Summary

Over the course of the last year our efforts have focused on screening a large collection of ethylmethanesulfonate (EMS) mutants of the soybean cyst nematode (SCN) resistant soybean cultivar Forrest for mutations in two candidate SCN resistant genes, GmRhg1 and GmRhg4, using TILLING (Targeting Induced Local Lesions in Genomes). One to three point mutations per 1.5kb of targeted sequence were identified by screening pooled DNAs representing 768 M2 plants. Mutant plants were phenotyped for alterations in SCN resistance and genotyped to confirm the presence of the mutation and zygosity of each individual. Pooled DNAs representing 3000 Forrest M2 plants are now available and additional screening is underway, including the third 1.5kb of targeted sequence for each gene. For protein-protein interaction studies, several yeast two-hybrid bait constructs were generated for GmRhg4. RNA for prey library construction has also been isolated from SCN-infected root tissues.

Research Activities and Progress

TILLING GmRhg Genes

TILLING data for GmRhg1 and GmRhg4 have been generated from the analysis of 768 DNA samples representing EMS-soybean cv. Forrest M2 lines. Two 1.5kb regions were TILLED for each gene (Figures 1 and 2). A total of 4 mutations were identified for GmRhg1 and a total of 5 mutations were identified for GmRhg4. For GmRhg1 these include a serine to phenylalanine substitution in the leucine-rich repeat domain (S255F), a proline to leucine substitution at amino acid 427 (P427L), a serine to phenylalanine substitution in the 5'end of the predicted kinase domain (S536F), and a silent mutation in the central region of the predicted kinase domain (R669=) (Figure 1). For the GmRhg4 gene, a silent mutation was identified in the leucine-rich repeat domain (Q263=). In addition, a mutation causing a premature stop codon in the leucine-rich repeat domain was identified at amino acid 263 (Q263*) likely resulting in a truncated non-functional GmRhg4 protein that would lack two leucine-rich repeats, the transmembrane domain, and predicted kinase domain. Additional mutations include a leucine to phenylalanine substitution just upstream of the transmembrane domain (L417F), a silent mutation at amino acid 516 (Q516=), and a leucine to phenylalanine substitution slightly upstream from the start of the predicted kinase domain (L537F) (Figure 2). SIFT predictions were made on all missense mutations. SIFT predictions with IC <3.25 are considered confident. Changes with a SIFT score <0.05 are predicted to be damaging to the protein. All GmRhg missense mutations identified had IC values <3.25 and thus the SIFT predictions can be considered confident. In addition to the GmRhg4 truncation mutant (Q263*), the missense mutation L537F was also predicted to be deleterious to the protein. Although the other missense mutations were not predicted to be deleterious to the protein the change could still alter the SCN resistance phenotype (e.g. if involved in recognition of a nematode molecule to trigger resistance).

Phenotyping and Genotyping of GmRhg TILLING Mutants

Phenotyping and genotyping of all nine GmRhg mutant lines was conducted, including silent mutations. Lines carrying silent mutations were included to serve as controls and also further confirm the TILLING approach in soybean as it is a new reverse genetic tool for this crop plant. Mutant seed was received and SCN screening was initiated in May 2005. Leaf samples were harvested from individual plants for DNA extraction and genotyping. A summary of the phenotyping and genotyping results can be found in Table 1. Information pertaining to the number of plants for each mutant line screened, the seed stock used, screen date, SCN female index, flower color (if noted), whether or not genomic DNA was isolated, zygosity of the individual plants based on genotyping, and whether or not seed will be recovered from the plant is indicated. Plants are drying down and seed will be harvested in November 2005. Each mutant line was screened in replicate against SCN PA3 (Hg-type 0; Race 3). Forrest (SCN-resistant) and Essex (SCN-susceptible) were included as controls. A female index of <10% is characteristic of the SCN PA3 resistant cultivar Forrest. A female index >10% would likely indicate a possible alteration in the resistance phenotype. For genotyping, dCAPs primers were designed flanking each individual point mutation. The primers were used in PCR and the amplified products then digested with the appropriate restriction enzyme and separated by gel electrophoresis (see a representative gel in Figures 3 and 4) to confirm the presence of the mutation and determine the zygosity of each plant. Because GmRhg4 is a dominant resistance allele, plants homozygous for a mutation deleterious to the protein would be required to see a SCN susceptible phenotype. In contrast, because GmRhg1 is a codominant resistance gene, plants heterozygous for a mutation that is deleterious to the protein would display a partially resistant phenotype. Plants homozygous for a mutation that is deleterious to GmRhg1 would be expected to display a completely susceptible SCN phenotype.

