Progress report for GNE21-258
Project Information
Objective 1: Determine the inheritance patterns of the HrpN-induced HR in one or more mapping populations of tobacco.
Objective 2: Identify one or more candidate resistance genes in a tobacco mapping population by performing a bulked segregant analysis.
Fire blight, caused by the bacterium Erwinia amylovora (Ea), is a devastating disease that primarily afflicts apple (Malus x domestica) and pear (Pyrus communis). The disease is characterized by wilting and death of flowers and shoots, production of bacterial ooze from plant parts, and canker formation on the tree trunk. Left unmanaged, fire blight can result in death of the entire tree. Outbreaks of fire blight in orchards are costly to growers, often resulting in millions of dollars in crop and tree loses (van der Zwet, Orolaza-Halbrendt, and Zeller 2012). Treatment of fire blight currently relies on application of antibiotics (e.g. streptomycin). However, heavy antibiotic use has led to the rise of streptomycin-resistant Ea strains that are causing recent outbreaks in the Northeastern US (Dougherty et al. 2021; Russo et al. 2008). Additionally, antibiotic application to flowers must be correctly timed, as antibiotics are not effective once Ea has gained access to internal plant tissues. This suggests the need to develop genetic sources of resistance as a complementary fire blight management strategy. Currently no apple cultivars with complete resistance to fire blight are available, and many of the most economically important apple and pear cultivars are highly susceptible to fire blight (van der Zwet, Orolaza-Halbrendt, and Zeller 2012; Kostick, Norelli, and Evans 2019).
HrpN is a proteinaceous virulence factor secreted by Ea that facilitates the bacterium’s infection of plant tissues (Wei et al. 1992). While HrpN functions in fire blight development on host plants, it has been shown to activate plant defenses in nonhost plants. HrpN, when secreted by Ea or infiltrated as a cell-free elicitor, induces a strong, rapid, and localized cell death defense reaction called the hypersensitive response (HR) in leaves of the nonhost plant Nicotiana tabacum (cultivated tobacco). The HR is associated with gene-for-gene recognition and plant disease resistance, leading us to hypothesize that tobacco possesses a HrpN-recognizing resistance (R) protein. Through the objectives proposed in this study, we aim to identify the tobacco R protein conferring HrpN recognition. By elucidating HrpN recognition in tobacco, this research will uncover gene(s) that can be transferred to apple and pear to confer HrpN recognition and potentially fire blight resistance.
The long-term goal of this project, development of fire blight-resistant apple and pear, will directly assist Northeast farmers by increasing productivity, reducing costs, and increasing net farm income. Crop losses due to fire blight outbreaks will decrease, increasing the productivity and profitability of grower operations. Additionally, growers will save on costs associated with annual chemical treatments, including product, labor, and transportation costs. These reduced costs will ensure that healthful produce is available to consumers at a reasonable price. Additionally, lowered antibiotic use provides a reduction of environmental risks in agriculture by reducing the prevalence of antibiotic-resistant microbes and avoiding off-target effects on beneficial microbes.
Research
Objective 1: Determine the inheritance patterns of the HrpN-induced HR in one or more mapping populations of tobacco.
Rationale
The strong cell death defense response induced by HrpN in tobacco is suggestive of R gene mediated recognition. As tobacco is a diverse crop species cultivated throughout the world, I can screen these accessions for natural variation in the HrpN response. A small number of accessions have likely accumulated genetic mutations in the HrpN-recognizing R gene through random human-mediated selection processes. Those accessions with a defective receptor will show a weak or no cell death response when its leaves are infiltrated with HrpN protein.
In preliminary studies, I screened 127 tobacco accessions collected from around the world by infiltrating their leaves with HrpN protein and visually assessing the cell death response. I identified 12 tobacco accessions that were deficient in their visual cell death response to HrpN. I then used an electrolyte leakage assay to identify three accessions with the quantitatively weakest cell death response to HrpN. The three unresponsive tobacco accessions likely have genetic mutations in one or more genes that are responsible for the HrpN-induced HR. Crossing of these accessions with a HrpN-responsive accession will help elucidate the heritability and segregation patterns of the trait.
Methods
F1 Generation and Phenotyping
To generate F1 populations, each unresponsive tobacco accession will be crossed with the HrpN-responsive tobacco accession TI 1608. To perform a cross, anthers are removed from an unopened flower that is within 24 hours of opening. The emasculated flower’s stigma is then immediately pollinated with an anther from the proper accession. 4 weeks after pollination, the seed pods are removed, and the seed is collected and stored in manila envelopes.
