading producer of navy beans, after North Dakota ( USDA - ERS, 2014 ). In the United States, black beans rank behind the leading market classes of pintos and navy beans in overal l production (USDA ERS, 2014). Michigan is the leading producer of organic dry beans in the U.S. (Table 2.01). According to the 2008 report of the National Agricultural Statistics Service with 1,960 ha harvested at a value of $3.9 million. In that year, Michigan accounted for 38.5% of the acres of organic dry beans harvested (USDA ERS , 2008). Organic dry bean production grew 81% by 2011 with Michigan accounting for 48% of U.S. production (Table 2.01). According to the USDA National Organic Program (NOP) in an approved manner consistent with the standards set by the program. Inputs are limited to those approved by the Organic Materials Review Institute (OMRI). Generally, sy nthetic certified organic production. Inputs, such as manure, composts and other approved soil additives the production of certified organic agricultural products. A producer must work through a certifying agency which reviews records and certifies that acceptable materials are used and practices are followed. In comparative studies of dry beans grown und er both organic and conventional production systems, Heilig and Kelly (2012) found that on average, yields in organic production were 20% lower compared with those raised under conventional systems in adjacent trials. Weeds are often 30 a problem in organic production systems as there are few approved controls aside from delayed planting and multiple mechanical cultivations. Weeds were controlled with mechanical methods including hand removal. Insects were similarly controlled with approved insecticides. As ide from potential varying efficacies of organic control methods, soil fertility was identified as a possible source of the differences in yield (Heilig and Kelly, 2012). Conventionally produced beans were provided artificial fertilizer at the recommended rate of 55 kg N ha - 1 (Warncke et al., 2009) which is readily available while fertility in the organic system relies on the natural processes involved in breaking down organic components to release nutrients potentially leading to reduced availability, in particular nitrogen. Dry bean genotypes that performed well in conventional production systems were also the best suited to organic production due to resistance to disease as well as adaptation to modern agricultural practices (Heilig and Kelly, 2012). G enotypes developed under conventional management, however, are not selected for their ability to fix N from the atmosphere as artificial fertilizer is applied to all conventional trial plots. Comparing 4 dry bean cultivars, Oliveira et al. (1998) found th at beans without supplemental N were able to fix as much N as plants fertilized with 60 kg N ha - 1 accumulated 198.8 mg N plant - 1 through fixation compared to 184.7 mg N plant - 1 with the addition of 60 kg N ha - 1 . Cover crops ar e often used in organic production to manage fertility and control weeds. Different cover crops contribute different amounts of N to total inorganic N in soil (Hill, 2014 ). Legume cover crops such as medium red clover ( Trifolium pratense L.) contributed l ess than 20 kg ha - 1 to over 100 kg ha - 1 through the growing season following incorporation, depending on location and year (Hill, 2014 ). Without a cover crop, total inorganic N available in the soil ranged from less than 20 kg ha - 1 to 80 kg ha - 1 (Hill, 20 14 ). Other cover crops studied, cereal rye 31 ( Secale cereale L.), and oilseed radish ( Raphanus sativus L.) soil N rates were intermediate (Hill, 2014 ). The higher levels of soil N observed are above recommendations (Warncke et al. , 2009), however, the amoun t of N in the soil is extremely variable depending on location and cover crop used. Additionally, total soil N measurements do not represent the N available to the plant at any given time. In those situations where soil N is too low to produce a competit ive dry bean crop, SNF should provide the balance. Improving the SNF ability of future dry bean cultivars should be a useful benefit for organic producers and commercial producers who want to use reduce fertilizer inputs. The current study was designed t o investigate the importance of symbiotic N fixation (SNF) ability in performance of dry bean genotypes grown under organic production systems and compare performance with traits associated with SNF measured under N - free conditions in the greenhouse. Mat erials and Methods Plant material Seventy - nine black and navy bean genotypes including elite breeding lines, commercial checks and one non - nodulating genotype were grown under organic conditions over three growing seasons (2011, 2012, and 2013) on certifi ed organic ground in Caro, Michigan, Wisner, Michigan and Frankenmuth, Michigan. Each season 18 black bean and 18 navy bean genotypes were planted but entries differed annually based on prior performance. As the study progressed, some lines were dropped f rom the study while newer breeding lines were added to replace them. Black and navy bean lines included cultivars and elite breeding lines from the bean breeding program at MSU along with selections from the a black bean recombinant inbred line population 32 enhanced SNF ability (St. Clair et al., 1988; Bliss et al., 19 89; Park and Buttery, 1989; Pereira et al., 1989; Chaverra and Graham, 1992; Thomas et al., 1984; Wolyn et al., 1991; Tsai et al. 1998). Puebla 152 was not included in the test due to a lack of adaptation to local growing conditions. The non - to calculate nitrogen fixation. Inoculation Prior to planting seed was treated with rhizobial inoculant consisting of fine buffered peat (American Peat, Technology, Aitkin, MN , USA ) carrying Rhizobium tropici strain CIAT899 which was prepared by culturing the rhizobia in yeast mannitol broth (as described in Somasegaran and Hoben, 1994) for three d prior to mixing with peat at a rate to allow sufficient wetting of the peat. Selecti on of Rhizobium tropici strain CIAT899 as the inoculant was based on its widespread use in research and ability to form a symbiosis with a wide range of dry bean genotypes (Graham et al., 2003). The resulting inoculant was incubated in the dark for 8 to 1 2 wk at room temperature. Seed was mixed with the peat inoculant and a small amount of water to adhere the peat inoculant to the seed. Treated seed was stored in a cooler until planting. Planting and C ultivation Seed was planted into two - row plots 6.1 m long with rows spaced 51 cm apart at a rate of approximately 14 seed per m in a lattice design with 4 repetitions. Two locations were planted each season at the Saginaw Valley Research and Extension Center (SVREC), Frankenmuth, MI and Wisner, MI in 201 1; Wisner, MI, and Caro MI in both 2012 and 2013. The Wisner location 33 was abandoned in 2013 due to excessive moisture throughout the season and at harvest. Field conditions are reported in Table 2.02. Weed control was by mechanical cultivation as needed early in the season and supplemented with hand weeding to control weeds especially within the plant row where the cultivator was unable to reach. Plots were cultivated 2 to 3 times each season, as needed. No fertilizer was applied at planting nor during the growing season. Potato leaf hopper was controlled as needed with Pyganic Crop Protection EC 5.0 ( McLaughlin Gormley King Company, Minneapolis, MN) at a rate of 586 ml ha - 1 resulting in 30 ml ha - 1 active ingredient. Data C ollection Days to flower was recorded as the number of days after planting when 50% of the plants in each plot had one open flower. Maturity was determined when 50% or more of the plants had reached physiological maturity, at which time plant height (cm), lodging (1=upright, 5=prost rate) and agronomic desirability (1=not desirable, 6=highly desirable) were recorded. Harvest When plants reached maturity, plots were direct harvested with a Wintersteiger AG plot combine (Winterstieger AG, Austria). Seed was air dried and cleaned with a Clipper Mill (A.T. Ferrell Company, Bluffton, IN, USA) before weighing. Seed moisture content at weighing was measured with a Dickey - john GAC 2500 moisture meter (Churchill Industries, Minneapolis, MN). Yield was calculated by adjusting values to 18% m oisture content. Seed size was determined by weighing a random sample of 100 seeds, adjusted to 18% moisture. 34 Nitrogen Analysis A 30 g seed subsample from each plot was placed in an envelope and placed into a dryer at 60º C for 1 wk. Seed was then grou nd in a Wiley Mill pass through a 1 mm mesh screen. Seed samples were then stored at room temperature until sent to the Stable Isotope Facility at UC PDZ Europa ANCA - GSL elemental analyzer (Sercon Ltd., Cheshire, UK ) Percent nitrogen derived from the atmosphere (%Ndfa) using the difference method by the following equation (Boddey, 1987): %Ndfa = (N yield - Fixer N yield - non - Fixer)/N yield - Fixer Where N yield = Seed Yield (kg) * %N of the respe ctive fixer and non - fixer (R99) . Greenhouse Assay To study the SNF ability of individual genotypes at flowering, the navy and black bean genotypes along with commercial checks were grown in the absence of N under greenhouse conditions. Seeds of the genoty pes being studied were sterilized by soaking in a 10% bleach solution for 2 min followed by two 2 - min rinses with sterile water. Six seeds were planted into each plastic 5.7 l nursery container which had been filled with a 2:1 mix, v:v of perlite to vermi culite which had been autoclaved. Seeds were watered in with tap water. At 3 d after germination, 500 ml YMB culture of R . tropic i CIAT899 was diluted in 20 l of tap water which had been adjusted to a pH of approximately 6.5 which resulted in a concentra tion of approximately 10 3 cells per ml. Each nursery container received 250 ml of the final rhizobial dilution. Inoculation was repeated 10 d after planting in the same manner to ensure sufficient population levels of rhizobia to effect symbiosis. Pots w ere placed randomly on a greenhouse bench with day length extended to 16 h with high pressure sodium lights. Two to three times 35 weekly each pot was watered with 500 ml full strength Broughton and Dilworth N free solution (Broughton and Dilworth, 1970) as needed to avoid drought stress. Plants were thinned to two plants per pot before emergence of the first trifoliate. The experiment was repeated three times. When one or both of the plants had at least one flower open, plant shoots were measured and cut with a razor blade at soil level. The perlite and vermiculite was carefully removed from the roots which were measured for maximum length and scored from nodulation on a scale of 0 to 6, with 0 representing roots without nodules and 6 being roots with a l arge number of fully developed and functioning nodules. Both root and shoot biomass samples were dried in a dyer at 60º C for 7 d at which time they were weight and then ground to pass through a 1mm screen on a Wiley mill. Samples were also sent the U.C. Davis SIF for N analysis as described previously for field grown seed. Total biomass was calculated by adding root biomass and shoot biomass. Similar to the calculation for %Ndfa, total biomass difference, shoot biomass difference, and root biomass diffe rence were calculated by the following equation for the respective trait: Difference = (mass (g) of fixer - mass (g) of non - fixer)/mass (g) of fixer Statistical Analysis A PROC GLM analysis using SAS 9.4 (SAS Institute Inc. 100 SAS Campus Drive, Cary, NC 27 513 - 2414, USA) to generate an ANOVA determined that there was a significant difference among years and sites. Consequently, each site/year was analyzed separately. PROC CORR was used to generate Pearson Correlations in SAS 9.4. Results and Discussion V ariability in precipitation and soil conditions at each site along with the rotation of genotypes as the study progressed ma d e it difficult to make comparisons of yield performance of particular 36 genotypes. Field stations were relied on for precipitation m easurements, however, variable precipitation patterns with localized rain events caused a discrepancy between what was measured at the field station and what was observed on the actual plot. Checks planted in all site/years showed variability in yield acro ss years and locations . In 2011 Zorro was the highest yielding line at both sites (2117 kg ha - 1 and 1519 kg ha - 1 ) while breeding lines yielded more in 2012 and 2013 (Table 2.03). In 2012 the elite breeding line B11361 was the highest yielding line at the Wisner location while B11302 was the highest yielding line at the Caro location. The Caro location in 2013. Zenith was evaluated as breeding line B10244 in t his study prior to registration. Overall the navy bean genotypes were not as competitive as the black bean Prior research on early nodulation found that while navy and black beans formed nodules in the same amount time, navy beans produced fewer nodules early in development than black beans (Heilig and Kelly, 2012 b ). Hungria and Philips (1993) found that dry beans with white seeds produced fewer flavonoids tha n genotypes with colored seed. These flavonoids are involved in early initiation of symbiosis and reducing the concentration may reduce the number of infection sites