GmRhg1 TILLING Mutants

We were able to confirm the presence of the three GmRhg1 missense mutations (S255F, P427L, and S536F) by genotyping individual plants. Homozygous mutants were identified for all three mutants. Unfortunately, each of the three mutant lines did not display any alteration in their resistance phenotype against SCN PA3. The M2 plant for the S255F mutation was determined to be homozygous (Figure 1) and all 20 M3 individuals screened were genotyped as homozygous. The M2 plant for the P427L mutation was determined to be heterozygous (Figure 1) and of the seven M3 individuals screened, four were genotyped as wild-type, one was heterozygous, and two were homozygous for the mutation (Table 1). The M2 plant for the S536F mutation was determined to be heterozygous (Figure 1) and as expected the M3 individuals were segregating. Of 17 individuals genotyped (Table 1), four were wild-type, nine were heterozygous, and four were homozygous for the mutation. Female indexes for individual plants of each of these three mutant lines all remained below 10%, similar to wild-type Forrest. As mentioned previously, these three missense mutations were not predicted to be deleterious to the GmRhg1 protein. In contrast, to the missense mutations, individual plants carrying the silent mutation R669= displayed a partially resistant phenotype with female indexes ranging from 14-20%. Further examination of this mutant line via genotyping revealed that this line was a contaminant resulting from a cross between Forrest and Essex which likely occurred during development of M2 populations in the field. Thus, TILLING identified a natural polymorphism between GmRhg1 Essex and Forrest alleles rather than an EMS mutation. The individual plants also had purple flowers, another indication of a cross with Essex which has purple flowers (Purple is dominant to White flower color). Two additional lines, M3-2175 and M3-2176, were identified by TILLING, but the mutation remains to be sequenced. Individuals for these two lines are SCN susceptible with female indexes ranging from 50-123% (in two separate experiments). Furthermore, the plants produce white flowers so it is unlikely they are contaminants due to crossing with Essex (purple flower). We are following up on these mutant lines to determine the exact site of the mutation.

GmRhg4 TILLING Mutants

We were able to confirm the presence of the GmRhg4 silent mutation Q263= and the truncation mutation Q263* by genotyping individual plants and homozygous mutants were identified (Table 1). However, both GmRhg4 missense mutations, L417F and L537F, could not be confirmed by genotyping. In addition, the silent mutation Q516= could not be confirmed. All of these plants genotyped as wild-type. One possible explanation is a mix-up with the seed stocks and this is being rechecked. The M2 plant for the Q263= mutation was determined to be heterozygous (Figure 2), however all six individuals screened were homozygous for the mutation. No alteration in the resistance phenotype was observed as one would expect. The most interesting mutation identified was the GmRhg4 Q263* mutant which should result in a truncated GmRhg4 protein lacking part of the LRR, the TM, and the predicted kinase domain (Figure 2). An individual plant homozygous for this mutation should not make a functional Rhg4 protein, thus if this is the correct candidate resistance gene we would expect a change from the resistant phenotype to susceptible. The M2 plant for the Q263* mutation was determined to be heterozygous, thus we would expect segregation of the mutation in the M3 plants. Of 17 plants phenotyped, all of them had female indexes below 10%, similar to wild-type Forrest (Figure 3C; Table 1). Initial genotyping of the 17 individual plants with dCAPs primer set M2 identified 12 heterozygous for the mutation, five wild-type, and no homozygous plants. This was not the expected 1:2:1 (Wt:Het:Homo) segregation pattern (Figure 3A). Either the homozygous mutants displayed a lethal phenotype, or the dCAPs primers were amplifying another gene that was then masking the homozygotes. To test the second possibility we re-genotyped these lines using the dCAPs primer set 423 designed for the Q263= mutant, where we knew we could see the homozygotes. Figure 4 shows that when we use the different set of primers we now uncover the 5 homozygotes. Interestingly, after sequencing the PCR fragments (~300bp) generated using the first set of primers we determined that it was a mixed amplicon containing sequences differing by 12 nucleotides that resulted in only two amino acid differences. Interestingly, this suggests that there is at least one other copy of a gene that shows similarity to GmRhg4. The possibility exists that these genes may be functionally redundant, so by knocking out one copy (e.g. Q263*) the second copy can still function as an SCN resistance gene and may explain why we do not see any change in the resistance phenotype in this mutant. We are currently exploring this possibility further.