Once seed has been collected, 40 individuals of each of the three F1 families will be grown to 6 weeks old in a plant growth chamber. Four leaves on each plant will be infiltrated with 100 µM HrpN142-188, a peptide corresponding to amino acids 142 to 188 of full-length HrpN protein. This peptide represents the major HR-inducing segment of HrpN. An advantage of using HrpN142-188 is that it is the simplest HR trigger available to test HrpN-mediated plant responses and eliminates other potential plant reactions triggered by full-length HrpN. Two days after infiltration, leaves are scored visually using a rating scale: 1 (no tissue necrosis), 2 (mild underside tissue collapse), or 3 (complete tissue necrosis). Each individual is then classified as strongly responsive (≥ 75% of infiltrations scored a 3), unresponsive (≥ 50% of infiltrations scored a 1), or moderately responsive (not meeting the criteria of strong or weak responders). As R gene mediated recognition of pathogen virulence factors is typically a dominant trait, it is expected that all individuals in each population will be strongly or moderately responsive to HrpN142-188.
01/22 Updates
The HrpN-responsive tobacco accession used in F1 generation was changed to TC 319. This accession was used for sequencing the tobacco genome. The use of TC 319 will lead to a stronger sequencing data analysis in Objective 2. TC 319 has been verified to respond to HrpN peptide.
The peptide HrpN142-188 has been replaced with HrpN140-176. This peptide has increased potency from HrpN142-188 and represents a more specific trigger of HR in tobacco.
For scoring of progeny in the F1 and subsequent generations, a binary classification system was adopted. Individuals were classified as responsive if at least one infiltration resulted in a rating of 3 (complete tissue necrosis). Individuals with infiltrations resulting in only ratings of 1 and 2 are classified as unresponsive. This will allow for easier grouping of individuals into HrpN-responsive and HrpN-unresponsive pools as outlined in Objective 2.
F2 Generation and Phenotyping
Plants in each of the F1 populations will be allowed to self-pollinate. Seed will be collected from the selfed plants as described previously. 200 individuals of each of F2 population will be grown and phenotyped with HrpN142-188 using the same methods as the F1 screen. To accommodate the large number of plants, some plants may be grown in rented space in a university greenhouse. In the F2 populations, a segregation of phenotypes is expected, where some plants are strongly or moderately responsive and others are unresponsive. A chi-squared test will be used to determine the segregation patterns of each population. The F2 which most closely resembles a Mendelian segregation pattern (e.g. 3:1) of strongly/moderately responsive: unresponsive individuals will be utilized for the bulked segregant analysis (BSA), as this indicates the trait is controlled by a small number of genes. However, the BSA may also be utilized in populations displaying more complex inheritance patterns, as long as individuals are sorted into contrasting phenotypic pools (see Objective 2).
01/22 Updates
To confirm the HrpN-responsive and HrpN-unresponsive designations of F2 individuals, each plant will be selfed to generate F3 populations. 10 individuals from each F3 population will be screened with HrpN140-176 and classified as HrpN responsive or unresponsive. F3 populations that contain no trait segregation (i.e. consisting of individuals only HrpN responsive or only HrpN unresponsive) will be utilized in the BSA, while F3 populations displaying trait segregation will be excluded. This will increase the confidence in assigning F2 individuals to HrpN responsive and unresponsive groups.
01/24 Updates
To better evaluate the responsiveness of the F3 families to HrpN140-176, a 'responsiveness' score was calculated for each family. Each individual was leaf infiltrated four times with HrpN140-176. After two days, each infiltration was scored using the 1,2,3 rating system described above. These values were then summed for an individual, giving a responsiveness score between 4 and 12. Finally, the individual scores were averaged across the 10 screened F3 individuals to give an overall responsiveness rating for each F3 family.
Objective 2: Identify one or more candidate resistance genes in a tobacco mapping population by performing a bulked segregant analysis.
Rationale
Once the segregation patterns have been determined in my F2 populations, a bulked segregant analysis (BSA) can be performed. A BSA consists of a pooled DNA sequencing strategy, where each pool corresponds to a specific phenotype. The pooling of many individuals in each sample allows the investigator to identify genetic mutations, typically single nucleotide polymorphisms (SNPs), that are strongly associated with the phenotype of interest. Additionally, the pooling of samples, rather than sequencing of individuals, saves sequencing costs and simplifies analysis.
Methods
Sample Grouping and DNA Extraction
F2 individuals will be grouped (i.e. bulked) into two phenotypic categories: HrpN-responsive (consisting of individuals identified as strongly or moderately responsive in Objective 1) or HrpN-unresponsive. DNA will be extracted individually from leaf tissue of 30-40 individuals in each phenotypic group using the Qiagen DNeasy Plant Mini kit. DNA quality will be assayed using a department-provided NanoDrop. For each group, the combined DNA sample will be created by mixing 2 µl of each individual sample.