and thus nodule number. The non - nodulating genotype R99 was the lowest yielding line on ly in 2012 at both locations, possibly reflecting the nitrogen level in the fields (Table 2.03). R99 was at or below average for percent seed N in all sites and years. Low to moderate coefficients of variation for percent seed N suggest that this trait w as not as affected by environmental variation as yield (Table 2.04) . Nitrogen yield reflects the amount of N harvested and is calculated by multiplying seed yield by percent seed N. R99 had the lowest N yield in all years and sites (Table 2.03). Nitrog en yield 37 largely followed the pattern of seed yield except in 2012 at the Wisner site when B09175 had the highest N yield, whereas Zorro had the highest seed yield. %Ndfa was determined for all entries using R99 as a non - fixing reference check (Table 2.04) . Overall, the highest %Ndfa was 71.7% for Zorro at the Frankenmuth site in 2011 while the lowest was 9.8% for the breeding line B10243 at the Wisner site in 2012 (Table 2.04). The relationship among yield over all site years appears to be rather limited as Pearson correlations show that yield in any particular site/year is not correlated to yield in any other site year (data not shown). The correlation was moderate between the Wisner site and Frankenmuth site in 2011 . The same pattern was found for N yi eld, which was expected as seed yield was used to calculate N yield. Correlations between %Ndfa and %N in seed were much more consistent across locations and years (Table 2.05). There was a high correlation for %Ndfa between Caro site 2013 (r=0.96, p=0.00 19) and Wisner and Frankenmuth in 2011 ( r=0.94 , p=0.0046). Similarly correlations were fairly high between Wisner 2012 and both sites in 2011 (r=0.88, p= 0 .0009 and r=0.82, p=0.0034) and Caro 2012 and both sites in 2011 (r=0.77, p=0.0009, and r=0.73, p=0.0 16). %Ndfa was moderately or strongly correlated with %N for most site/years, suggesting that there is an underlying genetic factor that controls this characteristic. Environmental variability had a higher impact on yield than it did on %Ndfa. There was no statistical difference among the site/years for percent seed N suggesting that this trait too is less dependent on environmental factors but genetically determined. The values for traits measured in the greenhouse under N free conditions are shown in Table 2.0 6 . As expected, the vast majority of N was derived from the atmosphere, ranging from 75.8 to 98.8 %Ndfa (Table 2.0 6 ). The only potential sources for N would be the N in the planted seed, N in tap water, or dust that fell on the surface of the me dia. Nodule ratings differed 38 (2.0) (Table 2.0 6 ). Many of the tr aits measured in the greenhouse were significantly correlated with each other (Table 2.07). Root N and shoot N are very highly correlated (r=0.95, p < 0.0001) which is not surprising as N status of the plant depends on the roots acquiring N whether it origi nates in the potting media or the rhizobia in the nodules. Root biomass and shoot biomass are similarly highly correlated (r=0.93, p < 0.001). Since N is a major nutrient which determines plant growth these parameters would be expected to be closely relate d. The visual score (0 to 6) of root nodules was moderately to highly correlated to all measured traits. A visual scoring of nodules is preferable to actual counts due to time restraints and cost. The root:shoot biomass ratio was inversely correlated to all other greenhouse traits (Table 2.07). As the shoot biomass increases, other traits, such as shoot N (r= - 0.3, p < 0.05) and %Ndfa (r= - 0.69, p < 0.0001) decrease. While N fixation is important for growth of dry bean, partitioning of N to the seed appears to be equally important. Biomass, and the proportion of biomass in the roots versus the shoot seems to play a role in yield and SNF traits. T he root:shoot ratio in the greenhouse showed a significant moderate negatively correlation with seed yield and N y ield in the field for three of five site years (Table 2.08). As the amount of root biomass increase d , seed yield and N yield decrease d . This partitioning among various organs may account for the differences seen in the field with respect to seed yield an d N yield and likely are not directly involved in the SNF process but more involved in the movement of N, whether fixed or from the soil, within the plant. Developing a screening method for traits related to SNF that requires less time and space than a fu ll field evaluation would be advantageous for breeders selecting genotypes with enhanced SNF. Assays conducted in the greenhouse require less time to complete, can be conducted 39 during the off - season, and can be controlled to a much greater degree than fie ld studies. Correlations between the traits measured in the greenhouse and those in the field can help to determine which traits are useful for making selections for enhanced SNF. One trait that was fairly consistently correlated to the field traits meas ured at all site/years was %Ndfa (Table 2.05). There was a significant moderate to high correlation between greenhouse %Ndfa and seed yield in three out of five site/years. The field %Ndfa was significantly correlated with greenhouse %Ndfa for all site/ye ars whereas N yield in the field was moderately correlated with greenhouse %Ndfa in three of five site/years. The %N in either the seed from the field trials or %N in shoot biomass is not significantly associated with yield traits. It is likely that %N in plant tissue is genetically independent of the processes involved in acquiring N. Since there is no storage form of N in the plant, N is found in the structural components of the biomass, with biomass growth being dependent on availability of N. A sui which is less costly and easier to measure than measuring actual N content. Shoot biomass difference was significantly correlated with seed yield in four of five site years ; moderately to highly correlated with field %Ndfa in five out of five site years; and moderately correlated to N yield in four of five site years. The only site year where the shoot biomass difference was not correlated with seed yield or N yield was the Wisner site in 2012. Drought stress was severe in 2012 and may be confounding the results for this site year. Other greenhouse traits showing promise for predicting field performance include nodule rating, shoot biomass, and root:shoot ratio. Given tha t SNF traits in the greenhouse correlate with yield parameters in the field, N may be a limiting factor in organic dry bean production and should be considered in crop rotation choices. In the field studies %Ndfa ranged from 23.5% to 71.2% for commercial g enotypes 40 whereas %Ndfa for advanced breeding lines was more variable ranging from 11.1% to 71.7% (Table 2.0 4 ). Genotypes such as Zorro and Zenith had relatively high yield along with higher levels of %Ndfa suggesting that there may be little advantage to e nhanced SNF in these genotypes. The range in %Ndfa seen in the advanced breeding lines indicates that there is considerable potential to select lines with enhanced SNF ability and retain yield potential. Conclusions Variability in performance of genotypes as well as variable environmental conditions make it necessary to evaluate dry bean genotypes over several seasons and multiple locations to better determine their yield and SNF potential. Not all dry bean genotypes have potential to fix a high %Ndfa, bu t variation in this trait is present in elite breeding lines that exceed fixation levels seen in commercial genotypes. These genotypes may prove to be worthy of release, such as Zenith, which surpassed the previously released cultivar Zorro in yield and % Ndfa in 2 of 3 years tested. 41 APPENDICES 42 APPENDIX A CHAPTER 2 TABLES Table 2.01. Hectares planted and production of organically produced dry bean in the United States . Data from USDA - NASS, 2015. http://quickstats.nass.usda.gov/ accessed June 2015. 43 Table 2.02. Field location, year, planting date, precipitation, soil type and soil chemistry of fields where black and navy bean genotypes were evaluated under organic conditions i n 2011, 2012, and 2013 in Michigan. Report generated at Enviro - Weather (http://www.agweather.geo.msu.edu/mawn/) for the months of June through September The 30 year average for 1 June through 30 September is 362 mm § Nearest weather station used for p recipitation data USDA Web Soil Survey Natural Resource Conservation Service Total Kjeldahl Nitrogen measured by the Soil and Plant Nutrient Lab at Michigan State University 44 Table 2.03. Seed yield and N yield of 79 black and navy bean genotypes grown under organic co nditions in Frankenmuth, Caro, and Wisner, MI in 2011, 2012, and 2013 . Saginaw Valley Research and Extension Center, Frankenmuth, MI. Highest and lowest yield in each site and year of advanced breeding lines. May not be the same line each year. 45 Table 2.04. Percent N derived from the atmosphere (%Ndfa) and percent N in seed of 79 black and navy bean genotypes grown und er organic conditions in Frankenmuth, Caro, and Wisner, MI in 2011, 2012, and 2013 . Saginaw Valley Research and Extension C enter, Frankenmuth, MI. Highest and lowest value in each site and year of advanced breeding lines. May not be the same line each year. 46 Table 2.05. Pearson correlations for %Ndfa (lower left) and %N (upper right) for 79 black and navy bean genotype s grown under organic conditions in Frankenmuth, Caro, and Wisner, MI in 2011, 2012, and 2013 . Correlation significant at 0.0001 . Saginaw Valley Research and Extension Center, Frankenmuth, MI. 47 Table 2.06. Traits measured on 79 black and navy dry bean genotypes grown under N free conditions in the greenhouse in East Lansing MI in 2011 an d 2012 . Ratio of shoot N (g) to root N (g). Rating of nodules on roots 6.0 being heavily nodulated and 0.0 having no nodules. § %Ndfa calculated on biomass using the difference method. Biomass difference calculated using the biomass (g) of each re spective plant part with the non - nodulating R99 used as a reference. Highest and lowest value in each trait of advanced breeding lines. May not be the same line for each trait. 48 Table 2.07 . Pearson Correlations of traits measured of 79 black and navy bean genotypes grown under N free conditions in the greenhouse in East Lansing, MI in 2011 and 2012 . Correlation significant at 0.0001 Diff = (biomass of N fixer (g) - biomass non - fixer (g))/biomass of N fixer (g) § Root/Shoot=root biomass (g)/shoot biomass (g) 49 T able 2.08. Pearson correlations for field traits and greenhouse traits measured for 79 black and navy bean genotypes grown i n the field under organic production systems and in the greenhouse in East Lans ing, MI under N free conditions in 2011 and 2012 . Correlation significant at 0.0001 . Calculated using the difference method. § Saginaw Valley Research and Extension Center, Frankenmuth, MI. Biomass difference calculated using the biomass (g) of each respective plant part with the non - nodulati ng R99 used as a reference. 50 APPENDIX B CHAPTER 2 SUPPLEMENTAL TABLES Table S 2 . 0 1 . Seed yield and N yield of 79 black and navy bean genotypes grown under organic conditions in 2011, 2012, and 2013 in Frankenmuth, Caro, and Wisner, MI. 51 Table S 2.0 1 ( ) 52 Table S 2.0 1 ( ) 53 Table S 2.0 1 ( ) Saginaw Valley Research and Extension Center, Frankenmuth, MI. Advanced breeding lines were rotated through the trial based on performanc e . § Selected genotypes from the Puebla 152/Zorro recombinant inbred line (RIL) population (See Chapter 2). 54 Table S 2. 02 . Percent N derived from the atmosphere (%Ndfa) and percent N in seed of 79 black and navy bean genotypes grown under orga nic conditions in 2011, 2012, and 2013 in Frankenmuth, Caro, and Wisner, MI. 55 Table S 2. 02 ( ) 56 Table S 2. 02 ( ) 57 Table S 2. 02 ( ) Saginaw Valley Research and Extension Center, Frankenmuth, MI. Advanced breeding lines were rotated through the trial based on performance . § Selected genotypes from the Puebla 152/Zorro recombinant inbred line (RIL) population (See Chapter 2). 58 Table S 2. 03 . Traits measured on 79 black and navy d ry bean genotypes grown under N free conditions in the greenhouse in East Lansing MI. 59 Table S 2. 03 ( ) 60 Table S 2. 03 ( ) 61 Table S 2. 03 ( ) Ratio of shoot N (g) to root N (g). Rating of nod ules on roots 6.0 being heavily nodulated and 0.0 having no nodules. § %Ndfa calculated on biomass using the difference method. Biomass difference calculated using the biomass (g) of each respective plant part with the non - nodulating R99 used as a refere nce. 62 LITERATURE CITED 63 LITERATURE CITED Bliss, F.A., P.A.A. Pereira, R.S. Araujo, R.A. Henson, K.A. Kmiecik, J.R. McFerson, M.G. Teixeira, and C.C. Da Silva. 1989. Registration of five high nitrogen fixing common bean germplasm lines. Crop Science. 29:240 - 241. Boddey R M . 1987 . Methods for quantification of nitrogen fixation associated with G ra mineae. Critical Reviews in Plant Science. 6 : 209 - 266. Broughton, W.J., Dilworth, M.J., 1970. Methods in legume - rhizobium technology: plant nutrient solutions. In: Somasegaran, P., Hoben, H.J. (Eds.), Handbook for Rhizobia, Methods in Legume - Rhizobium Technology. Springer - Verlag, New York, Inc. Page 340. Chaverra, M.H. and P.H. Graham. 