Colletotrichum lindemuthianum (Sacc. & Magnus) Lams.-Scrib. is the causal agent of a devastating fungal disease of dry beans known as anthracnose, which occurs in almost all bean growing regions worldwide (Sharma et al., 2007). This results in a decreased seed quality and can also cause seed yield losses up to 100% if disease onset occurs at early stages (Junaid et al., 2014). The world produced 27 million MT of dry beans in 2020, worth an estimated $30 billion (US) of which dry beans in the Americas contributed 30% of the total (https://www.fao.org/faostat). United States is one of the largest dry bean producers in the world, with a total annual production of 1.6 million MT, and a production value of close to USD $1 billion in 2020 (USDA-NASS, 2020). The most prevalent race in the United States is race 73, to which most Middle American cultivars from the market classes commonly grown in the US (i.e., pinto, navy, black, great northern and small red) are susceptible (Balardin et al., 1997). However, disease response varies within each market class depending on the genetic background of each genotype. Increasing the number of resistant genotypes will decrease overall economic losses and provide higher seed quality.

Twenty-four resistance loci have been identified on eight chromosomes, Pv01, Pv02, Pv03, Pv04, Pv07, Pv08, Pv09 and Pv11, in various dry bean cultivars and germplasm (González et al., 2015; Zuiderveen et al., 2016). The mode of inheritance of most anthracnose resistance genes follows a typical dominant, single gene model. One of the most effective pyramiding of anthracnose resistance genes involves the Co-1² and Co-4² alleles, which are on Pv01 and Pv08, respectively. In addition to the Co-genes, the physiological state and the age of the bean plant may contribute to anthracnose resistance in an age-related resistance (ARR) manner. According to Develey-Rivière and Galiana (2007), the mechanisms behind ARR in plants vary, are species-specific, and can be due to specific gene regulations/expression levels in different plant tissues and development stages. Thus far, increased levels of proteins related to resistance, such as phenylalanine ammonia-lyase (PAL), chalcone synthase (CHS), and chalcone isomerase (CHI), which accumulate in the plant tissue over time, have been reported in some species (Nuss et al., 1996).

A classic example of ARR was reported in the cereal rust patho-system. Two well studied genes for ARR, classified as adult plant resistance genes, include the genes, Lr34 and Sr2 (reviewed in Ellis et al., 2014). These two cereal rust resistant genes provide partial, broad-spectrum resistance. Sr2, a resistance gene for stem rust (Puccinia graminis sp. Tritici Eriksson & Henning), provides both partial resistance as well as a boost for other R genes. Lr34 provides partial resistance against leaf rust (caused by P. triticina Eriksson), stripe rust (P. striiformis Westendorp), and powdery mildew (Blumeria graminis (de Candolle) Speer). It was cloned in 2009 (Krattinger et al, 2009) and is related to the ABC class of trans-membrane transporters. Both Sr2 and Lr34 have been extensively used with R genes in breeding programs to pyramid resistance.

ARR has been extensively reported in various plant species including dry beans. Bateman and Lumsden (1965) reported an increase of resistance to Rhizoctonia solani Kuehn (isolate R-B) in the hypocotyls of red kidney beans two weeks after emergence, and full resistance after four weeks. The conversion of pectin to calcium pectate caused increased resistance, due to the inability of R. solani (isolate R-B) to hydrolyze calcium pectate and subsequently destroy the cell wall. Beebe and Pastor-Corrales (1991), reported similar observations in seedling assays of the Andean cultivar ICA-Llanogrande, a cargamanto (large cream/mottled seed) climbing bean (growth type IV) cultivar from Colombia, and Rio Negro, a black bean cultivar from Brazil (Voysest, 2000), in which both cultivars were susceptible to anthracnose at the seedling stage, while older plants appeared to be resistant. Another study by Bigirimana and Höfte (2001) detected increased resistance to C. lindemuthianum races 385 and 401 over time in the cultivar Prelude (a snap bean cultivar), where 6 to 8-day old plants showed moderate or no resistance, but developed full resistance 12 days after emergence. The authors speculated that similar resistance mechanisms are involved as described in Mezzola et al. (1994), for resistance to bacterial blight (Xanthomonas oryzae pv. oryzae (Ishiyama) Swings, van den Mooter, Vauterin, Hoste, Gillis, Mew & Kersters) on rice (Oryza sativa L.) which was mediated by gene Xa21, and was not expressed at early development stages. In addition to Xa21 related mechanisms, Bigirimana and Höfte (2001) speculated that developmentally regulated cell death genes may have a role in ARR. However, none of those studies looked further into the plant defense mechanisms at the physiological/biochemical level.

