Prevalence and Host Resistance to Common Bean Rust Disease in Western and Central Kenya

,


Introduction
Te common bean (Phaseolus vulgaris L.) is an important and versatile component of food, nutrition, and economic systems for rural and urban populations around the world [1]. Beans are rich in proteins, vitamins A, B 6 , C, K, folic acid, and essential minerals such as calcium, potassium, iron, manganese, copper, and phosphorus [2]. Such nutrients are useful in complementing carbohydrate-rich foods such as cereals, tubers, and root crops. According to FAOSTAT [3], the global production of dry and green beans in 2019 was 28.9 million tons and 26.9 million tons, respectively. Te per capita consumption of common beans in Kenya is relatively high, at approximately 14 kg-66 kg per year [4,5]. Despite the importance of common beans as a pulse and vegetable crop, relatively low yields have been reported across years, and this can be explained by abiotic and biotic stresses, e.g., pests and disease [6].
Rust, caused by Uromyces appendiculatus, is among the major diseases decimating common bean felds wherever it occurs [7,8]. Te pathogen has high virulence variability and is distributed throughout the globe, constraining common bean production in humid subtropical and tropical regions and creating intermittent severe epidemics in moist temperate areas [8]. Te occurrence of common bean rust is infuenced by factors such as altitude, agronomic practices, temperature, relative humidity, leaf surface moisture, and host factors [9][10][11][12]. Bean rust disease frst appears on the upper and lower leaf surfaces as circular chlorotic or white spots that form reddish-brown pustules and yellow tissue surrounding single large or small groups of uredia [13]. Under these favourable conditions, the pathogen causes premature leaf yellowing, senescence, and total leaf fall resulting in 65-100% yield losses in common beans [14,15]. Most farmers mainly rely on the use of chemical and cultural methods, which are expensive for many small-holder farmers [16,17]. Host plant resistance is, therefore, considered a sustainable method of managing the disease.
Incorporation of rust disease resistance genes into common bean cultivars grown in Kenya was achieved more than two decades ago and resulted in the release of resistant cultivars under the Grain Legume Project (GLP). However, due to the broad pathogenic variability of the rust fungus [18] and lack of focus on the pathogen in current breeding programs, informal reports have shown that the disease is slowly re-emerging resulting in losses among small-holder farmers in Kenya. A study by Odogwu et al. [19] reported high rust incidence and severity in the neighbouring country, Uganda. Terefore, there is a need for monitoring the changing virulence patterns in Kenya to prevent potential epidemics in the future and for the deployment of durable bean rust resistance genes. In addition, periodic collection and characterization of bean rust is essential, as it informs on virulence diversity, the dynamics of epidemics, and the development of common bean cultivars.
Resistance against bean rust disease is mainly conditioned by 14 major dominant genes, which are derived from the two common bean gene pools (Andean and Mesoamerican) [8,13]. Te gene pools refer to the domestication centres of wild beans identifed by using the phaseolin seed protein [20], diferent allozymes [21][22][23], and various classes of molecular markers [24][25][26]. Tese markers are still useful in understanding the common bean germplasm; for example, the phaseolin protein molecular marker was utilized by Arunga and Odikara [24] to designate Kenyan French beans into the two common bean gene pools. Furthermore, various DNA assays (random amplifed polymorphic DNAs and sequence characterized amplifed regions) linked to the major rust resistance genes have been developed and utilized in the identifcation of resistance genes and for markerassisted selection [8]. Te classifcation of the common bean germplasm is important in rust resistance breeding because combining genes from both gene pools is a major strategy in integrated management of the disease [27,28]. Te objectives of this study were, therefore, to determine the distribution and factors infuencing the occurrence of bean rust in Kenya and to identify cultivars with genes that confer resistance to this fungus.

Materials and Methods
Te study entailed a feld survey to assess the prevalence, severity, and factors infuencing the occurrence of rust in cultivated common beans in fve counties in Kenya. Secondly, feld and greenhouse trials were carried out to evaluate a panel of common bean genotypes for resistance to rust.  [29] that formed the study units, and each feld was visited once. Purposive and simple random sampling based on intensity of bean production, crop stage, and spatial and ecological location was used, targeting 30 felds in each county. Fields with bean plants at fowering to pod-formation growth stages were selected randomly at intervals of fve to 10 km along the main roads. Following the methodology of Odogwu et al. [30], the size of each sampled feld was estimated and the crop stage established. Equidistant steps following an inverted "V" outline were made at the edge of the feld from which the sample plants were selected. At each predetermined pace, the plant nearest to the right foot was taken as the sample unit. Assessment of disease was done on 20 plants of the same cultivar sampled within each feld. Evaluations were done on a cultivar found in a sample feld. Whenever necessary, the number of randomly selected single plants per feld was adjusted to suit the feld size and crop distribution.