1992. Cultivar variation in traits affecting early no dulation of common bean. Crop Science. 32:1432 - 1436. Graham, P.H., J.C. Rosas, C. Estevez de Jensen, E. Peralta, B. Tlusty, J. Acostos - Gallegos, and P.A. Arraes Pereira. 2003. Addressing edaphic constraints to bean production: The bean/cowpea CRSP P roject in perspective. Field Crops Research. 82: 179 - 192. Heilig, J.A. and J.D. Kelly. 2012a. Performance of dry bean genotypes grown under organic and conventional production systems in Michigan. Agronomy Journal. 104(5):1485 - 1492. Heilig, J.A. and J .D. Kelly. 2012b. Utilizing growth pouches to screen black and navy dry bean breeding lines for early nodulation. Annual report of the Bean Improvement Cooperative. 55:67 - 68. Hill, E.C. 201 4 . Cover crop influence on nitrogen availability, weed dynamic s, and dry bean (Phaseolus vulgaris L.) characteristics in an organic system. PhD Dissertation. East Lansing: Michigan State University. 112 p. Hungria, M., and D. A. Phillips. 1993. Effects of a Seed Color Mutation on Rhizobial nod - Gene - Inducing Flav anoids and Nodulation in Common Bean. Molecular Plant - Microbe Interactions. 6(4):418 - 422. bean. Journal of Plant Registrations. 9(1):15 - 20. Kelly, J.D., G.V. Va Journal of Plant Registrations. 3:226 - 230. Oliveira, W.S., L.W. Meinhardt, A. Sessitsch, and S.M. Tsai. 198. Analysis of Phaseolus - Rhizobium interactions in a subsistence farming s ystem. Plant and Soil. 204:107 - 115. 64 Park, S.J., and B.R. Buttery. 1989. Identification and characterization of common bean (Phaseolus vulgaris L.) lines well nodulated in the presence of high nitrate. Plant and Soil. 119: 237 - 244. Park, S.J. and B. R. Buttery. 2006. Registration of ineffective nodulation mutant R69 and non nodulation mutant R99 common bean genetic stocks. Crop Science. 46(3):1415 - 1417. Pereira, P.A.A., R.H. Burris and F.A. Bliss. 1989. 15 N - Determined dinitrogen fixation poten tial of genetically diverse bean lines ( Phaseolus vulgaris L.). Plant and Soil. 120:171 - 179. Somasegaran, P. and H.J. Hoben (Eds), 1994. Handbook for Rhizobia, Methods in Legume - Rhizobium Technology. Springer - Verlag, New York, Inc. St. Clair, D.A. , D.J. Wolyn, J. DuBois, R.H. Burris, and F.A. Bliss. 1988. Field comparison of dinitrogen fixation determined with nitrogen - 15 - depleted and nitrogen - 15 - enrichded ammonium sulfate in selected inbred backcross lines of common bean. Crop Science. 28:773 - 7 78. Tsai, S.M. R.O. Nodari, D.H. Moon, L.E.A. Camargo, R. Vencovsky, and P.Gepts. 1998. QTL mapping for nodule number and common bacterial blight in Phaseolus vulgaris L. Plant and Soil. 204:135 - 145. Thomas, R.J., J.R. McFerson, L.E. Schrader, and F.A. Bliss. 1984. Composition of bleeding sap nitrogen from lines of field - grown Phaseolus vulgaris L. Plant and Soil. 79:77 - 88. USDA - ERS, 2008. United States Department of Agriculture - Economic Research Service. Vegetables and Pulses Data. http://www.ers.usda.gov/data - products/vegetables - and - pulses - data.aspx . Accessed June 2015. USDA - ERS, 2014. United States Department of Agriculture - Economic Research Service. Veget ables and Pulses Data. http://www.ers.usda.gov/data - products/vegetables - and - pulses - data.aspx . Accessed June 2015. Warncke, D., J. Dahl, and L. Jacobs. 2009. Nutrient recommendations for field crops in MI. Michigan State University Extension Bulletin E2904. Wolyn, D.J., D.A. St Clair, J. DuBois, J.C. Rosas, R.H. Burris, and F.A. Bliss. 1991. Distribution of nitrogen in common bean (Phaseolus vulgaris L.) genotypes selected for differences in nitrogen fixation ability. Plant and Soil. 138:303 - 311. 65 CHAPTER 3 NITROGEN FIXATION ABILITY OF A PUEBLA 152/ZORRO RIL POPULATION EVALUATED UNDER GREENHOUSE AND FIELD CONDITIONS Abstract Variability in symbiotic nitrogen fi xation (SNF) can be found within Phaseolus vulgaris L. The but it is poorly adapted to cultivation at northern latitudes due to long season and indeterminate type I II growth habit. The recombinant inbred line (RIL) population developed by crossing Puebla enhanced SNF ability. The recombinant inbred line (RIL) population w as evaluated in the greenhouse under N free conditions, and under low N conditions in the field in East Lansing (EL), MI and in the field in Isabela, Puerto Rico (PR). Site year averages for p ercent N derived from the atmosphere (%Ndfa) ranged between 12. 7 % up to 66.6 %, although individual RILs ranged up to 90.5 %Ndfa. Traits measured in the greenhouse such as shoot biomass and biomass difference correlated moderately with %Ndfa traits measured in the field. Introduction Common or dry bean ( Phaseolus v ulgaris L.) is capable of fixing N through an association with Rhizobium spp. but common bean is often considered poor at fixing N (Buttery et al., 1992; Bliss, 1993; Piha and Munns, 1987; Fageria et al., 2014) in comparison to soybean ( Glycine max (L.) Me rr. ) and chickpea ( Cicer arietanum L.). The range in percent N derived from the atmosphere (%Ndfa) dry bean is reported to average 40 %Ndfa ranging up to 73 %Ndfa while soybean averages 68 %Ndfa but can reach as high as 95 %Ndfa in laboratory settings (H erridge 66 et al., 2008). In field studies, common bean average d 36 %Ndfa compared with crops such as soybean and chickpea reaching 58 %Ndfa and 65 %Ndfa, respectively (Herridge et al., 2008). Piha and Munns (1987) conducted an acetylene reduction assay to determine the activity of the nitrogenase in the nodules of the species. They noted that dry bean evolved more H + during fixation than soybean or chickpea representing a reduction in efficiency of SNF for bean. They also compared the size and number of th e nodules and discovered that the nodules of dry bean were smaller but more numerous than soybean or chickpea. Piha and Munns (1987) also noted that the period between germination and flowering of dry bean was much shorter, 27 days in their study, compare d to chickpea, which averaged 34 days to flowering. In addition, the interval between flowering and physiological maturity was much shorter for chickpea suggesting that the chickpea simply had more time in a vegetative state to establish nodules and fix N before the strong sink strength of the seed for photosynthate competes with the nodules for resources (Piha and Munns, 1987). In an effort to better calculate the contribution of pulse crops to soil N levels, Walley et al. (2007) used published data from the Northern Great Plains area to investigate the contributions of N fixation in pulse crops such as pea ( Pisum sativum L. ), lentil ( Lens culinaris Medik.) , chickpea, dry bean and faba bean ( Vicia fabia L.) . Similar to other findings Walley et al. (2007) determined that dry bean had the lowest average percent nitrogen derived from the atmosphere (%Ndfa) (40 %Ndfa) and also had the highest amount of variability year to year and by location among the pulse crops studied. Faba bean was the highest fixer achi eving 84 %Ndfa. Values reported for %Ndfa are also quite variable in dry bean. Pereira et al. (1989) investigated the N fixation ability of a wide range of dry bean genotypes grown in the field in Hancock, Wisconsin. The average %Ndfa was 21.6%, however, the range was from 5.7% for the navy 67 between %Ndfa and seed yield as some lines that exhibited a higher %Ndfa were not necessarily the best adapted or the highes t yielding lines (Pereira et al., 1989). The ability of the genotype to partition fixed N into seed is an important trait that cannot be overlooked as selecting on high %Ndfa may result in genotypes that are not efficient in partitioning. N fixation in dry bean is a quantitatively inherited trait. St. Clair and Bliss (1991) demonstrated that enhanced SNF ability was heritable in a population of inbred backcross lines (IBLs) developed from Puebla 152 and Sanilac, which was the recurrent parent. The goal was to produce lines with the agronomic characteristics of Sanilac with enhanced SNF ability. Select IBLs were intercrossed to generate F 3 families which were then evaluated for %Ndfa (St. Clair and Bliss, 1991). Several F 3 families did have a higher %N dfa (51.2 %Ndfa to 60.8 %Ndfa) than Puebla 152 (50.4 %Ndfa) but none of those lines yielded higher than Puebla 152, (St. Clair and Bliss, 1991). It is possible to select progeny with enhanced SNF ability though yield may not be correlated to SNF traits suc h as %Ndfa (St. Clair and Bliss, 1991; Elizondo Barron et al., 1999; Bliss, 1993; Buttery et al., 1992). To better understand the importance of partitioning of fixed N into the seed, Wolyn et al. (1991) utilized lines selected from the two IBL population s derived from Puebla 152 and either Sanilac track the partitioning of N through development. Those lines with the highest %Ndfa at R3 did not produce the hig hest %Ndfa at maturity. Line 24 - 17 had 17.5 %Ndfa at R3, which was the 2 nd lowest of the 5 lines studied while at maturity produced the highest value (44.6 %Ndfa; Wolyn et al., 1991). Maturity was positively correlated with N fixation, though some lines, such as line 24 - 21 matured earlier than Sanilac but had a higher %Ndfa which suggest s that it is possible to 68 select lines that were early maturing with enhanced SNF ability than the recurrent parent Sanilac (Wolyn, et al., 1991). s partitioning and discrimination of 15 N can result in over or underestimation of %Ndfa. Lazali et al. (2014) looked at six RILs selected from a population developed by crossing BAT477 and DOR364. The individual lines were selected for their tolerance or sensitivity to P deficiency. They found that there were differences across P levels and among the genotypes in their discrimination against 15 N (Lazali et al., 2014). Looking at different portions of the plant - roots, shoots, and nodules, they found that a higher proportion of 15 N remained in the roots, specifically the nodules, while the proportion of 15 N/ 14 N was lower in the shoots. Thus, measuring 15 N in the shoot might result in an overestimation of %Ndfa. Lazali et al. (2014) did find a significant correlation between 15 N in nodules and P sensitivity and N fixed by the plant. In addition to the genotypes discriminating against 15 N, different strains of Rhizobia also discriminate differently for 15 N. Yoneyama et al. (1986) found that not only did the 10 Rhizobium strains vary in their SNF ability, the amount of 15 N in the shoot of the three inoculated by different strains. In the laboratory, it is possible to control the specific strain of Rhizobium, however, in the field nodule occupancy is likely to vary even within the same plant. Yoneyama et al. (1984) noted that the amount of 15 N varied by the plant organ with stems and petioles having considerably less 15 N than leaves. Studying the kidney bean genotypes , 15 N was +9.3 and +8.5 in the nodules, +1.5 and - 1.0 for pods, and - 0.6 and +2.8 for stems, respectively (Yoneyama et al., 1984). The selection of the plant part in calculations cou ld result in a very different estimation of %Ndfa. 69 While Puebla 152 is poorly adapted to production at northern latitudes due to late maturity and vigorous indeterminate type III growth habit it has been recognized as a valuable source of enhanced SNF ch aracteristics. The purpose of this study was to generate a RIL population with Puebla 152 to investigate characteristics associated with SNF in a black bean mapping population adapted to production in northern latitudes. Materials and Methods Plant Mate rial A recombinant inbred line (RIL) mapping population consisting of 122 lines was generated and used in the quantitative trait loci (QTL) analysis of traits associated with SNF. The landrace selection Puebla 152 was selected as the donor parent due to i ts enhanced SNF ability (Chaverra and Graham, 1992). Puebla 152 is a small black seeded genotype belonging to the Mesoamerican gene pool and originates in Puebla, Mexico (St. Clair and Bliss, 1991). Puebla 152 is poorly adapted to production in northern latitudes due to its late maturity and type III growth habit. The commercial cultivar Zorro (Kelly et al., 2009) was selected as the other parent because it is efficient and has a type II growth habit. Zorro is also a small seeded black bean cultivar ada pted to production in northern latitudes, and is not known to possess enhanced SNF ability. In the fall of 2007 Zorro was crossed with Puebla 152 in the Plant Research Greenhouses at Michigan State University in EL, MI. Seed was harvested and planted in the greenhouse in February 2008 to grow out F2 plants. F2 seed was space planted in the field in June 2009 at the Saginaw Valley Research and Extension Center (SVREC) in Frankenmuth, MI. A single pod was harvested from each of 150 plants in September 20 09. A single seed from each pod was 70 planted in the greenhouse in September 2009 to begin two generations of single seed decent (SSD). Seed from each plant was harvested separately in May 2010 resulting in F 5 seed which was carried forward as a line. See d of 122 F 4: 5 lines was increased in June 2010 in the field at SVREC. Selfed seed of F 4:5 lines was used for all further evaluations in the field and greenhouse. Field Trials For field trials in 2011, 2012, and 2013 122 RIL lines , parents, and five comme rcial genotypes were planted. The checks included the parents, Zorro and Puebla 152, and bean cultivar ) and