Preliminary/Interesting Findings

• None of the identified GmRhg TILLING mutants showed an altered resistance phenotype to soybean cyst nematode
• Both GmRhg1 and GmRhg4 are still candidate SCN resistance genes
• TILLING mutant R669= confirmed previous findings that a GmRhg1 susceptible allele from Essex (i.e.) results in a slight decrease in soybean resistance to SCN (i.e. partial resistance)
• We show that TILLING is a powerful tool to identify both natural and induced polymorphisms in soybean
• TILLING has led to the identification of possibly one other functional homolog of GmRhg4

The identified Forrest GmRhg4 TILLING mutant Q263* should be a null allele due to an EMS-induced point mutation that results in a premature stop codon in the leucine-rich repeat domain of the GmRhg4 protein. Phenotypic characterization of this soybean mutant for SCN resistance showed that the mutation had no effect on soybean resistance to SCN. At first glance, these results suggest that GmRhg4 may not be the correct candidate resistance gene. However, careful observation of the genotyping results for this mutant identified partial sequence of a second gene with similarity to GmRhg4. The possibility exists that these genes are functionally redundant and may explain why we did not observe an altered SCN resistance phenotype in this mutant. Another possibility is that GmRhg4 is duplicated throughout the soybean genome, but not all are functional copies, which is often the case for plant disease resistance genes. The copy of the GmRhg4 TILLED may be a non-functional copy. Therefore, we can not yet rule out GmRhg4 as an SCN resistance gene. To investigate this possibility further, we identified a restriction fragment length polymorphism to differentiate between these two genes. Genomic DNA was isolated from Essex and Forrest and digested with appropriate enzymes for Southern blot analysis using a probe designed to the GmRhg4 leucine-rich repeat domain. Using this approach we have confirmed the presence of at least two GmRhg4 genes. To further characterize the sequence of these genes we have isolated and are in the final stages of sequencing two different GmRhg4 cDNA clones isolated from SCN-infected Essex and Forrest root tissues. Sequence comparisons will be made between the Essex and Forrest alleles. This is an interesting finding that warrants further investigation.

Work Planned for Coming Year

Because our recent findings suggest that more than one copy of GmRhg4 (and possibly GmRhg1) may exist in Forrest we have put a hold on protein-protein interaction studies until the correct genes conferring SCN resistance are confirmed. We are planning to finish the TILLING screens for GmRhg1 and GmRhg4 to include the final 1.5kb of targeted sequence, as well as additional pooled DNAs for each gene representing a total of 3000 Forrest M2 plants. The possibility exists that the TILLED GmRhg4 is a non-functional copy. Alternatively, the two copies may be functionally redundant. Based on the sequences of the isolated GmRhg4 cDNA clones we can try to design gene-specific primers for TILLING. Isolation of a null allele in a second copy of GmRhg4 can then be crossed to the GmRhg4 Q263* mutant to make a double knock out plant. We will also determine the mutations in the TILLING mutant lines M3-2175 and M3-2176 that show altered SCN resistance. In addition to TILLING, we propose to focus our efforts on GmRhg4 and rescreen a Forrest BAC library for additional BAC clones that may carry copies of GmRhg4. The BAC clones will undergo additional characterization and their position on the soybean genetic map determined. Genomic subclones and cDNAs can then be used for more thorough complementation experiments to confirm their role in SCN resistance.

Manuscripts, Abstracts, Presentations, Intellectual Property

Xiaohong Liu, Aziz Jamai, Khalid Meksem, and Melissa G. Mitchum. Elucidating the Molecular Mechanism of Soybean Resistance to Soybean Cyst Nematode. MU Life Sciences Week 2005, Poster Presentation.