Data Collection and Analysis.
Bean rust incidence was recorded from 20 sampled plants of the same cultivar within the sample feld. Rust disease severity was rated using a modifed CIAT 1-9 scale, adopted from Van Schoonhoven and Pastor-Corrales [31]. Tis scale considers nine infection types, where 1 � no visible pustule, 2 � pustules covering 1% of leaf area, 3 � few pustules covering 2% of leaf area, 4 � intermediate pustules covering 5% of leaf area, 5 � small pustules covering 8% of leaf area, 6 � pustules covering 10% of leaf area, often surrounded with chlorotic halos, 7 � large pustules covering 15% of leaf area, surrounded with chlorotic halos, 8 � large pustules covering 20% of leaf area surrounded with chlorotic halos, and 9 � very large pustules covering more than 25% of leaf area, often with defoliation.
A disease score of 1-3 was regarded as resistant, 4-6 as intermediate, and 7-9 as susceptible. Te global positioning system (GPS) readings of latitude, longitude, and altitude were recorded for each feld using a GPS map camera lite application (version 1.0.7). In addition, information regarding factors afecting disease prevalence was recorded in a feld book based on the farmers' responses. Tese factors included the cropping system (intercrop or sole crop), common bean cultivar under production, seed source (farmer-saved seeds, local market, or certifed seed from merchants), previous crop planted, and other cultural practices (fungicide use, crop debris management, crop spacing, and management of volunteer plants). At harvest maturity, seeds were collected from the visited farms for the purpose of screening for resistance to rust. Infected common bean leaves were collected from each sampled feld for subsequent single-spore isolation and multiplication for further screening for rust resistance.
Te GPS survey data from each sample location on feld coordinates were used to develop the bean rust disease severity map. Rust incidence and severity data were subjected to analysis of variance using GenStat [32] Discovery Edition 14.0 statistical software. In this analysis, location (counties), cropping system, cultivar, source of seeds, debris management, previous crop, fungicide use, and management of volunteer plants were considered fxed factors. Multiple mean comparisons for rust disease incidence and severity for all felds surveyed were performed using Tukey's studentized range test at α � 0.05.

Experimental Site.
Evaluation of the resistance profles of the genotypes under feld and greenhouse conditions was conducted at the University of Embu research feld, located at a latitude of 0°30′S and a longitude of 37°27′E. Te area's mean temperature is 19°C, with a maximum of 25°C and a minimum of 10°C, and an average annual rainfall of 1,120 mm [33].

Plant Materials.
Te common bean germplasm used in this study comprised of 77 bean genotypes obtained from farmers in the surveyed counties, which represented major bean-growing areas in western and central Kenya, the Kenya Agricultural and Livestock Research Organization (KALRO) seed unit, and the French bean improvement program at the University of Embu. Te common bean genotypes consisted of 13 landraces, 20 French bean cultivars, 29 dry bean cultivars, 3 breeding lines, and 12 bean rust diferential cultivars. Codes UN1 to UN8 were used to identify the eight landraces that were unnamed. GLP X92 (susceptible to rust) and the 12 diferential cultivars/lines were used as checks because information on their resistance genes and gene pools was available [8].

Field Experimental Layout and Data Collection.
Te feld experiment was conducted from May to July 2021, during the long-rain cropping season. Te experiment was set up as a randomized complete block design with three replicates. Twenty-one seeds from each entry were sown in a 2meter-long row, with inter and intrarow spacing of 30 cm and 10 cm, respectively. A susceptible cultivar, GLP X92, was planted as a spreader row after every fve entries at a relatively high plant density to ensure increased disease pressure. Disease inoculation was based on natural infection. Bean rust disease severity was recorded using the modifed CIAT 1-9 disease rating scale adopted from Van Schoonhoven and Pastor-Corrales [31].