PR0443 - 151 (a black bean from PR that performed under low fertility - sized whit e bean cultivar from PR; Beaver et al., 2008), TARS - LFR1 (small red bean developed in PR; Porch et al., 2014), and PR1147 - 6 (black seeded breeding line developed in PR) were included in EL in 2012 and 2013. In addition the non - ark and Buttery, 2006) was included as a reference to percent nitrogen derived from the atmosphere (%Ndfa). Field plots for the SNF population consisted of four - 6 m rows spaced 50 cm apart. The outer two rows were planted to an erect, non - vining commer border effects. All lines in the RIL population are black, so the utilization of the white seeded border facilitated removal of border seed mixed with the sample. The center two rows of each plot were the yield row s where the experimental lines were planted. The plot was located on the old Soils Farm located on the campus of Michigan State University, East Lansing, MI. The field was selected because of the low soil nitrogen. Soil type was a Capac Loam and Riddles - Hillsdale Sandy Loam ( USDA - NRCS, 2013 ) . The same field was utilized for 71 this study in 2011, 2012, and 2013. Corn was the crop grown on the field in 2010. Soil test results are given in Table 3.01. Prior to planting, pre - emergence herbicides Sonalan ( Syngenta Crop Protection LLC, ), Eptam 7E Selective Herbicide (S - ethyl dipropylthiocarbamate) (Gowan, Yuma, AZ), and Dual (S - Metolachlor) (Dow AgroSciences), were applied at a rate of 69 g ha - 1 active ingredient, 403 g ha - 1 active ingredient, and 351 g ha - 1 active ingredient respectively. A postemergence application of the herbicides Raptor Herbicide (Imazamox) (BASF Corporation, Research Triangle Park, NC) at a rate of 5.7 g ha - 1 active ingredient and Basagran (Sodium salt of bentazon* (3 - (1 - methylethyl) - 1H - 2,1,3 - benzothiaddiazin - 4 (3H0one 2,2 - dioxide) (BASF Corporation, Research Triangle Park, NC) at a rate of 115 g ha - 1 active ingredient were applied to control weeds prior to flowering of the beans. Potato leaf hoppers ( Empoasca fabae ) was controlled as needed with Warrior (Lambda - - ( + ) - cyano - (3 - phenoxyphenyl)methyl - 3 - (2 - chloro - 3,3,3 - trifluoro - 1 - propenyl) - 2,2 - dymethylcyclopropane carboxylate) (Syngenta Crop Protection LLC) at a rate of 12 g ha - 1 active ingredient. Seed was p lanted using a four row air planter. No fertilizer was applied to the field at any time to maintain low N conditions. Seed was treated with rhizobial inoculant consisting of fine buffered peat (American Peat Technology, Aitkin, MN , USA ) carrying Rhizobium tropici strain CIAT899 which was prepared by culturing the rhizobia in yeast mannitol broth for three d prior to mixing with peat to allow sufficient wetting of the peat. The resulting inoculant was incubated in the dark for 8 to 12 wk at room temperatur e. Seed was mixed with the peat inoculant and a small amount of water to adhere the peat inoculant to the seed. Treated seed was stored in a cool room until planting. 72 The study was repeated in Isabela, PR in winter 2012 and 2013. The field site was ma intained as a low fertility site. Soil type was an acidic Coto Clay, very fine, kaolinitic, isohyperthermic typic eutrustox. Seed was planted in 3 m rows spaced 60 cm apart with 3 replicates and a single row was planted for each genotype. Data C ollecti on Plant stand was measured by randomly placing two 1 - m rulers randomly in each plot along the yield rows and counting the number of seedlings falling within the 1 m length. Plant stand was measured at the first trifoliate stage. Days to flower was recor ded as the number of days after planting when 50% of the plants in each plot had one open flower. Maturity was determined when 50% or more of the plants had reached physiological maturity, at which time plant height (cm), lodging (1=upright, 5=laying on t he ground), and agronomic desirability (1= poor, 6=superior) were recorded. A Minolta SPAD 502 Meter (Konica Minolta, Inc., Tokyo, Japan) was used to measure chlorophyll content at early pod fill. The last completely expanded leaf was selected to measur e chlorophyll content. Harvest When plants had reached harvest maturity, plots were pulled mechanically and raked into piles. Piles were weighed to measure biomass, then threshed with a Wintersteiger AG plot combine (Winterstieger AG, Austria). Seed was air dried, cleaned with a Clipper Mill (A.T. Ferrell Company, Bluffton, IN, USA) before weighing. Seed moisture was measured with a Dickey - john GAC 2500 moisture meter (Churchill Industries, Minneapolis, MN). Yield was calculated by adjusting values to 1 8 % moisture content. A biomass sample was collected from each plot, weighed the day of harvest, dried for 7 d in a dryer at 60º C, and weighed again to determine 73 moisture content at harvest. At maturity in Isabela, PR, 2 m of row was hand harvested and t hrashed with similar methods to measure seed weight and yield for seed harvested in EL, MI. Nitrogen Analysis A 30 g subsample from each plot was placed in an envelope and placed into a dryer at 60º C for 1 wk. Seed and biomass was then ground in a Wiley Mill pass through a 1 mm mesh screen. Seed and biomass samples were then stored at room temperature until sent to the Stable Isotope PDZ Europa ANCA - GSL elemental analyzer interfaced to a PDZ Europa 20 - 2 0 isotope ratio mass spectrometer (Sercon Ltd., Cheshire, UK) atmosphere. To determine %Ndfa, the following equation was used (Ramaekers et al., 2013): %Ndfa = ( 15 N non fixing re ference plant - 15 N fixing legume ) x 100 ( 15 N non fixing reference plant B) With B representing the 15 N of the fixing legume grown under N free conditions in the greenhouse where it relies completely on SNF for its nitrogen requirement. Since e ach genotype studied was also evaluated under N free conditions in the greenhouse each genotypes respective 15 N was used. Greenhouse Assay To study the SNF ability of the RILS at flowering, the same 122 RILs and five commercial checks were grown under greenhouse conditions. Seeds of the genotypes being studied were sterilized by soaking in a 10% bleach solution for 2 min followed by two 2 - min rinses with sterile water. Six seeds were planted into each plastic 5.7 L nursery container which had been fil led with a 2:1 mix, v:v of perlite to vermiculite which had been autoclaved. Seeds were watered in with tap water. At 3 d after germination, 500 ml YMB culture of R . tropici CIAT899 74 was diluted in 20 L of tap water which had been adjusted to a pH of appr oximately 6.5 which resulted in a concentration of approximately 10 3 cells per ml. Each nursery container received 250 ml of the final rhizobial dilution. Inoculation was repeated 10 d after planting in the same manner to ensure sufficient population leve ls of rhizobia to effect symbiosis. Pots were placed randomly on a greenhouse bench with day length extended to 16 h with HPS lights. Two to three times weekly each pot was watered with 500 ml full strength Broughton and Dilworth N free solution (Brough ton and Dilworth, 1970) as needed to avoid drought stress. Plants were thinned to two plants per pot before emergence of the first trifoliate. The analysis was repeated three times. When one or both of the plants had at least one flower open, plant shoo ts were measured and cut with a razor blade at soil level. The perlite and vermiculite was carefully removed from the roots which were measured for maximum length and scored from nodulation on a scale of 0 to 6, with 0 representing roots without nodules a nd 6 being roots with a large number of fully developed and functioning nodules. Both root and shoot biomass samples were dried in a dyer at 60º C for 7 d at which time they were ground to pass through a 1mm screen on a Wiley mill. Samples were also sent the U.C. Davis SIF for 15 N analysis. Biomass difference for shoot, root, and whole plant were calculated using the non - nodulating R99 as a reference as follows: Biomass Difference = Mass (g) fixer - Mass R99 (g)/Mass (g) fixer Statistical Analysis A PROC GLM analysis using SAS 9.4 (SAS Institute Inc. 100 SAS Campus Drive, Cary, NC 27513 - 2414, USA) to generate an ANOVA showed that there was a significant difference among 75 years and locations. Consequently, each year and location was analyzed separately. PROC CORR was used to gen erate Pearson Correlations. Results and Discussion Mean seed yield of the RILs ranged from a low of 1807 kg ha - 1 in EL in 2012 to a high of 3026 kg ha - 1 in 2013 and from 835 kg ha - 1 in 2012 to 984 kg ha - 1 in 2013 in PR (Table 3.02). In all years precipi tation was below average at EL, and precipitation in 2012 in EL was less than 2011 and 2013 resulting in significant water stress (Table 3.01). Average seed yield follows the same trend as precipitation with the 2012 season producing the lowest yields com pared to 2011 and 2013 (Table 3.02). In all years and locations, Zorro yielded more than Puebla 152 except in EL in 2013 suggesting that Zorro is better adapted to both locations. The low yields observed for Puebla 152 resulted from an overall lack of ada ptation at both locations and from its inability to mature within the normal growing season. The highest yielding RILs differed between years and location and yielded more than both Zorro and Puebla 152 which suggests that there was considerable transgress ive segregation for yield at both locations. Yield ranged from low 97 kg ha - 1 in PR in 2013 to a high of 4304 kg ha - 1 in EL in 2011. The non nodulating check R99 outyielded Zorro parent in 2012 in EL. Yields in PR varied each season with overall yields in 2012 lower than 2013 but variability was very high (CV~60%) both years (Table 3.02). R99 yielded significantly below the average both years in PR and was equivalent to the local check Verano indicative of low N levels in the test site. The two lines PR04 43 - 151 and TARS - LFR1 bred for low fertility conditions in PR were the highest yielding lines both years providing supporting evidence for low soil fertility levels in PR. The test site at Isabella PR was specifically chosen as a low fertility site and b eans had been continuously grown on the site for 76 five years prior the 2012, so the soil had become heavily infected with root rot pathogens which introduced added variability to the yield results. Percent N in seed ranged from 2.8 to 4.6% and from 0.38 to 1.7% in biomass over years and locations (Table 3.03). Values were more consistent and did not follow the highly variable trends observed for seed yield. Among the parents and checks, Puebla 152 tended to have higher N values in the seed (3.6%) and R 99 had the lower values (2.8%). The RILs showed a wider range on both extremes than either parent. Comparing years and locations, % N in seed and biomass was not significantly different year to year (data not shown). Since there were year and site differ ences for yield, %N was presented similarly for consistency. It is notable that the CV for seed N were rather small whereas those for biomass were considerably larger. Those RILs with maturities similar to or later than Puebla 152 had not completely drie d down by harvest and thus had a larger biomass at harvest compared to more efficient RILs , which were more similar to the efficient parent Zorro, and produced lower biomass. (Table 3.04). Each method offers advantages and disadvantages while each can possibly lead to incorrectly estimating the actual N fixed (Peoples et al., 2009). Both the difference method and 15 N natural abundance depend on the characteristics of the non - fixing reference plant with regard distributio n within the plant. In this study the non - fixing navy bean R99 was utilized as a reference as it is the same species as the Puebla 152 and Zorro parents (Singh et al., 1991). The difference method tends to estimate a higher %Ndfa than natural abundance ( % Ndfa - 15 N ) and was more consistent year to year in maintaining a similar trend among the checks with Puebla 77 152 generally fixing a greater portion of N from the atmosphere than Zorro in all but one year, EL 2011. Aside from the difficulties associated wi th measuring %Ndfa there were RILs with greater %Ndfa than either the Puebla 152 or Zorro parent (Table 3.04). One line (B11519) had a higher %Ndfa than Puebla 152 in all but 1 location/year. Values ranged a high of 68.4% for the highest fixing RIL in P .R. in 2012 down to a low of 0 % in E L in 2012. Given that Puebla 152 is late maturing compared to Zorro (Table 3.06) some RILs are better adapted to production in northern latitudes, while combining the enhanced SNF abilities of Puebla 152. Considering the N in seed (g), however, R99 produces more N in seed than Zorro in 2 of 5 location/years (Table 3.05). Comparing the results from 2011 (a favorable growing season) to 2012 (a dry season) when water stress was lower, Zorro had a greater seed N yield , w hereas the relationship was reversed in the dry year (Table 3.05). This inconsistency in performance year to year suggests that other factors are having an effect on N dynamics within the plant and response to stresses such as drought can alter SNF abilit y. In some situations such as drought in EL in 2012, R99 accumulated more seed N than any other genotype studied resulting in low estimations of %Ndfa using the difference method. Nitrogen derived from the atmosphere was also low using the 15 N natural ab cross discriminated against 15 N under P stress. While not a legume, Robinson et al. (2000) found that in wild barley ( Hordeum spontaneum C. Koch) discriminated against assimil ating 15 N from the soil under drought conditions. Perhaps these differences in affinities for 15 N depending on genotype and environmental stress contribute to inaccuracies in measuring %Ndfa. Both Lazali et al. (2014) and Robinson et al. (2000) found tha t not only was 15 movement within the plant was limited with a higher proportion of 15 N remaining in the roots. 