Anthracnose resistance studies in dry beans are usually carried out using young plants (8 days; Castellanos et al., 2015) at the unifoliate leaf stage. This allows for rapid screening of high numbers of genotypes. Our objective was to identify resistance in older plants (similar to ARR), that could be used in dry bean breeding programs. To evaluate the plant defense system in older dry beans, and unveil potentially new and unutilized defense mechanisms, the plants used for this study were substantially older (approx. 14-20 days, trifoliate leaf stage) than in the original method of anthracnose inoculation by Castellanos et al. (2015). The genetic location of the associated resistance genes was identified using the Middle American Diversity Panel (MDP), and the observed resistance was further characterized by additional greenhouse studies with varying lengths of optimal disease conditions, and field testing to confirm the field applicability of the resistance.

The phenotyping of older plants (trifoliate stage) during the initial round of MDP screening focused on the identification of genomic regions contributing to ARR which could be targeted for use in pyramiding disease resistance. The standardized unifoliate screening method would not allow detection of ARR as infection in this method occurs during unifoliate leaf expansion. Previous research in dry beans indicated a major developmental shift occurs between unifoliate leaf expansion and trifoliate leaf expansion (Bateman and Lumsden, 1965; Beebe and Pastor-Corrales, 1991; Voysest, 2000; Bigirimana and Höfte, 2001) and thus was the focus of this research. Another major developmental shift occurs when plants change from a vegetative to reproductive stage which was not evaluated in this study. The use of a condensed scale to score the MDP limited the comparison of the results to previously published research. However, the differences in disease rating scales gave a more robust picture of disease reaction. The 1-3 disease severity rating considered the general appearance of the plant and produced better ratings for surviving/recovering plants after a five day or longer exposure time under favorable disease conditions.

The SNP markers associated with anthracnose ARR were located on Pv10 and on Pv01 with the primary factor near the end of Pv10. Major anthracnose resistance genes have been identified on Pv01 but not on Pv10. López et al. (2003) did mention an insignificant anthracnose resistance locus on Pv10 for race 7 but no physical location was identified. Zuiderveen et al. (2016) also identified an insignificant anthracnose resistance locus on Pv10 for race 7 near 3.78 Mb (G19833 v1.0). The identification of the ARR associated SNP, S10_72250 near the ZPR1 zinc protein finger protein is highly indicative of its involvement in anthracnose ARR. The ZPR1 protein is primarily localized in the cytoplasm but can accumulate in the nucleus when the cell is activated with mitogens. ZPR1 is an essential protein critical for normal cellular proliferation; however, its specific role is not known but it is hypothesized to function during regulation of pre-ribosomal RNA expression. ZPR1 interacts with the cytoplasmic domain of the inactive EF1A (elongation growth factor 1A) receptor and may inhibit its tyrosine kinase activity. The traditional role of EF1A is the delivery of aminoacyl-tRNA complex to ribosomes for protein elongation; however, EF1A is a multifunctional protein (Sasikumar et al., 2012; Momčilović and Fu, 2014). It is involved in the export of charged tRNA from the nucleus, arrangement of filamentous actin, as well as chaperone activity. As a chaperone it prevents thermal aggregation of malate dehydrogenase and may have been involved in salt tolerance in Arabidopsis (Shin et al., 2009). EF1A is highly expressed in young developing tissue but expression changes during plant development and in various tissues (Carneiro et al., 1999; Xu et al., 2007; Ransom-Hodgkins, 2009). EF1A expression is affected by environmental stimuli. In rice, expression increased after exposure to cold, drought, salt, or heat (Li and Chen, 1998). In barley (Hordeum vulgare L.) and maize (Zea mays L.), EF1A expression increased after exposure to cold stress (Dunn et al., 1993; Berberich et al., 1995). In soybeans, exposure to drought, cold, salt or ABA increased expression (Chung et al., 2009). In another system, various isoforms of EF1A were upregulated during the hypersensitive response (HR) after induction in grapes (Vitis vinifera Linnaeus) (Dietrich et al, 2010). The reported differences support the role of EF1A in reactions to stress and with its interaction with ZPR1, make ZPR1 a possible target for exploiting anthracnose ARR.

The other resistance regions identified in this study had a minor influence on anthracnose ARR in this study, each explaining less than 1% of the phenotypic variation. The interval containing the peak SNP, S01_48462800, encoded several proteins with a known role in disease resistance including Cytochrome P450 (reviewed by Pandian et al., 2020) and may be the same resistance loci identified by Zuiderveen et al. (2016) and fine mapped by Murube et al. (2019). Co-1 resides on Pv01 and may fall within this interval. The interval containing the peak SNP S01_60066578 included a CRINKLY4 related protein. Recently Richard et al. (2021) identified a truncated CRINKLY4 kinase as a candidate for Co-X, corroborating our results.