Screening for Resistance under Greenhouse Conditions.
Ten viable bean rust isolates obtained during the survey were purifed through single-spore isolation [8]. An individual unopened pustule including a 25 mm 2 surrounding leaf tissue for each isolate was separately cut and the spores were collected and transferred to susceptible seedlings of cultivar GLP X92. Te single-pustules were collected and multiplied on the susceptible variety for three consecutive cycles and then characterized into physiological races using a set of 12 diferential cultivars, according to Steadman, Pastor-Corrales, and Beaver [34]. Four races identifed as 29-1, 29-3, 61-1, and 63-1 and an additional set of mixed isolates was used to evaluate the response of the germplasm to bean rust. Ten seeds of each common bean germplasm panel and 12 diferential series (as standard checks) were sown on seedling trays flled with sterile soil and laid out in a randomized complete block design with three replicates. Te disease inoculum was introduced on 8-10-day-old plants with about two-thirds of the primary leaves expanded by hand spraying viable U. appendiculatus urediospores at a concentration of 2.0 × 10 4 urediospores per ml of distilled water. Inoculated plants were then transferred to a screenhouse maintained at 20 ± 1°C and a relative humidity >95% under a 12-hour light/dark regime for approximately 48 hours, after which the plants were transferred to a greenhouse at 20 ± 5°C for about 14 days.

DNA Analysis for Gene Pool Afliations.
Young leaves were collected from each of the 77 common bean genotypes, and DNA was extracted using the Mahuku DNA extraction protocol [35]. Te phaseolin protein SCAR marker was used in PCR amplifcation [35]. A 10 μl reaction volume in FrameStar ® Break-A-Way PCR tubes containing 1X Dream Taq bufer (containing 2 mM MgCl 2 ), 0.2 mM dNTPs, 0.5 μM of each reverse and forward primer, 0.1 U Taq Polymerase (Termo Fisher Scientifc), and 1.5 ηg/μl of genomic DNA were used. Te PCR procedure was as follows: an initial denaturation step at 94°C for 3 min, followed by 35 cycles of the following three steps: denaturation at 94°C for International Journal of Agronomy 10 s, 55°C annealing for 40 s, an extension at 72°C for 2 min, and a fnal extension step at 72°C for 5 min. To each PCR product, 2 μl of 6x DNA loading dye (NEB) was added. A 50 bp DNA ladder (https://www.thermofsher.com/order/ catalog/product/10416014) was loaded in the frst well; then, PCR product contents were loaded in subsequent wells on a 1.5% agarose gel prestained with 5 μm of ethidium bromide in 1x sodium borate bufer and run at 100 volts for 3 hours. Te DNA bands were then viewed under ultraviolet light (UVP ® GelDoc-it system) and scored for the presence of either two or three fragments of diferent sizes.

Results and Discussion
4.1. Prevalence, Incidence, and Severity of Bean Rust. Bean rust disease was observed across the fve surveyed counties, with varying degrees of incidence and severity. Rust severity scores ranged from 1 to 9, with an incidence between 0 and 100%. Te mean rust severity map revealed the distribution of rust across the surveyed counties ( Figure 1). Incidence and severity of bean rust varied signifcantly (P < 0.001) among counties (Table 1) and the altitude ( Table 2) of the regions. Te overall mean rust incidence for the counties surveyed was 55.2%, with an overall mean severity of 3.03. Te mean rust recorded was as follows: Bungoma (prevalence 100%; incidence 70.8%; severity 3.99), Uasin Gishu (prevalence 96.7%; incidence 61.20%; severity 3.12), Kakamega (prevalence 100%; incidence 57.30%; severity 3.00), Kirinyaga (prevalence 93.3%; incidence 48.3%; severity 2.69), and Embu (prevalence 83.3%; incidence 38.3%; severity 2.34). In this study, the incidence and severity of common bean rust varied by location, depending on environmental conditions and crop husbandry practices. Te high incidence and severity of bean rust may be attributed to the agronomic practices adopted in the studied production areas among smallholder farmers. For instance, due to the use of susceptible cultivars and poor bean residue management, the bean rust incidence and severity were high in some individual felds studied in Bungoma, Kakamega, and Uasin Gishu counties, which could be explained by specifc cultural activities compounded by high relative humidity due to high rainfall received in the counties in 2020 [36]. Lower bean rust incidence and severity were recorded in low altitude areas of <1,200 m above sea level, especially in lower parts of Embu County that occasionally receive low rainfall and high temperatures, which do not favor the occurrence of bean rust disease. Areas with altitudes of more than 1,200 m above sea level had a high bean rust incidence and severity. Tis may be attributed to high rainfall and relative humidity that favors infection and development of bean rust disease [12].