78 In 2011 and 2012 Zorro had higher seed N yield than Puebla 152, 107.5 kg ha - 1 and 74.8 kg ha - 1 compared to 79.4 k g ha - 1 and 33.9 kg ha - 1 , respectively (Table 3.05). The opposite trend was seen in 2011 and 2012 for biomass, Puebla 152 yielded 278.4 kg ha - 1 and 101.4 kg ha - 1 respectively compared to Zorro which yielded 50.0 kg ha - 1 and 15.8 kg ha - 1 , respectively (Tabl e 3.05). The resulting N harvest index (NHI) is substantially lower for Puebla 152, 24.2 % in 2011 and 26.2 % in 2012 and much higher in Zorro at 68.3% and 82.1 % in 2011 and 2012, respectively (Table 3.05). Zorro is much more efficient than Puebla 152 a nd partitions a greater proportion of N accumulated into the seed, which is highly desirable in a modern dry bean cultivar. Given the total amount N in plant biomass and seed in the same field, Puebla 152 (357.8 kg ha - 1 in 2011 and 135.3 kg ha - 1 in 2012) compared to Zorro (157.6 kg ha - 1 in 2011 and 90.6 kg ha - 1 in 2012) Puebla 152 likely obtained a significantly higher proportion of N from the atmosphere but is very inefficient compared to Zorro. In 2011 the NHI of RIL B11560 was 84.1% and the second hig hest B11617 at 77.6 %. In 2012 B11617 had the highest NHI at 85.5 % which was higher than Zorro in each year. Both B11617 and B11560 had slightly below average yield (data not shown). Increasing the amount of fixed N is not useful if that N is not subseq uently partitioned into the seed. Genotypes such as Puebla 152 (NHI = 26.2%) leave the majority of N fixed in the field within the straw residue. Overall, N yield was lower in PR compared to EL for checks, including those developed in PR (Table 3.05). Aver age flowering and maturity were earlier in PR than EL, the shorter vegetative phase, combined with less time between flowering and harvest resulted in less time to fix N and could account for some of the differences. In addition, the field in PR had a lon g history of dry bean production resulting in higher root disease pressure which may have further contributed to lower N yields in PR. 79 Generally, Puebla 152 had the longest days to flower and maturity of all checks in EL (Table 3.06). In PR, Puebla 152 had similar maturity to Zorro and the other checks. The shorter day length in PR may have helped Puebla 152 to initiate flowering in a manner similar to the day length insensitive checks which resulted in similar maturity (Table 3.06). In EL, however, long days in summer may have contributed to the late maturity of Puebla 152. For both flowering and maturity transgressive segregation was seen as there were RILs that exceeded Puebla 152 as well as RILs which flowered and matured earlier than Zorro (Table 3.0 6). In the N - free greenhouse analysis none of the RILs accumulated as much biomass as Puebla 152 (Table 3.07). Puebla 152 also had the highest % N in the shoot, shoot:root ratio, and biomass difference (Table 3.07). Since the only N available to the p lant was that present in the seed at planting, in the tap water, or in dust deposited on the potting media it would be expected that the vast majority of the N found in the plant was derived from the atmosphere which resulted in the high average 94.8 %Ndfa . As N is fixed in the nodules it must be mobilized through the plant into the biomass and the amount of biomass accumulated depends on the amount of N available. This circular relationship is further enforced by the fact that photosynthate is necessary to provide the energy for nodule function. Puebla 152 produces a shoot to root ratio of 3.28 compared to Zorro at 2.63 and R99 at 0.9. Perhaps a mechanism employed by dry bean plants is to support root growth to mine N from the soil when N is limiting wh ile resources are only invested in the shoot as N becomes less limiting. Ideally, greenhouse screening could be a useful tool in selecting genotypes with enhanced SNF ability avoiding the need for expansive and costly field studies. Looking at correlati ons of greenhouse traits to field traits some greenhouse traits may be helpful in selecting genotypes for improved performance in the field. Interestingly greenhouse traits did not correlate with yield in 80 the field except in EL in 2012 (Table 3.08). Trait s moderately and inversely related with yield were: shoot weight (r= - 0.4, p < 0.0001), shoot N (r= - 0.3, p=0.0006), root weight (r= - 0.335, p < 0.0001), nodule rating (r= - 0.201, p =0.0244 ), shoot difference (r= - 0.305, p =0.0005 ), root difference (r= - 0.323, p=0.000 2), and total biomass difference (r= - 0.312, p=0.0004). At this location, 2012 was the driest year and plants were exposed to drought stress for much of the season. All of these traits are measuring the ability of the genotype to accumulate biomass with s hoot weight and root weight being positively correlated (r=0.82, p < 0.0001) in the greenhouse (Table 3.10). Growth of biomass may not support an increase in yield parameters if the biomass that is accumulated is not translocated into the seed and in fact, h arvest index (HI) is positively correlated to yield in the field (Table 3.09). Puebla 152 has a low HI (ranging from 0.07 to 0.27 in EL from 2011 to 2013) compared to the more efficient Zorro (ranging from 0.31 to 0.45 in EL from 2011 to 2013). Under dro ught conditions a plant must mobilize resources from the biomass into the seed, however, if the biomass is a stronger sink than the seed the result could be lower yield resulting in the negative correlations observed. Many of the greenhouse traits are mod erately correlated with Ndfa calculated using the difference method. Shoot weight in the greenhouse is associated with Ndfa difference in 2 of 3 field seasons (r=0.27, p < 0.05 in 2011 and r=.23, p < 0.01 in 2013) (Table 3.08). Looking at the relationship amo ng field traits %N in seed is inversely correlated to seed yield in 2 of 3 years (Table 3.09). One might expect that as the yield increases the concentration of N in the seed might decrease. The same trend is seen in the greenhouse analysis on plants grow n without additional N. Shoot weight and %N are inversely correlated (r= - 0.30, p<0.0001) (Table 3.10) suggesting that the N use efficiency is an important component of yield when improving traits for SNF. Root, shoot, and total biomass was inversely corre lated with 15 N (r= - 0.33, - 0.27, 81 and - 0.32, p < 0.0001, respectively) (Table 3.10). The lower the 15 N , the greater %Ndfa which stands to reason that those plants with the lowest 15 N are fixing more N and thus able to accumulate more biomass. Similarly, 1 5 N was highly negatively correlated, (r= - 0.66, p < 0.0001) with %Ndfa difference. This confirms that those genotypes with lower 15 N fix more N. Traits measured in the greenhouse were also significantly correlated to SNF traits measured in the field in PR (Table 3.11). Greenhouse shoot weight was correlated with %N in seed in both 2012 and 2013 (r=0.38, p < 0.0001 and r=0.38, p < 0.05). Biomass difference in the greenhouse was correlated with %Ndfa calculated using either the natural abundance (r=0.39, p < 0.0 001) or difference method (r=0.40, p<0.0001) in 2012 and natural abundance (r=0.33, p<0.05) and difference method (r=0.18, p<0.05) in 2013 (Table 3.11). The consistency with which biomass difference is correlated makes this trait suited to use as a selecti on tool when breeding for lines with enhanced SNF. Shoot weight in the greenhouse was positively correlated with all traits measured in the greenhouse (Table 3.10). Shoot weight of genotypes when grown under N - free conditions may be a useful trait to use in selecting plants with improved SNF traits such as %Ndfa in the field and %N in seed (Table 3.08). Conclusions There is considerable variability for SNF characteristics within the Puebla 152/Zorro RIL population. Several RILs combined enhanced N fixat ion ability with plants better adapted to agronomic conditions in northern latitudes. These RILs could be useful in developing lines better able to acquire a larger proportion of their N needs from the atmosphere. Drought may confound evaluation of SNF a s %Ndfa appeared to be lower in years with limited precipitation. Traits measured in the greenhouse may be useful to select for field traits, though stress such as 8 2 drought may confound the results making field evaluation over several season important in de veloping a better understanding of the traits being studied. 83 APPENDICES 84 APPENDIX A CHAPTER 3 TABLES Table 3.01. Planting date, precipitation, and soil characteristics of plots in East Lansing, MI, where Puebla 152/Zorro RILs were grow n. Year Date Planted Precipitation (mm) pH % N 2011 6/13/2011 315 5.3 0.03 2012 6/19/2012 172 6.5 0.067 2013 6/4/2013 298 6.5 0.08 30 year average 361 Report generated at Enviro - Weather (http://www.agweather.geo.msu.edu/mawn/) for the months of June through September for the Hancock Turfgrass Research Center, East Lansing, MI. Total Kjeldahl Nitrogen measured by the Soil and Plant Nutrient Lab at Michigan State University, East Lansing, MI. 85 Table 3.02. Yield of commercial checks and 122 Pu ebla 152/Zorro RILs grown in East Lansing in 2011 to 2013 and Puerto Rico in 2012 and 2013. East Lansing Puerto Rico Commercial Checks 2011 2012 2013 2012 2013 ------------------------------------ kg ha - 1 -------------------------------- -- Puebla 152 2204 918 2272 447 714 Zorro 3222 1933 2159 1225 811 R99 2913 2736 1755 261 458 Medalist 2194 1719 513 636 PR0443 - 151 3273 1767 2316 1297 2739 Verano 2197 2008 233 367 TARS - LFR1 2140 2060 1633 2200 PR1147 - 6 1678 1831 872 1953 RILs Highest yielding 4304 2732 3042 1692 2514 lowest yielding 1076 763 1405 228 97 Test M ean 3026 1807 2223 835 984 LSD (p < 0.05) 90 0.6 751 . 3 46 2.6 871.3 1225.2 CV% 17.8 24.9 12.9 64.8 58.1 Highest and lowest value in each trait of advanced breeding lines. May not be the same line for each trait. 86 Table 3.03. Percent N in the biomass and seed of commercial checks and 122 Puebla 152/Zorro RILs grown in East Lansing in 2011 to 2013 and Puerto Rico in 2012 and 2013. East Lansing, MI Puerto Rico Commercial 2011 2012 2013 2012 2013 Checks Seed Biomass Seed Biomass Seed Seed Seed -------- ---------------------------------------------- % N ------------------------------------------------------------- Puebla 152 3.60 0.81 3.74 1.22 3.70 3. 22 3.6 1 Zorro 3.31 0.54 3.88 0.49 3.67 3. 37 3.29 R99 2.82 0.79 3.90 0.93 3.63 2.97 3.52 Medalist 3.80 0.79 3.45 3. 06 2.98 PR0443 - 151 2.89 0.70 3.66 0.74 3.41 3.0 2 2.98 Verano 3.98 0.95 4.15 3.45 3.67 TARS - LFR1 3.58 0.56 3.81 3.07 3.37 PR1147 - 6 4.28 0.87 4.19 3.66 3.29 RILs Highest 4.41 1.13 4.74 1.71 4.60 3.92 3.91 Lowest 2.45 0.38 2.99 0.48 2.94 2.24 2.94 Test M ean 3.35 0.66 3.86 0.90 3.8 3.30 3.37 LSD (p < 0.05) 0.60 0.23 0.46 0.42 0.39 0.55 0.44 CV% 10.8 20.9 7.0 27.9 6.3 10.4 8.1 Highest and lowest value in each trait of advanced breeding lines. May not be the same line for each trait. 87 Table 3.04. Percent N derived from the atmosphere (%Ndfa) calculated using the natural abundance method ( 15 N ) and the difference method for checks and 122 RILs of the Puebla 152/Zorro RIL population grown in East Lansing, MI in 2011 to 2013 and Isabella, Puerto Rico in 2012 and 2013 . %Ndfa values for some lines were negative and are not included in this table . Highest and lowest yield in each trait of advanced breeding lines. May not be the same line for each trait. 88 Table 3.05. Amount of N ( kg ha - 1 ) in seed and biomass, and N harvest index (NHI) of 122 Puebla 152/Zorro RILs grown in East Lansing, MI and Puerto Rico in 2011, 2012, and 2013. East Lansing 2011 East Lansing 2012 East Lansing 2013 Puerto Rico 2012 Puerto Rico 2013 Commercial Checks Seed Biomass Total NHI Seed Biomass Total NHI Seed - kg N ha - 1 - % - kg N ha - 1 - % - kg N ha - 1 - Puebla 152 79.4 278.4 357.8 24.2 33.9 101.4 135.3 26.2 84.1 14.7 24.8 Zorro 107.5 50.0 157.6 68.3 74.8 15.8 90.6 82.1 79.6 41.5 26.1 R99 82.5 91.1 173.6 47.8 107.5 51.0 158.4 68.6 64.0 7.8 16.1 Medalis t 83.8 49.7 133.6 66.5 59.4 15.9 18.8 PR0443 - 151 95.0 75.6 160.8 55.4 64.6 18.9 83.5 77.3 78.4 38.9 81.9 Verano 88.2 43.7 131.8 65.9 83.0 13.0 13.3 TARS - LFR1 76.6 16.0 92.6 83.1 78.3 50.0 74.1 PR1147 - 6 71.9 30.7 102.7 70.6 76.7 32.0 64.2 RILs Highest 151.3 251.3 318.0 84.1 104.2 223.6 259.0 85.5 125.0 54.8 86.8 Lowest 44.0 24.5 119.7 15.4 31.2 12.9 6 2.2 14.4 52.1 7.8 3.6 Test M ean 101.9 93.8 195.7 53.5 69.5 50.6 119.5 61.0 83.7 27.4 32.9 LSD (p < 0.05) 41.8 49.1 72.8 12.5 31.5 42.0 49.6 16.5 19.1 n.s. 41.4 CV% 24.5 31.2 22.2 14.0 26.9 49.1 24.4 15.9 1 4.1 6 4.7 74.3 Highest and lowest yield in each trait of advanced breeding lines. May not be the same line for each trait. 