The predictive utility of each of the four peak significant SNPs was evaluated by comparing the means between the two allelic states for each peak SNP ( Figure 2 ). Three of the four peak SNPs, S01_48462800, S01_60066578, and S10_8708314, had significantly different means for each allelic state using a student’s t-test (0.012 0.00, and 0.00 probability respectively). A closer look at the phenotypes for each allelic state indicates SNP S01_60066578 may be a good predictor for breeding as all genotypes with an adenine at position S01_6006578 exhibited a resistant reaction. This is consistent with the expected dominant mode of inheritance for most anthracnose resistance genes reported. SNPs, S01_48462800 and S10_8708314, are not as good a predictor as S01_6006578 since the phenotypes were not all resistant. These two SNPs may not be as close to the genes regulating the race 73 resistance response in their respective locations, or perhaps the mechanism behind the resistance is not simple dominance. The disease means are not significantly different for S10_72250 and thus it would be difficult to use it to predict a resistant phenotype. As stated by Lo et al. (2015) significance analyses are made based on assumptions of the underlying distribution and predictive analyses are based on knowledge of the underlying distribution. Therefore, variables that are highly significant may not be highly predictive. Further study is required to understand the mechanism underlying resistance at this location.

Differences in the Andean reference genome G19833 and the Middle American genome UI111 suggests careful thought should be used when selecting a reference genome for each study. Many SNP datasets are being generated without using the scaffold sequences. This study focused on genotypes in the Middle American gene pool and thus a reference genome genotype in the Middle American gene pool was selected. The use of the Andean reference would have eliminated S10_72250 as it was on scaffold 73 ( Table 3 ). A variation in the reference genome assemblies place SNP S01_48462800 on chromosome 1 in UI111 but on chromosome 2 in G19833. The association results would have been different removing one highly significant SNP and shifting another highly significant SNP to another chromosome. Structural differences between the reference genomes and the lack of phenotypic data make it difficult to compare underlying genes and their expression. With the availability of multiple reference genomes in dry bean, careful evaluation of the genotypes used in the study should be undertaken to select the most appropriate reference genome.

Most known anthracnose resistance genes to race 73 derive either from Andean genotypes, with the multiallelic Co-1 resistance gene family, such as A193 with Co-1 (Mendoza et al., 2001), Kaboon with Co-1² (White Kidney; Zuiderveen et al, 2016), and Widusa with Co-1⁵ (Gonçalves-Vidigal and Kelly, 2006). The Co-4² gene, derived from the Andean genotype SEL1308 (Alzate-Marin et al., 1997), is also known to provide anthracnose resistance to race 73. Middle American genotypes, with genes from the Co-3 resistance gene cluster derived from the genotypes Mexico 222 (White; Co-3; Méndez-Vigo et al., 2005; Rodríguez-Suárez et al., 2008), BAT93 (Tan; Co-9; Geffroy et al., 1999) and Ouro Negro (black; Co-10; Alzate-Marin et al., 2003) provide anthracnose resistance to several anthracnose races, however, from the Co-3 resistance gene cluster only Co-10 provides resistance to race 73. All these resistance genes provide protection against anthracnose infection from early stages onwards and are not known to be age-related. The results from this research study suggest that some resistance to anthracnose may be ARR, and that breeders can target ARR for pyramiding resistance.

The extended greenhouse testing confirmed the presence of anthracnose resistance or intermediate resistance in 13 of the 16 MDP lines and further characterized their resistant behavior. The 1-9 scale used in the extended greenhouse testing is more comparable to other studies and facilitated the identification of the type of ARR. The additional exposure times of disease pressure also showed some genotypic variation across time for the genotypes and allowed the genotypes to be further classified as highly resistant, highly resistant with abrupt resistance breakdown, only early resistance, early susceptibility progressing to short-lived resistance, or a susceptible rating ( Table 2 ). Resistance was not confirmed in either Maverick or F07-499-9-3. Both genotypes were near the breakpoint between resistant and susceptible using the condensed 1-3 scale during the trifoliate screening. Both genotypes had multiple plants with no symptoms but also some plants with symptoms and healthy regrowth tissue. Several reasons can explain the variation. It is possible the variation was due to escapes. Based on the number of plants that would have been escapes in just these two particular lines, it’s unlikely the mixed phenotype is due to escapes. A second possible explanation is these two genotypes were not homozygous for resistance. One example of this type of variation in dry beans was rust resistance in ‘Stampede’. Stampede was released as a genotype resistant to rust with multiple genes for rust resistance. A new race of rust appeared and ‘Stampede’ was found to segregate for resistance. The resistance gene for the new race of rust was not under any selection pressure and allelic frequency had reached nearly 50:50. A third explanation is the seed source was not pure. A fourth possibility is the difference in statical measure usage. Use of the medians in the initial trifoliate screening eliminated the outliers but would not account for segregation. However, segregation would affect the LSMeans in the extended testing. Another possibility is the presence of a mechanism which slows disease progression and suggests partial or quantitative resistance. Understanding the various resistance mechanisms will allow breeders to further utilize that knowledge for cultivar development.