Efects of Cultural Practices on Bean Rust Prevalence and
Severity. Some common bean production practices significantly infuenced the incidence and severity of bean rust in the surveyed regions (Tables 3 and 2). Tis inference is consistent with the previous fndings that show that the environment is a major factor afecting the distribution of biotic stressors on pulse crops [37]. Production of common beans as a sole crop or intercrop did not infuence disease incidence or severity. Similarly, the source of seeds used for planting and previous crops grown had no signifcant infuence on the incidence and severity of bean rust in the surveyed counties. Te insignifcant infuence of the cropping system, source of planting material, and previous crop grown on prevalence of bean rust may be explained by the fact that bean rust could be infuenced by the interaction of a set of factors such as ideal environmental conditions, host plant susceptibility, and high virulence of the pathogen.
Fungicide use signifcantly (P < 0.01) afected the incidence and severity of bean rust, with reduced disease in felds sprayed with fungicides such as Dithane M45® (Mancozeb) and Funguran® (copper hydroxide-770 g/kg). However, the occurrence of rust in some felds in the surveyed counties in the central region despite fungicide treatment suggests inefective application of fungicides or possibly that the pathogen in those areas has developed resistance to the fungicides being used. Tis fnding emphasizes the need to evaluate the efectiveness of the available fungicides for efcacy and informed use of fungicides in the management of bean rust among smallholder farmers.
Incidences and severity of bean rust were cultivardependent, with the most susceptible cultivars being Kisii, Sungura, GLP-24 (Canadian Wonder), and Kablanketi, whereas the most resistant cultivars were Vanilla, Embean 14, and KAT B11. Limited cultivar selection among common bean farmers in Kenya, resulting in the use of cultivars susceptible to rust, contributed to high-rust incidence and severity in the surveyed counties. Te signifcant infuence of common bean cultivars under production on the occurrence and severity of bean rust observed in felds cultivated with landraces and commercial cultivars is due to their inherent genetic structure. Common bean cultivars have been reported to have a range of resistance to bean rust disease depending on their genetic composition under feld conditions [19,38].
Strategies used in the management of common bean residues, management of volunteer plants, and crop spacing had signifcant efects on mean rust incidence and severity (P < 0.05). Te bean rust pathogen cannot survive without its common bean host, being an obligate parasite [28], and this could explain the signifcant infuence of diferent strategies  International Journal of Agronomy used by farmers in managing volunteer plants and bean debris on bean rust incidence and severity. Bean plant debris may bear viable rust spores, and this infuences the occurrence and severity of bean rust. Using bean debris in making trash-lines, preparing compost manure, and leaving it on the soil surface signifcantly contributed to the high incidence and severity of bean rust in farmers' felds compared to those who reported to practice soil incorporation and had signifcantly lower rust. Tese fndings agree with the recommendation for the elimination of bean residue through strategies such as soil incorporation to aid in the management of bean rust disease. High mean rust incidence  Means followed by the same letter within a column are not signifcantly diferent from one another (P < 0.05). a, b, and c following the values are supposed to show the diferences in the means.
International Journal of Agronomy and severity were observed at close spacing, possibly due to increased relative humidity and enhanced pathogen inoculum spread, which could favor bean rust development.

Bean Rust Races.
Four bean rust races 29-1, 29-3, 61-1, and 63-1 were obtained from single spores (Table 4). Races 29-1, 29-3, and 61-1 were previously reported in Kenya [18,39], and this highlights their predominance and importance in genotype screening for resistance in breeding programs. Te common bean rust race 63-1 identifed in this study has not been previously documented in Kenya, and this points out the necessity for comprehensive collection and characterization of bean rust isolates into physiological races using sequence analysis and BLAST technology.