89 Table 3.06. Flowering, maturity, and plant height of 122 Puebla 152/Zorro RILs grown in East Lansing, MI and Puerto Ri co in 2011, 2012, and 2013. Days to Flowering Days to Maturity Height East Lansing Puerto Rico East Lansing Puerto Rico East Lansing --- Days --- -- cm -- Checks and Parents 2011 2012 2013 2012 2013 2012 2013 2012 2013 2012 2013 Puebla 152 54 57 45 39 40 109 101 82 88 40 35 Zorro 51 51 42 40 39 98 94 81 88 65 40 R99 41 46 43 39 40 100 94 82 88 60 45 Medalist 47 44 38 42 98 94 82 84 65 50 PR0443 - 151 47 49 44 40 38 9 7 93 82 84 55 35 Verano 45 43 39 39 104 95 78 88 55 45 TARS - LFR1 46 42 38 39 96 91 82 87 55 40 PR1147 - 6 47 43 38 39 98 95 81 84 68 45 RILs Highest 63 60 48 43 43 111 103 90 94 90 73 Lowest 42 45 41 35 35 95 90 81 84 35 25 Test M ean 48 51 44 39 39 101 96 82 87 62 43 LSD (p < 0.05) 4 .1 4.7 2.6 2 .6 3 .1 6 .1 2 .1 4.1 n.s. 20.5 14.8 CV% 5.1 4.4 3. 1 4.1 4.2 2.7 2.5 3.2 4.7 15.9 17.2 Highest and lowest yield in each trait of advanced breeding lines. May not be the same line for each trait. 90 Table 3.07. SNF traits measured in the greenhouse in East Lansing, MI on 122 Puebla 152/Zorro RILs grown in the greenhouse growing in N free conditions. Commercial Checks %N in biomass Shoot Weight Root Weight %Ndfa Difference Shoot/ Root Ratio Biomass Diff - Shoot Biomass Diff - Root Biomass Diff - Total Nodule Rating 15 N (g) Puebla 152 2.52 12.38 0.335 98.1 3.28 95.0 81.3 91.6 4.0 - 3.480 Zorro 3.06 4.44 0.136 94.4 2.63 83.5 59.8 77.0 4.5 - 3.620 R99 1.14 0.58 0.022 0.0 0.90 0.0 0.0 0.0 0.0 0.443 Medalist 3.43 2.94 0.101 91.8 2.30 74.8 45.6 66.1 4.0 - 3.290 PR0443 - 151 2.41 6.42 0.155 97.8 1.81 92.0 74.6 85.6 4.5 - 3.420 RILs § High 3.47 9.60 0.276 98.0 3.75 94.0 79.2 90.2 6.0 - 2.910 Low 2.08 2.76 0.081 88.7 1.79 58.6 35.1 42.6 2.0 - 3.850 Test M ean 2 . 63 5.58 0.144 94.8 2.72 86.5 63.1 80.0 4.6 - 3.223 LSD (p < 0.05) 0 . 57 2.41 0.075 7.4 0.78 8.0 16.0 11.0 1.9 0.49 CV% 14.2 28.8 34.3 5.1 18.5 6.1 16.1 9.0 20.2 - 9.3 Difference= (biomass of N fixer (g) - biomass non - fixer (g))/biomass of N fixer (g) . B value used in the natural abundance equat ion . § Highest and lowest yield in each trait of advanced breeding lines. May not be the same line for each trait. 91 Table 3.08. Pearson correlations between traits measured in the field in 2011 - 2013 and in the greenhouse on the Puebla 152/Zorro RIL p opulation grown in East Lansing, MI . Correlation significant at 0.0001 Difference= (biomass of N fixer (g) - biomass non - fixer (g))/biomass of N fixer (g) 92 Table 3.09. Pearson Correlations between traits measur ed in the field on the Puebla 152/Zorro RIL population grown in East Lansing, MI . Correlation significant at 0.0001 93 Table 3.10. Pearson Correlations between traits measured on the Puebla 152/Zorro RIL population grown under N - free conditions in the greenhouse in East Lansing, MI . Correlation significant at 0.0001 Difference= (biomass of N fixer (g) - biomass non - fixer (g))/biomass of N fixer (g) 94 Table 3.11. Pearson Co rrelations between traits measured in the field on the Puebla 152/Zorro RIL population grown in Isabela, Puerto Rico and in the greenhouse in East Lansing, MI . Correlation significant at 0.0001 95 APPENDIX B CHAPTER 3 SUPPLEMENTAL TABLES Table S 3. 01 . Yield of commercial checks and 122 Puebla 152/Zorro RILs grown in East Lansing in 2011 to 2013 and Puerto Rico in 2012 and 2013. 96 Table S 3. 01 ( ) 97 Table S 3. 01 ( ) 98 Table S 3. 01 ( ) 99 Table S 3. 0 2 Percent N in the biomass and seed of commercial checks and 122 Puebla 152/Zorro RILs grown in East Lansing in 2011 to 2013 and Puerto Rico in 2012 and 2013. 100 Table S 3. 0 2 ( ) 101 Table S 3. 0 2 ( ) 102 Table S 3. 0 2 ( ) 103 Table S 3. 03 . Percent N derived from the atmosphere (%Ndfa) calculated using the natural abundance method ( 15 N ) and the difference method for checks and 122 RILs of the Puebla 152/Zorro RIL population grown in East Lansing, MI in 2011 to 2013 and Isabella, Puerto Rico in 2012 and 2013 . 104 Table S 3. 03 ( ) 105 Table S 3. 03 ( ) 106 Table S 3. 03 ( ) 107 Table S 3. 03 ( ) 108 Table S 3. 03 ( ) %Ndfa values for some lines were negative and are not included in this table. 109 Table S 3. 04 . Amount of N ( kg ha - 1 ) in seed and biomass, and N harvest index (NHI) of 122 Puebla 15 2/Zorro RILs grown in East Lansing, MI and Puerto Rico in 2011, 2012, and 2013. 110 Table S 3. 04 ( cont d ) 111 Table S 3. 04 ( cont d ) 112 Table S 3. 04 ( cont d ) 113 Table S 3. 04 ( cont d ) 114 Table S 3. 04 ( cont d ) 115 Table S 3. 05 . SNF traits measured in the greenhouse in East Lansing, MI o f 122 Puebla 152/Zorro RILs grown in the greenhouse growing in N free conditions. 116 Table S 3. 05 ( ) 117 Table S 3. 05 ( ) 118 Table S 3. 05 ( ) 119 Table S 3. 05 ( ) 120 Table S 3. 05 ( ) Difference= (biomass of N fixer (g) - biomass non - fixer (g))/biomass of N fixer (g). B value used in the natural abundance equation. 121 L ITERATURE CITED 122 LITERATURE CITED Plant Registrations. 2(3):187 - 189. Bliss, F.A. 1993. Breeding common bean for improved biological nitrogen fixation. Plant and Soil. 152:71 - 79. Broughton, W.J., Dilworth, M.J., 1970. Methods in legume - rhizobium technology: plant nutrient solutions. In: Somasegaran, P., Hoben, H.J. (Eds.), Handbook for Rhizobia, Methods in Legume - Rhizobium Technology. Springer - Verlag, N ew York, Inc. Page 340. Buttery, B.R., S.J. Park, and D.J. Hume. 1992. Potential for increasing nitrogen fixation in grain legumes. Canadian Journal of Plant Science. 72:323 - 349. Chaverra, M.H., and P.H. Graham. 1992. Cultivar variation in traits af fecting early nodulation in common bean. Crop Science. 32(6):1432 - 1436. Fageria, N.K., L.C. Melo, E.P.B Ferreira, J.P. Oliveira, and A.M. Knupp. 2014. Dry matter, grain yield, and yield components of dry bean as influenced by nitrogen fertilization and rhizobia. Communications in Soil Science and Plant Analysis. 45:111 - 125. Herridge, D.F., M.B. Peoples, and R.M. Boddy. 2008. Global inputs of biological nitrogen fixation in agricultural systems. Plant Soil 311:1 - 18. Journal of Plant Registrations 3(3):226 - 230. Lazali, M., A. Bargaz, G. Carlsson, S.M. Ouname, and J.J. Drevon. 2014. Discrimination against 15 N among recombinant inbred lines of Phaseolus vu lgaris L. contrasting in phosphorus use efficiency for nitrogen fixation. Journal of Plant Physiology 171:199 - 204. Park, S.J. and B.R. Buttery. 2006. Registration of ineffective nodulation mutant R69 and non nodulation mutant R99 common bean genetic stoc ks. Crop Science. 46(3):1415 - 1417. Peoples, M.B., M.J. Unkovich, and D.F. Herridge. Measuring symbiotic nitrogen fixation by legumes. In Agronomy Monograph 52: Nitrogen Fixation in Crop Production. D.W. Emerich, H.B. Krishnan (eds.) Pereira, P.A.A., R.H. Burris and F.A. Bliss. 1989. 15 N - determeind dinitrogen fixation potential of genetically diverse bean lines ( Phaseolus vulgaris L.). Plant and Soil 120:171 - 179. P iha, M.I. and D.N. M unns. 1987. Nitrogen fixation potential of b eans ( Phaseolus vulga ris L.) compared with other grain legumes under controlled c onditions . Plant and Soil. 98:169 - 82. Porch, T.G., J . S. Beaver, G . Abawi, C . Estévez de Jensen, and J . R. Smith . 2014. Registration of a small - red dry bean germplasm, TARS - LFR1, w ith multipl e disease resistance and superior performance in low nitrogen soils . Journal of Plant Registrations 8:177 182. 123 Ramaekers, L. C.H. Galeono, N. Garzon, J. Vanderleyden, M.W. Blair. 2013. Identifying quantitative trait loci for symbiotic nitrogen fixation capability and related traits in common bean. Molecular Breeding. 31:163 - 180. Robinson, D., L.L. Handley, C.M. Scrimgeour, D.C. Gordon, B.P. Forster, and R.P. Ellis. 2000. Using stable isotope natural abundances ( 15 13 C) to integrate stress respo nses of wild barley ( Hordeum spontaneum C. Koch.) genotypes. Journal of Experimental Botany. 51(342):41 - 50. Singh, S.P., P. Gepts, and D.G. Debouck. 1991. Races of common bean (Phaseolus vulgaris, Fabaceae). Economic Botany. 15(3):379 - 396. St. Cla ir, D.A. and F.A. Bliss. 1991. Intrapopulation recombination for 15 N - determined fixation ability in common bean. Plant Breeding 106:215 - 225. USDA - NRCS , 2013. United States Department of Agriculture - National Resources Conservation Service. A vailable at http://websoilsurvey.nrcs.usda.gov/app/ Accessed September 2013) Walley, F.L., G.W. Clayton, P.R. Miller, P.M. Carr, and G.P. Lafond. 2007. Nitrogen economy of pulse crop production in the Northern G reat Plains. Agronomy Journal. 99:1710 - 1718. Wolyn, D.J., D.A. St Clair, J. DuBois, J.C. Rosas, R.H. Burris, and F.A. Bliss. 1991. Distribution of nitrogen in common bean (Phaseolus vulgaris L.) genotypes selected for differences in nitrogen fixation ability. Plant and Soil. 138:303 - 311. Yoneyama, T., N. Yamada, H. Kojima, and J. Yazaki. 1984. Variations in 15 N abundances in leguminous plants and nodule fractions. Plant and Cell Physiology. 25(8): 1561 - 1565. Yoneyama, T., K. Fujita, T. Yoshida, T. Matsumoto, I. Kambayashi, and J. Yazaki. 1986. Variation in natural abundances of 15 N among plant parts and 15 N/ 14 N during N 2 fixation in the legume - rhizobia symbiotic system. Plant Cell Physiology. 27(5): 791 - 799. 124 CHAPTER 4 QTL ANALYSIS OF SYMBIOTI C NITROGEN FIXATION IN THE PUEBLA 152/ZORRO DRY BEAN RIL POPULATION Abstract Dry bean ( Phaseolus vulgaris L) is able, through symbiotic N fixation (SNF) to acqui re N from the atmosphere; but dry bean is generally considered a poor N - fixer. Considerable div ersity within dry bean germplasm has been identified and several studies have shown that SNF can be enhanced through selection. More recently quantitative trait locus (QTL) analysis and genome wide association studies (GWAS) have been used to identif y reg ions of the genome associated with SNF traits. In the current study a mapping population of 122 recombinant inbred lines was genotyped with single - nucleotide polym orphism (SNP) markers developed through the BeanCAP, to construct a genetic map spanning 972 cM and containing 430 SNPs. The population was grown in the field in East Lansing, MI (EL) and Isabela, Puerto Rico (PR) and in the greenho use (GH) under N free co nditions to evaluate for yield, nodule development, biomass growth, agronomic traits, and N fixation. A total of 19 QTL associated with SNF traits were identified on all 11 chromosomes except Pv02 and large clusters of QTL were discovered on Pv01, Pv06, an d Pv08. Many of the QTL associated with %Ndfa, N harvest index, and %N in biomass were also associated with candidate genes expressed in the nodules and roots. Candidate genes such as Phvul.006G146400, which is a chitin elicitor receptor kinase that is in volved in recognition of rhizobia in the early establishment of the symbiotic relationship. Other candidate genes are transcription factors, such as Phvul.006G034400 that is associated with 15 N, is a MADS - box family gene and is expressed in young and mat ure green pods. The majority of QTL associated with genes expressed in the root or nodule are derived from Puebla 152 while QTL 125 associated with genes with enhanced expression in stems and pods are associated with Zorro. This follows a pattern where Pue bl a 152 has superior SNF , whereas Zorro is highly efficient in partitioning the fixed - N into the seed. The QTL described serve as potential targets for improvement of SNF characteristics in adapted commercial dry bean genotypes. Introduction Symbiotic nitro gen fixation (SNF) is a complex trait. Not only must the plant be able to form compatible symbioses with the appropriate rhizobacteria, but it must also form sufficient nodule mass and effectively move fixed N through the plant to the seeds. Nodule numbe r has been shown to vary among dry bean genotypes (Pereira et al., 1993). There was a considerable significant correlation ( r 2 =0.64, p<0.01) between nodule number and N fixed in a population of dry beans bred for enhanced N fixation (Pereira et al., 1993) . C onsiderable variation in this trait exists within dry bean germplasm (Wolyn et al., 1991; Pereira et al., 1993; Fageria et al., 2014). Improvement should be possible as dry bean appears to be responsive to selection for improved SNF by selecting direct ly or indirectly for fixed N (Wolyn et al., 1991; Elizondo Barron et al., 1999). St. Clair and Bliss (1991) selected four inbred backcross lines from a Puebla 152/Sanilac population which showed superior acetylene reduction assay (ARA) levels. These pla nts were intercrossed and the F 3 progeny were tested for their ability to fix N. The majority of the 25 resultant progenies were superior N fixers when compared to Sanilac. Several of the lines studied fixed N similar to high N - fixing parent Puebla 152 w hile having agronomic traits similar to Sanilac which would make them more amenable to direct harvest (St. Clair and Bliss, 1991). 