DNA Sequencing
The two pooled DNA samples will be submitted to the Penn State Genomics Core Facility for library construction and DNA sequencing. Libraries will be assembled using the Illumina DNA PCR-free prep kit. Genome sequencing will be done using two runs of NextSeq High Output 150x150 paired-end sequencing. As the tobacco genome is 4.5 gigabases, this will generate ~30x coverage of reads across the genome. 30x coverage provides sufficient sequencing data to detect genetic variation between the phenotypic groups (Schultink et al. 2019).
Data Analysis
Computational analyses will be conducted using the Penn State’s Roar supercomputer, a free high-performance research cloud. Reads will be mapped to the tobacco reference genome using BWA-MEM (Li and Durbin 2010). Calling of SNPs and other genetic polymorphisms will be performed using GATK (McKenna et al. 2010). BWA-MEM and GATK are freely available software that are commonly employed in the analysis of DNA sequencing data. SNPs and other polymorphisms will be filtered based on the following criteria: a differential abundance of at least 25% between the two pools, a minimum mapping quality of 10, and their presence in the coding sequence or promoter of genes with structural features of known defense genes (e.g. receptor-like kinases and receptor-like proteins). The genome location of polymorphisms which meet these criteria will be used to identify up to 10 high quality gene candidates that may be mediating tobacco’s defense response to HrpN.
A graphical overview of Objectives 1 & 2 can be found here: Experiment Diagram
01/22 Objective 1
F1 generation and phenotyping
Four F1 families, representing crosses of different HrpN-unresponsive accessions with the responsive accession TC 319, were screened with HrpN140-176. In the TC 319 x TI 126 cross, every individual tested (n=20) was rated as HrpN responsive, suggesting that the HrpN response has a genetic basis and is a dominant trait.
F2 Generation and Phenotyping
An F1 individual from the TC 319 x TI 126 cross was selfed to produce an F2 population. 220 F2 individuals have been screened with HrpN140-176. Of these, 174 individuals were classified as HrpN responsive and 46 as HrpN unresponsive. This result is in line with an expected 3:1 ratio of HrpN responsive:HrpN unresponsive individuals (Chi-sqaure probability=0.16). Thus, HrpN recognition appears to be governed by a single gene.
Approximately 160 F2 individuals are currently being grown in a greenhouse for seed collection. Screening of the F3 lines from these individuals is expected to conclude in summer 2022.
01/23 Update-Objective 1
A total of 346 F2 individuals have now been screened with HrpN140-176. 258 individuals were classified as HrpN responsive and 88 as HrpN unresponsive. This result further supports the expected 3:1 ratio of HrpN responsive:HrpN unresponsive individuals (Chi-square probability=0.85).
I have screened 64 F3 families-31 from responsive F2 parents and 33 from unresponsive F2 parents. Of these, 18 and 16 F3 families had consistent phenotypes with their responsive and unresponsive F2 parent phenotypes, respectively. Thus, these F2 parents will be utilized in the bulked segregant analysis. I am currently in the process of screening 29 more F3 families from responsive F2 parents and 27 more from unresponsive F2 parents. At the current rate, I will have at least 25 responsive F2 parents and 25 unresponsive F2 parents that can be utilized for sequencing by early Spring. One more round of F2/F3 screening may be needed to ensure adequate numbers of individuals in each phenotypic category.
01/24 Update-Objective 1
F3 screening concluded in spring 2023. A total of 113 F3 families were screened- 53 from responsive F2 parents and 60 from unresponsive F2 parents. For each F3 family, 'responsiveness' to HrpN140-176 was calculated (see 01/24 update in Methods).
01/24 Update-Objective 2
Sample Grouping
The results of the F3 phenotyping were used to inform selection of F2 parents to be used in sequencing. For sequencing, a minimum of 30 F2 individuals of each phenotypic class (HrpN-responsive and HrpN-unresponsive) were needed. Using the F3 responsiveness scores, we identified 30 F2 individuals in each category with progeny that gave a consistent phenotype. F3 families from the chosen responsive F2 parents had an average responsiveness score of 11.4, while F3 families from the chosen unresponsive F2 parents had an average responsiveness score of 5.6. Thus, we have increased confidence in the phenotypic assignments of our F2 individuals because the trait showed strong heritability into the next generation.
DNA Extraction
Young leaf tissue (stored at -80 C) from the 60 F2 individuals chosen for sequencing was utilized for DNA extraction. All extractions were performed with the Promega Wizard Genomic DNA Purification Kit. The amount of DNA in each extraction was determined through NanoDrop and Qubit assays. Approximately 10 ng of DNA from each of the 30 individual preps in each phenotypic class were pooled to create the responsive and unresponsive DNA samples.