Profles of Common Bean Resistance to Rust Based on Field and Greenhouse Evaluations.
Field and greenhouse screening of the common bean germplasm in Kenya revealed high variability in response to rust (Table 5). Te low disease pressure under feld conditions could be attributed to a low initial inoculum, high chances of disease escapes, and unfavourable environmental conditions in the feld [40]. Tis variability in host resistance to diferent races of bean rust indicates the possibility of resistance genes inherent in the genotypes. Genotypes such as MU#13, UN2-Darkgreen, UN6-Nakholo, Kat X56, and KMR-11 (Angaza) exhibited high resistance under feld and greenhouse conditions and are therefore potential parental genotypes in common bean breeding for the region. According to Wagara and Kimani [41], genotype variability in response to bean rust can be exploited in common bean improvement programmes as sources of resistance. As similarly reported by Arunga [42] and Kamiri et al. [43], MU#13, a local French bean breeding line, is resistant to a number of bean rust races and anthracnose. Tis genotype can be exploited in French bean improvement for disease resistance to counter local races. However, there is a need for the characterization of these resistance sources and the development of high-throughput molecular markers to aid in marker assisted breeding for rust resistance.
A consistent reaction to bean rust was observed among the diferential cultivars harbouring Ur-3, Ur-3+, Ur-5, Ur-11, Ur-14, and Ur-CNC resistance genes under feld conditions and to races 29-1, 29-3, 61-1, and 61-3 as well as mixed isolates. Tis emphasizes their importance in breeding for resistance to bean rust in Kenya. Most genotypes exhibited a susceptible reaction to rust, and this may be attributed to the broad pathogenic variability of U. appendiculatus, as similarly reported by Hillocks et al. [44]. Terefore, there is a need for pyramiding resistance genes into common bean germplasm as a breeding strategy in bean rust disease management. Additionally, multiyear/ multiseason evaluation for bean rust resistance across different altitudinal ranges in central and western Kenya would be necessary for targeted deployment of resistance genes.
A sample with a profle of two fragments of 249 bp and 270 bp was considered as belonging to the Mesoamerican gene pool, and a sample with three fragments of 249 bp, 264 bp, and 285 bp was considered to belong to the Andean gene pool [45]. Based on this, 37 genotypes were classifed as Mesoamerican, whereas 40 genotypes belonged to the Andean gene pool. Common bean genotypes of Andean origin such as Enclave, Kat X56, Kablanketi, and KMR 11 (Angaza) and Mesoamerican genotypes such as MU#13, Manakelly, and UN6-Nakholo were resistant to all the races evaluated. However, the Andean genotypes Hawaii, Julia, Amy, Samantha, and UN3-Yellow small were susceptible to all the Andean bean rust races. Paralleled common bean host reactions relative to the bean rust pathogen suggest hostpathogen coevolution, which explains the occurrence of Uromyces appendiculatus as a biotroph comprising diferent pathotypes. Generally, genotypes of the Mesoamerican gene pool exhibited high resistance to bean rust compared to those of the Andean gene pool, supporting probable pathogen coevolution with the common bean host. Furthermore, the Andean genotypes as well as some Mesoamerican genotypes were susceptible to the Andean races used in this study, complementing the fndings by Acevedo et al. [38]. High resistance among the Mesoamerican genotypes   International Journal of Agronomy

Conclusions
Bean rust is prevalent in central and western Kenya. Te choice of resistant cultivars for production, the management of crop residue, and the use of fungicides can desirably be used in managing bean rust diseases. Farmers need to be informed on the appropriate cultural practices to employ in reducing the incidence and severity of common bean rust. Te use of resistant cultivars can be utilized in managing bean rust instead of fungicides, which are expensive and potentially hazardous to the environment. Cultivars such as Kat X56, Enclave, and KMR-11 can be used by farmers, considering their high resistance to bean rust. Breeding for resistance can utilize local germplasm such as UN-2, UN-6, MU#13, Kat X56, and KMR-11, as well as one or more of the Mesoamerican genes such as Ur-3, Ur-3+, Ur-5, Ur-11, Ur-14, and Ur-CNC in common bean improvement.

Data Availability
Some of the data used to support the fndings of this study are included in the article. Additional data are available from the corresponding author upon request.

Disclosure
While the research reported here was funded by Kirkhouse Trust, the design, execution, and interpretation of the research remain wholly the responsibility of the authors.

Conflicts of Interest
Te authors declare that there are no conficts of interest regarding the publication of this paper.