Field testing confirmed the resistance observed in GTS-900 and Remington and by extension suggests Huron, Agassiz, Topaz, Grand Mesa, CDC Pintium, and CDC WM-2 would exhibit high levels of anthracnose resistance under field conditions, but this remains to be confirmed. The field-testing results did not mirror the greenhouse results. In the greenhouse, plants were inoculated at two different stages, the unifoliate (10 days after planting) and first trifoliate (20 days after planting). The field inoculum was applied later (~35 days after planting). The expectation was the trifoliate reaction in the greenhouse would more closely match the field reaction; however, the greenhouse unifoliate reaction was a better indicator of how the genotypes performed in the field. Typically, the unifoliate reaction detects the presence of R genes that tend to be pathogen and race specific. These genes tend to have a large effect and were more likely the ones observed in the field. The ARR genes tend to be “resistance boosters” or result in broad spectrum partial resistance, neither of which are easily observed under field conditions. The environmental conditions in the greenhouse were tightly controlled, thus generating a more specific picture of the resistance reaction.

In this study, MDP genotypes which exhibited ARR to anthracnose race 73 were mostly from the pinto market class of beans (i.e., Agassiz, CDC Pintium, CDCWM-2, Grand Mesa, GTS-900, Remington, and Topaz). However, the navy bean, Huron, also expressed ARR. Three red genotypes exhibited ARR to anthracnose: CENTA Pipil (small red), F07-014-22-2 (red) and Merlot (small red). Two genotypes, Deorho (small red) and UI-911 (black) were fully resistant compared to the resistant checks. The benefits of the newly identified resistant genotypes include reductions in fungicide spray regimes, minimizing the environmental impact, as well as decreasing losses in seed yield and quality, all results that provide an economic benefit to the growers. Further tests are needed to demonstrate the extent of these benefits. Certainly, these genotypes will be valuable new sources for anthracnose resistance for breeding programs. Two clusters of resistance genes on Pv01 and Pv10 were shown to be associated with a resistant phenotype to race 73 of C. lindemuthianum within a region on Pv10 containing the primary gene for ARR resistance. However, only anthracnose race 73 has been tested so far and it remains to be seen if the genes in these genomic regions can lead to durable resistance or tolerance to other races.

With losses of up to 100% seed yield, anthracnose is a devastating disease on dry beans and resistance is usually controlled by major genes clustered in several genomic regions. Typically, seedlings are screened for anthracnose resistance and resistance that develops during later vegetative growth stages is missed. A Genome-Wide Association Study for anthracnose age-related resistance identified two chromosomes containing race 73 resistance using a Middle American diversity panel. Two identified genomic regions were described previously on Pv01, likely corresponding to the multiallelic Co-1 loci and Co-X. Two new regions were identified on Pv10, one of which contains a ZPR1-like gene and is responsible for 23% of the observed variation. The SNPs identified in this new genomic region will be validated and targeted for future marker development to combine multiple anthracnose resistance genes with different modes of action into a single cultivar. The anthracnose disease testing results carried out in the greenhouse were validated using representative cultivars in the field. Field testing confirmed anthracnose resistance to race 73 in two pinto bean cultivars (GTS-900 and Remington). Field testing also confirmed fewer applications of fungicide were effective in preventing significant differences in anthracnose development when these resistant cultivars were used.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: http://arsftfbean.uprm.edu/beancap/research/, HapMap_469-Pinto UI111-geno-GATK-cleaned-imputed-sorted-hpmp-hmp, version: 1.0.

SS, JO and RC conceived the experiments. SS carried out the greenhouse phenotyping on the MDP. WP, DS and RC completed the field testing, initial statistical analysis on field results, and confirmed greenhouse phenotyping on the field-tested lines. AO and PM contributed the MDP SNP dataset. KS completed the GWAS conducted the final statistical analysis and wrote the manuscript. All authors contributed to the article and approved the submitted version.

This project was supported by the United States Department of Agriculture – Agricultural Marketing Service (USDA-AMS) and the North Dakota Department of Agriculture through the Specialty Crop Research Block Grant Program (16-SCBGP-ND-0029). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the USDA. Additional partial support came from the Northarvest Bean Growers Association and Agriculture and Agri-Food Canada. This work used resources of the Center for Computationally Assisted Science and Technology (CCAST) at North Dakota State University, which were made possible in part by NSF MRI Award No. 2019077.

Thank you to the NDSU Pulse Pathology Program for providing C. lindemuthianum race 73 for greenhouse inoculations.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.