126 In addition to the ability to fix N, the efficient use of N is important. Fageria et al. (2013) noted variability among the 20 dry bean genotypes for nitrogen use efficiency (NUE). Values ranged from 7.3 mg mg - 1 seed in genotype BRS Valente for each mg N applied to 21.2 mg mg - 1 for line CNFP 7624. However, Fageria et al. (2013) did not mention if the potting mix used was ste rilized nor if the plants were nodulated, as their N - free treatment yielded nearly as much seed N (43.6 g kg - 1 for the zero N treatment compared to 46.9 g kg - 1 in the fertilized treatment) though none was intentionally added. The source of this N is fixat ion. Thus, traits associated with partitioning likely interact with SNF to achieve enhanced yield. Several studies have been conducted to map QTL for various SNF traits (Nodari et al., 1993; Tsai et al., 1998). In an attempt to look for QTL involved in the interaction between host and bacteria, Nodari et al. (1993) used an F 3 population from BAT93 (Mesoamerican derived genotype with fewer nodules and resistance to common bean blight (CBB)) and Jalo EEP558 (Andean selection with high nodule number, s usceptible to CBB) . Four QTL , which explained a total of 52% of the phenotypic variation for nodule number , were discovered. One locus appeared to have an effect on both nodule number CBB resistance which is not surprising since many stages in the interac tion with pathogenic bacteria are similar to interactions with beneficial bacteria. This region, on Pv07 contributed by the BAT93 parent, was associated with CBB resistance but with low nodule number (Nodari et al., 1993). Tsai et al . (1998) used a simil ar population by crossing the high nodulating dry bean, Jalo EEP558 with the low nodulating BAT93 to inv estigate nodule number and CBB resistance inheritance under contrasting N conditions. Both parents contributed positive alleles to nodule number and CB B resistance in the F 2 derived F 3 RILs. Given that the low nodulating parent (BAT93) contributed alleles with a positive effect on nodule number and the CBB susceptible parent similarly contributed positive 127 alleles for CBB resistance. Ramaekers et al. (20 13) used 85 RILs developed from G2333 x G19839 to investigate characteristics associated with SNF in both the greenhouse under N free conditions and in the field. A total of 204 markers, including SSRs, SNPs, and an isozyme markers were used in the geneti c analysis. They measured traits such as leaf chlorophyll content, shoot dry weight, total biomass N, seed yield and total N in seed. Many QTL were discovered for SNF traits, such as SPAD (a measure of chlorophyll, and hence N level in leaves) at differ ent growth stages (R 2 = 11.49 % to 35.53 % ), %N in the shoot, root, and plant (R 2 = 16.3 % to 21.01 % ), and total N in the shoot, root, and plant (R 2 = 14.69 % to 20.87 % ), in the greenhouse . QTL for nodule number (R 2 = 17.25 % and 16.72 % ), two QTL for nodule dry wei ght (R 2 = 12.97 % and 19.07 % ), and one QTL for %Ndfa at harvest ( R 2 =18.79 % ) were detected in the field. D ifferent QTL were found between the field and greenhouse experiment but overlapped QTL for SPAD QTL on Pv01 and two QTL on Pv07 for SPAD at pod filling were detected in the greenhouse and the field. The QTL reported have low to moderate effect on the phenotype but could prove useful in developing markers for marker assisted selection ( MAS ) . More recently a genome wide association study (GWAS) found sev eral QTL associated with SNF traits in the Andean Diversity panel of common bean (Kamfwa et al., 2015a). The panel was grown under N free conditions in the greenhouse as well as under low N in the field. QTL for %Ndfa in the shoot in field studies were f ound on Pv02, Pv03, Pv07, Pv09, Pv10, and Pv11 with an R 2 ranging from 0.11 to 0.22. Chromosomes Pv02, Pv03, Pv07 and Pv09 contained QTL responsible for Ndfa in the greenhouse with R 2 ranging from 0.09 to 0.20 (Kamfwa et al., 2015a). One SNP, ss715648916 on Pv09 , was associated with multiple QTL for SNF including Ndfa in the seed, Ndfa in the shoot, %Ndfa in the shoot, %N in the seed, chlorophyll content, shoot biomass, and %N in shoot biomass. The candidate gene associated with this SNP, 128 Phvul.009G13620 0, is a leucine - rich repeat receptor like kinase (LRR - RLK) and may prove useful as a target for enhancing SNF in dry bean. It is clear that variation exists in SNF characteristics in dry bean. Many studies have identified regions in the genome that are associated with SNF in Andean populations or in Andean x Middle American populations. The objective of the current study was to investigate the genetic components of SNF in a Middle American black bean RIL population and develop genetic markers for use in MAS to develop genotypes with enhanced SNF . Germplasm with superior SNF characteristics that is adapted to modern conventional and organic agricultural practice s is currently not available. Lines which combine improved agronomic traits and superior SNF ability developed in this study will prove useful in increasing %Ndfa in commercial dry bean breeding materials. Materials and Methods The phenotyping of the plant material, experimental design and data collection used in the QTL study were previously desc ribed in Chapter 3. Single Nucleotide Polymorphism (SNP) Genotyping The Puebla 152/Zorro population was genotyped using the SNP array developed by the BeanCAP (www.beancap.org) project. Analysis was conducted at the Soybean Genomics and Improvement USDA Laboratory (USDA ARS, Beltsville, MD, Agricultural Research Center) following Hyten et al., (2010). The Illumina platform was used following the Infinium HD Assay Ultra Protocol (Illumina Inc.). The Infinium II assay protocol involves making and incubation of amplified DNA, fragmenting the amplified DNA for preparation of the bead chip, and hybridizing the samples to the BARCBean6K_3 BeadChip with 5389 SNPs. The DNA is 129 then extended, stained , and imaged. GenomeStudio Genotyping Module v1.8.4 (Illumina, Inc .) was used to call SNP alleles. Manual adjustments were then made. Linkage Map Construction Data was filtered to remove markers with no calls. Also, SNPs with more than 20% missing data and non - informative markers were removed. The remaining 1,116 SNP ma rkers were used for map construction using Joinmap 4 (Van Ooiijen, 2006). Prior to mapping, markers were sorted into their respective linkage groups according to the reference genome (Schmutz et al., 2014). Maximum likelihood was used with a chain length minimum of 20,000 simulations each cycle for a total of 4 cycles with a 5000 simulation burn in. Markers which were 100% identical were eliminated. Nearest neighbor stress and fitness tests were inspected to evaluate convergence with likely positions. Markers with elevated stress values were eliminated. Marker order on each linkage group was verified using the known locations of markers by comparing the completed map to the physical positions in the reference genome (Schmutz et al., 2014) using the f ixed orders option in Joinmap 4 to orient the linkage groups. QTL Analysis Multiple QTL mapping (MQM) was conducted using MapQTL 5 (Van Ooijen, 2004). The LOD threshold was determined by running a permutation analysis set to 10,000 permutations for each trait. The 95 th percentile of permutations for all traits for the genome wide group was selected l at Phytozome 10.2 ( Goodstein et al., 2012) was utilized to identify candidate genes in the Phaseolus vulgar is genome. Linkage map figures were generated with MapChart 2.3 (Voorrips, 2002). The multiple interval mapping (MIM) option in WinQTLCART 2.5 (Wang et al., 2012) was used to test for QTL x QTL interactions. 130 Results and Discussion The resulting genetic map retained 430 SNP markers with a genome size of 972 cM (Figure 4.01). The map approaches the size previously determined by Freyre et al. (1998), which was 1200cM. Hoyos - Villegas et al. (2015) estimated the genome size of the AP630 pinto bean map to b e 1499 cM using SNP markers . The genome size of the G2333 x G19839 map was estimated at 1,601 cM using SSRs, SNPs, and an isozyme markers (Ramaekers et al . , 2013) . The average coverage was one marker for each 2.26 cM which was intermediate to values report ed in the AP630 population, with an average distance of 3.6 cM between SNPs (Hoyos - Villegas et al., 2015) and the SEA5/CAL9 population which had an average distance of 0.64 cM between SNPs (Mukeshimana et al., 2014). Chromosomes Pv03 and Pv10 had the lowes t number of markers while Pv05 had a disproportionately large number of markers (Supplemental Figure 4.01). The markers remaining prior to map construction were similarly distributed leaving limited coverage of some chromosomes. A QTL for canopy height , HT1.1, w as detected on Pv01 in 2012 and one in 2013 in EL (Table 4.01 ). While the peak LOD was found at two distinct positions, 49 . 6 Mb and 48 . 5 Mb , respectively, they likely refer to the same QTL. A QTL for days to flower , DF1.2, was discovered on Pv0 1 located near position 48.5 Mb in both years in EL. Mukeshimana et al. (2014) , Hoyos - Villegas et al. (2015) and Kamfwa et al. ( 2015b ) also found a similar QTL on Pv01 for days to flower located at position 43 . 7 Mb . A total of six QTL for yield were disc overed, two on Pv01, two on Pv03, and two on Pv11. The same yield QTL , SY1.1, on Pv01 was found in 2012 and 2013 in EL and accounted for 13.0% and 17.1% of the variability, respectively (Table 4.01). The QTL was located at 48 .5 Mb . A QTL for yield , SY3. 3, on Pv03 were also discovered in 2012 and 2013 in P R (R2=14.6 % and 12.1 % , respectively) at positions 131 39 .4 Mb and 39 .6 Mb and these were attributed to Puebla 152 paren t. Hoyos - Villegas et al. (2015), Mukeshimana et al. (2014), and Kamfwa et al. (2015b) f ound the same QTL for yield in this region in very different genetic populations evaluated in diverse locations . These two loci were located 0. 1 6 Mb apart and likely refer to the same QTL. A QTL for yield was also discovered by Hoyos - Villegas et al. (201 5) on Pv03 located at 33 .6 Mb with an R 2 of 12.2 % . Kamfwa et al. (2015b) also identified a QTL for yield on Pv03 in the Andean diversity panel located at position 38 .3 Mb which is intermediate in position from that discovered in the current study and that reported by Hoyos - Villegas et al . (2015). Mukeshimana et al. (2014) identified a QTL for yield in the SEA5/CAL96 population grown in Karama, Rwanda, located at position 4.0 Mb . Two QTL for yield were also discovered on Pv11 in 2011 and 2012 in EL (Table 4.01). These QTL were located 431 kb apart at 47 .6 Mb and 48 . 0 Mb , respectively and likely refer to the same SY11.1 QTL. Analysis for traits measured in the greenhouse was conducted on the bulk of 4 reps and are presented in Table 4.01. QTL for shoot N, shoot and root weight, total biomass, nodule rating, shoot difference, total biomass difference, and shoot:root ratio were found on Pv01, Pv05, Pv08, and Pv11. Overlapping QTL for shoot N (R 2 =22.5 % ), shoot weight (R 2 =13.7 % ), and total biomass (R 2 =12.7 % ) w ere detected on Pv01 between 42 .2 M b and 51 .3 M b and all three QTL were contributed by Puebla 152. These traits are related to each other, since total biomass includes shoot biomass, with both depending on N availability so it is not unexpected to have ad jacent QTL for each trait. Kamfwa et al. (2015a) also identified a QTL on Pv01 at 48 .1M b for shoot biomass in an Andean bean panel in a similar greenhouse study with an R 2 of 8.0 % . This QTL is within 0. 35 Mb of the QTL discovered in this study. A QTL , S RR5.1 for shoot:root ratio was located on Pv05 at 39 .0 M b and was contributed by the Zorro parent. Zorro is the adapted, 132 efficient parent which partitions biomass into seed much more efficiently than Puebla 152. A single QTL, RWT8.1 for root weight, was fo und in the greenhouse assay on Pv08 at 9 .6 M b. Five QTL were found on Pv11, and two for shoot difference (R 2 =46.2 % ) and for total biomass difference (R 2 =40.9 % ) colocalized at 4 .5 M b. These QTL explain a high proportion of the variation for these traits and may be useful in investigating harvest index. A QTL for shoot weight ( SWT11.1, R 2 =13.7 % ) and another QTL for shoot:root ratio ( SRR11.1, R 2 =17.3 % ) colocalized at 5 .1 M b. A single QTL , NoR11.1 for nodule rating (R 2 =11.9 % ) was located at 8 .2 M b on Pv11 . Five QTL for SNF t raits were found on Pv01 including 15 N (D15N1 .1) , %Ndfa difference (%Ndfa1.1) , N harvest index (NHI1.1) , and two for seed N (SN1.1) which were discovered in 15 N was located at 48 .5 M b with R 2 =13.6 % . One of six QTL for %Ndfa difference w as located at the same position. In 2012 a QTL for N harvest index was found at the same position and is attributed to Zorro and accounts for 11.7% of the variation. Two QTL for seed N were discovered on Pv01, one each year in 2012 and 2013. Kamfwa et a l. (2015a) found a QTL for biomass at 48.1 Mb. Ramaekers et al. (2013) also found a large cluster of QTL on Pv01 , but the use of different marker systems from the current study limit s the ability to make comparisons . Several of the traits associated with this QTL cluster were for %N in the shoot, % N in the plant, total N harvest, and total %Ndfa. S imilarly, two QTL for seed N and shoot N were found in this study on Pv01 (Table 4.01, Figure 4.01). A cluster of six QTL colocalized on Pv08 between posit ions 9 .5 M b and 12 .2 M b (Table 4.01). A QTL for N harvest index (NHI 8.1) was located at 9 .5 M b (R 2 =10.3 % ) and originated from Zorro. Another QTL for seed N (SN 8.1) was found at 10 .2 M b (R 2 =16.1 % ) and also originated 133 from Zorro (Table 4.01). Two QTL for seed N (SN 8.1) were also discovered, one in 2011 (position 10 .2 M b R 2 =15.1 % ) and 2013 (position 12 .2 Mb , R 2 =12.7 % ). While the peak LOD (4.3 and 3.6, respectively) occur at different positions, they are relatively close and are likely the same QTL . A QTL for %Ndfa difference ( % Ndfa8.1), R 2 = 10.6 % ) was located at 12 .2 M b on Pv08 which originates with the Zorro parent and may be explained by the superior partitioning ability of Zorro compared to Puebla 152. Four QTL were discovered on Pv11, with two each colocalizing. At position 1 .5 M b %Ndfa difference (Nd faD11.1 , R 2 =18.8 % ) was discovered, colocalizing with %N seed ( %NS11.1, R 2 =10.2 % ). Seed N and another for %Ndfa difference ( %Ndfa11.2 ) was located at 39 .8 M b. Both QTL were associated with Zorro (Table 4.01). A QTL for root weight ( RWT8.1) in the greenhouse was found on Pv08 at position 9 .5 M b and flanked by SNP ss715647419 at 9 .1 Mb and SNP ss715648550 at position 12 .2 M b in the RIL population. A similar QTL was found by Ramaekers et al. (2013) in the G2333 x G19839 population at 84.66 cM in their study, wh ich when compared to map positions in the Red Hawk/Stampede population would place it at approximately 54 .9 M b. Based on the cM positions in the Red Hawk/Stampede map is within 30 cM of the QTL discovered in the current study. Use of different markers in t his study make direct comparison difficult. Other QTL for SNF traits were found on Pv03 (N yield , NY3.1 ), Pv04 (N yield , NY4.1 ), Pv05 (N yield ( NY5.1 ) and %N seed ( %NS5.1) ), Pv07 (N yield), Pv09 (N yield (N Y9l.1) and %Ndfa difference (%Ndfa9.1 ) ), and Pv10 ( total N harvest , T N 10.1 ) in the field study in EL . Six QTL for SNF traits were discovered on Pv06 based on data from the field in EL and PR. A single QTL for %N in biomass (%NB6.1) was discovered which accounted for 12.6% of the variation. This QTL was located near the SNP ss715645785 and was located at position 26 .0 M b 134 (Table 4.01; Figure 4.01). This QTL colocalized with three other QTL, for seed N (SN6.1) , N harvest index (NHI 6.1) , and %Ndfa difference (%Ndfa 6.2) (Figure 4.01). The seed N (SN6.1) QTL was located at the same location and accounted for 13.7% of the variation. The QTL for N harvest in dex (NHI6.1) was also located at the same location and accounted for 14.1 % of the variation. The fourth QTL in this group accounted for 12.8% of the variation for %Ndfa difference. These traits are related in that they appear to be involved in how N is partitio ned within the plant. The second group of QTL on Pv06 contains two QTL, one for %Ndfa ( %Ndfa6.1 ) and one for 15 N ( D15N6.1 ) 15 N QTL was located at position 13 .1 M b and accounted for 10.2% of variation. The QTL for %Ndfa, located at 13 .2 M b, accounted for 13.7 % of variation and was contributed by Zorro (Figure 4.01). Several QTL for multiple dive rse traits such as shoot N, shoot weight (SWT1.1) , and total biomass (BM1.1) in the greenhouse and days to flower (DF1.1 and DF1.2) , canopy height (HT1.1) , lodging (LDG1.2) , seed N yield (SN1.1) and yield (SY1.1) in the field are clustered on Pv01 (Table 4.01). This region is gene rich and has several protein kinases such as Phvul.001G222600, Phvul.001G222700, as well as a nodulin transporter gene - Phvul.001G223600. Several QTL for traits including nodule rating (N o R11.1) in the greenhouse shoot weight (SWT11.1) , shoot:root ratio (SRR11.1) , in the greenhouse and t he yield QTL (SY11.1) from the field are located on Pv11 (Figure 4.02). A nodule Cysteine - rich (NCR) secreted peptide is located at position 8 .3 M b is approximately 51 .4 k b from the QTL for nodule rating. This QTL is derived from Puebla 152 and may suggest that this parent has an allele which would be useful in increasing SNF in future dry bean varieties. 135 Several of the QTL discovered are located at or very near genes transcribed in the roots or nodules and may serve as potential targets for breeding dry bean gen otypes with enha nced SNF . A phosphofructokinase gene, Phvul.005G165000 is located at the position of the QTL SRR5.1, on Pv05 for s hoot :root ratio (Phytozome 10.2). This gene is involved in metabolism and the expression profile indicates that this gene is highly expresse d in the nodules and young pods. The QTL associated with nodule rating found on Pv11 includes the gene Phvul.011G085200.1. This gene is a xyloglucan transglucosylase/hydrolase (XTH) and is highly expressed in the roots (Phytozome 10.2). This class of ge ne has been implicated in cell wall loosening during fruit growth and ripening (Munoz - Bertomeu et al., 2013). This gene may be important in growth and development of nodules on the root of dry bean. Another candidate gene, Phvul.006G146400, is a chitin elicitor receptor kinase expressed in the nodules and is found 12 kb from the SNP (ss715645785) (Phytozome 10.2) associated with %N in biomass, %Ndfa difference ( % Ndfa6.2 ), and N harvest index (NH I6.1 ) on Pv06. The NOD factors produced by rhizobia are chitin - related molecules (Eckardt, 2008) suggesting that this QTL may be involved with the initial interactions between dry bean host and N fixing rhizobia. This allele originates with Puebla 152 sugges ting that perhaps this parent has characteristics that allow it to more effectively perceive and respond to rhizobia in the soil. A MADS - box gene, Phvul.006G146600, is located 18 .5 kb from the SNP associated with these QTL and is similarly expressed in th e roots and nodules. This transcription factor could also be involved in some aspect of regulation of SNF. Other QTL were associated with genes highly expressed in pods such as the QTL D 15 N 6.1 for 15 N on Pv06 at position 13 .1 M b. The candidate gene, Ph vul.006G034400 belongs to the MADS - box family of genes which are involved in gene regulation. This gene is located at 14.0 136 Mb . A candidate gene associated with a different QTL for 15 N ( D 15 N8.1 ) on Pv08 located at position 12.3 Mb is a MYB family transcr iption factor ( Phvul.008G107000) located at 12 .3 M b and is also expressed in pods and stems and may play a role in partitioning N into seed. A total of 19 QTL associated with SNF characteristics in a dry bean RIL population were discovered on all chromos omes except Pv02. The number of QTL ranged from a single QTL per chromosome to six QTL on Pv06 . Chromosomes Pv01, Pv06, and Pv08 have clusters of SNF QTL and may serve as good targets for improving SNF in dry bean. Candidate genes associated with these QTL include transcription factors, transferases, and receptors involved in sensing rhizobacteria. The QTL discovered in this study may prove useful for developing future dry bean cultivar s with improved SNF . 137 APPENDIC ES 138 APPENDIX A CHAPTE R 4 TABLES AND FIGURES Table 4.01. Quantitative trait loci (QTL) for biomass , agronomic and S N F traits in the Puebla 152/Zorro RIL population grown in the field in EL and PR in 2011 - 2013 and in greenhouse under N free condi tions in East Lansing, MI. 139 Table 4.01 ( d ) 140 Logarithm of odds. Percent of the phenotypic variation explained by the QTL. § Positive values indicated allele contributed by Puebla 152 parent, negative values indicat e allele came from Zorro parent. Biomass difference calculated using the biomass (g) of each respective plant part with the non - nodulating R99 used as a reference. 141 Figure 4.01. Dry bean chromosomes Pv01, Pv03, Pv04, Pv05, Pv06, Pv07, Pv08, Pv09, Pv10, and Pv11 showing QTL for Symbioti c N - fixation (SNF) from the N - free greenhouse analysis and in the field in East Lansing, MI and Isabela, Puerto Rico in 2011 to 2013 . 142 Figure 4.01. ( ). 143 Figure 4.01 ( ) 144 Figure 4.01 ( ) 145 F igure 4.01 ( ) 146 Fi gure 4.01 ( ) For traits measured in the greenhouse: S HT N=Shoot N, SW T =Shoot Weight, BM =Total Biomass, S R R=Shoot:Root ratio, N o R=Nodule Rating, SD - Shoot Difference, TD=Total Biomass Difference. For traits measured in the field: % NB=Percent N in Biomass, D 15 N= 15 N , %Ndfa =%Ndfa, NY=N Yield, NHI=Nitrogen Harvest Index, % NS=%N in seed, SN=Seed N, NH I =N Harvest, PN=%N S in Seed. 147 APPENDIX B SUPPLEMENTAL FIGURES FOR CHAPTER 4 Figure S 4. 0 1 . Linkage map of the Puebla 152/Zorro populatio n. 148 Figure S 4. 0 1 ( ) 149 Figure S 4. 0 1 ( ) 150 Figure S 4. 0 1 ( ) 151 L ITERATURE CITED 152 LITERATURE CITED Eckardt, N.A. Chitin signaling in plants: Insights into the perceptions of fungal pathogens and rhizobacterial symbionts. The Plant Cell. 20:241 - 243. Elizondo Barron, J., R.J. Pasini, D.W. Davis, D.D. Struthman, and P.H. Graham. 1999. Response to selection for seed yield and nitrogen (N2) fixation in common bean (Phaseolus vulgaris L.). Field Crops Research. 62:119 - 128. Fageria, N.K., L.C. Melo, and J. de Oliveira. 2013. Nitrogen use efficiency in dry bean genotypes. Journal of Pl ant Nutrition. 36:2179 - 2190. Fageria, N.K., L.C. Melo, E.P.B Ferreira, J.P. Oliveira, and A.M. Knupp. 2014. Dry matter, grain yield, and yield components of dry bean as influenced by nitrogen fertilization and rhizobia. Communications in Soil Science and Plant Analysis. 45:111 - 125. Freyre, R., P.W. Skroch, V. Geffroy, A.F. Adam - Blondon, A. Shirmohamadali, 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 integrat ed linkage map of common bean. 4. Development of a core linkage map and alignment of RFLP maps. Theor. Appl. Genet. 97:847 856. Goodstein , D.M , S . Shu, R . Howson, R . Neupane, R . D. Hayes, J . Fazo, T . Mitros, W . Dirks, U . Hellsten, N . Putnam, and D . S. Rokhs ar, Phytozome: a comparative platform for green plant genomics , Nucleic Acids Res. 2012 40 (D1): D1178 - D1186 Found at http://phytozome.jgi.doe.gov/pz/portal.html# Hoyos - Villegas, V., W. Mkw aila, P.B. Cregan, and J.D. Kelly. 2015. Quantitative trait loci analysis of white mold avoidance in common bean. Crop Science. 55:1 - 14. Hyten, D.L., Q. Song, E.W. Fickus, C.V. Quigley, J.S. Lim, I.Y. Choi, E.Y. Hwang, M. Pastor - Corrales, P.B. Cregan. High - throughput SNP discovery and assay development in common bean. 2010. BMC Genomics. 11:475. Kamfwa, K., K.A. Cichy, and J.D. Kelly. 2015a. Genome - wide association analysis of symbiotic nitrogen fixation in common bean. Theoretical Applied Genetics. DOI 10.1007/s00122 - 015 - 2562 - 5 Kamfwa, K., K.A. Cichy, and J.D. Kelly. 2015b. Genome - wide association study of agronomic traits in common bean. Plant Genome. 8(2):1 - 12. Mukeshimana, G., L. Butare, P.B. Cregan, M.W. Blair, and J.D. Kelly. 2012. Quanti tative trait loci associated with drought tolerance in common bean. Crop Science. 54:923 - 938. Munoz - Bertomeu, J., E. Miedes, and E.P. Lorences. 2013. Expression of xyloglucan endotransglucosylase/hydrolase (XTH) genes and XED activity in ethylene treat ed apple and tomato fruits. Journal of Plant Physiology. 170(13):1194 - 1201. 153 Nakedde, T., 2015. and mapping of QTL associated with fusarium root rot resistance and root architecture traits in black beans ( P. vulgaris ) . MS Thesis. East Lansing, MI: Mic higan State University Nodari, R.O., S.M. Tsai, P. Guzman, R.L. Gilbertson, and P. Gepts. 1993. Toward an integrated linkage map of common bean. III. Mapping genetic factors controlling host - bacteria interactions. Genetics. 134:341 - 350. Pereira, P.A .A., B.D. Miranda, J.R. Attewell, K.A. Kmiecik, and F.A. Bliss. 1993. Selection for increased nodule number in common bean ( Phaseolus vulgaris L.). Plant and Soil. 148:203 - 209. Ramaekers, L., C.H. Galeano, N. Garzon, J. Vanderleyden, and M.W. Blair. 2 013. Identification of quantitative trait loci for symbiotic nitrogen fixation capacity and related traits in common bean. Molecular Breeding. 31:163 - 180. Schmutz, J., P.E. McClean, S. Mamidi, G.A. Wu, S.B. Cannon, J. Grimwood, J. Jenkins, S. Shu, Q. Song, C. Chavarro, M. Torres - Torres, V. Geffroy, S.M. Moghaddam, D. Gao, B. Abernathy, K. Barry, M. Blair, M.A. Brick, B. Chovatia, P. Gepts, D.M. Goodstein, M. Gonzales, U. Hellsten, D.L. Hyten, G. Jia, J.D. Kelly, D. Kudrna, R. Lee, M.M.S. Richard, P.N. Miklas, J.M. Osorno, J. Rodrigues, V. Thareau, C.A. Urrea, M. Wang, Y. Yu, M. Zhang, R.A. Wing, P.B. Cregan, D.S. Rokhsar, and S.A. Jackson. 2014. A reference genome for common bean and genome - wide analysis of dual domestications. Nature Genetics. 46(7) :707 - 713. St. Clair, D.A. and F.A. Bliss. 1991. Intrapopulation decombination for 15 N - determined dinitrogen fixation ability in common bean. Plant Breeding. 106:215 - 225. Tsai, S.M. R.O. Nodari, D.H. Moon, L.E.A. Camargo, R. Vencovsky, and P.Gepts. 1 998. QTL mapping for nodule number and common bacterial blight in Phaseolus vulgaris L. Plant and Soil. 204:135 - 145. Voorrips, R.E., 2002. MapChart: Software for the graphical presentation of linkage maps and QTLs. The Journal of Heredity 93 (1): 77 - 78. Van Ooijen, J.W. 2006. Join Ma p 4. Software for the calculation of genetic linkage maps in experimental populations. Kya zma B.V., Wageningen, the Netherlands. Voorrips, R.E., 20 02. MapChart: Software for the graphical presentation of linkage maps and QTLs. The Journal of Heredity 93 (1): 77 - 78. Wang, S., C.J. Basten, and Z.B. Zeng. 2012. Windows QTL Cartographer 2.5. Department of Statistics, North Carolina State University, Ra leigh, NC. Wolyn, D.J., D.A. St Clair, J. DuBois, J.C. Rosas, R.H. Burris, and F.A. Bliss. 1991. Distribution of nitrogen in common bean (Phaseolus vulgaris L.) genotypes selected for differences in nitrogen fixation ability. Plant and Soil. 138:303 - 31 1.