DNA Sequencing
Whole-genome sequencing of the two samples was performed in Summer 2023 at the Huck Genomics Core Facility at Penn State. 150 base pair Illumina paired-end sequencing was utilized. 192 billion bases were sequenced in the responsive sample and 181 billion bases were sequenced in the unresponsive sample, representing ~40X coverage for each sample based on the 4.5 billion base pair size of the tobacco genome.
Data Analysis
The reads of each sample were mapped to the tobacco reference genome assemble using BWA-MEM. Each sample had ~99% of reads align to the genome assembly. Next, short DNA polymorphisms like single-nucleotide polymorphisms (SNPs), insertions, and deletions were called in each sample using the program 'bcftools'. The effect of each polymorphism (e.g. missense mutation) was annotated using the program 'snpEff'. In total, ~10 million DNA polymorphisms were identified across the two samples. Only ~40,000 (0.4 %) of these are protein-coding altering variants, while the vast majority of the variants are located in intergenic space. The protein-coding altering variants are of interest due to their potential to alter plant traits. Thus, one or more of those variants could underly the HrpN responsiveness phenotype.
To identify variants that could underlie our trait of interest, we calculated a variant allele frequency for all of the protein-coding altering variants in each sample, based on the percentage of sequencing reads that are supporting a given variant (versus ones that support the non-variant or reference sequence). Because the reference genome of tobacco is derived from the same variety used as the responsive parent in our initial cross, we expect variants with a higher variant allele frequency in the unresponsive sample to be potential candidates for underlying the HrpN response trait. When filtering for variants with a frequency difference of 0.5 or higher with good mapping quality, this narrows the list down to 260 variants. In this collection of variants, many of them reside in genes that are known to be involved in defense, such as receptor-like kinases.
We are performing additional analyses to select 10-20 high quality gene candidates that may underlie the response to HrpN. For example, we are utilizing the chromosome assembly of the tobacco genome to identify the genomic regions that are associated with our trait of interest. This can help narrow down further some of the protein-coding variants identified. When we have identified our final candidates, we will proceed with functional analyses to determine if any of them are involved in the HrpN response trait.
Education & Outreach Activities and Participation Summary
Participation Summary:
Results from the work proposed here will be disseminated at regional and national conferences. Specifically, I will give an oral presentation the 2022 Northeast Division meeting of the American Phytopathological Society. This will target local researchers in academia, government, extension, and industry. Sharing of my research here will help illustrate the potential of utilizing nonhost species for resistance genes to other plant disease researchers. My work will also be presented at the third International Symposium of Fire Blight on Rosaceous Plants to be held in summer of 2022. This is the premier conference for fire blight-related research, drawing in hundreds of researchers from around the world. My work will be of value to government and extension personnel who are interested in new ways to combat fire blight. This conference could help open the door to collaborations in future efforts to generate fire blight-resistant apple and pear germplasm.
Additionally, the results from this study will be published in a high-quality molecular plant pathology journal, such as Molecular Plant-Microbe Interactions. This will reach other plant pathology researchers across the world. To target extension personnel and apple and pear farmers in Pennsylvania, I will publish a news release on the publication for Penn State’s extension newsletter. This release will highlight the results of my research and its potential for fire blight management. As another avenue to directly reach growers, I will give a short talk at Penn State’s Fruit Research and Extension Center in summer 2023 at its biannual open house. This event attracts tree fruit growers and extension personnel from around Pennsylvania. Both extension outreach opportunities will be crucial in fostering continued positive relationships between Penn State University and Pennsylvania’s tree fruit producers. By being informed of this research at an earlier stage, growers may be more receptive to later outputs of the project, such as fire blight-resistant apple and pear varieties.
The opportunities for farmer outreach will expand past the duration of this project. New apple and pear varieties will be generated containing the tobacco resistance gene. Work with other researchers, extension personnel, and enthusiastic growers will be necessary to conduct field trials of the new varieties to test for fire blight resistance and productivity. If trials are successful, these varieties could be marketed to northeast growers as fire blight-resistant germplasm. Additionally, the results of this research will continue to be disseminated through conference presentations and manuscript publication.
01/22 Progress
I plan on presenting this spring at the 2022 APS Northeastern Division meeting. Because the 2022 International Symposium of Fire Blight on Rosaceous Plants is being held in Germany, I may present instead at the national APS meeting.
01/23 Progress
In early August 2022, I presented a poster on this research at the national APS meeting (Plant Health) in Pittsburgh, PA. The poster is attached below.
01/24 Progress
With the research phase wrapping up, we are beginning to write up the results of the study for publication in an academic plant pathology journal. Additionally, I will be giving seminars to my department and graduate program to communicate the initial results of the study.
Information Products
- Exploring natural varaiation in tobacco nonhost immune responses to Erwinia amylovora (Conference/Presentation Material)