TILLING for allergen reduction and improvement of quality traits in peanut (Arachis hypogaea L.)

Determination of Gene Copy Numbers, and Gene-Specific Amplification

Prior to designing PCR primers for Ara h 1, two genomic clones of Ara h 1 were found in GenBank. The first accession [GenBank: AF432231] was reported by Viquez et al. [17] and is identical to the cDNA sequence of accession L34402 whose encoded protein is designated Ara h 1.0101 by IUIS [2] (isoform Ara h 1.01). A second genomic clone [GenBank: AY581852] was reported by Li et al. [18] and is nearly identical to accession L38853 whose protein is referred to by Chassaigne et al. [19] as isoform 2. For clarity we will refer to this isoform as Ara h 1.02 even though this is not an official IUIS designation. PCR amplification using primers 1306 and 1307 (Table 2) produced two PCR products appearing as a doublet on agarose gel (2,241 bp for Ara h 1.01, and 2,031 bp for Ara h 1.02; Figure 1). Amplicons from gene-specific PCR were 2,211 bp for Ara h 1.01 and 1,666 bp for Ara h 1.02 (Figure 1; Table 1). Analysis of Ara h 1 PCR products from A. hypogaea and its diploid progenitors showed the presence of both genes in A. hypogaea, but only one copy in each diploid. The primer pair specific to Ara h 1.01 (1306/1308; Table 2) amplified only in A. hypogaea and A. ipaensis (B genome), while the primer pair specific to Ara h 1.02 (1306/1309; Table 2) amplified only in A. hypogaea and A. duranensis (A genome; Figure 1). Using the known sequence information, Southern blot analysis of genomic DNA from A. hypogaea was carried out to confirm that no additional copies of Ara h 1 are present in the peanut genome. Genomic DNA digested with HindIII, which has no cut sites within either gene, yielded two nearly overlapping fragments of approximately 6.5 kb each when probed with a full-length Ara h 1.01 probe (PCR product of primers 1306/1308). DNA was also digested with EcoRI, which has one cut site in each copy of Ara h 1. Southern blot analysis revealed four fragments, two from each homeolog, as expected. Lastly, the DNA was cut with AseI, which cuts Ara h 1.01 (two adjacent cut sites within the second intron), but not Ara h 1.02. As expected, three fragments were produced (Figure 2). EcoRI-digested plasmids carrying either Ara h 1.01 or Ara h 1.02 were also loaded as controls; the probe recognized both copies of the gene (data not shown).

Primer no.	Description	Sequence (5'-3')
813	5' Ara h 2	GGAGTGAAAAAGAGAAGAGAATA
817	3' Ara h 2	TCAAGATGGTTACAACTCTGCAGCAACA
815	5' Ara h 2.01	CGATTTACTCATGTACAATTAACAATAGAT
816	5' Ara h 2.02	ATCACCTTAAATTTATACATATTTTCGG
371	3' Ara h 2	CAGCAACAAAACATAGACAACGCC
1306	5' Ara h 1	GAGCAATGAGAGGGAGGGTT
1307	3' Ara h 1	CCTCCTTGGTTTTCCTCCTC
1308	3' Ara h 1.01	TTCTCAGGAGACTCTTTCTCAGG
1309	3' Ara h 1.02	CCTCCTCTTCTTCCCACTCTTG
1048	3' AhFAD2	CTCTGACTATGCATCAG
1055	5' AhFAD2A	GATTACTGATTATTGACTT
1101	5' AhFAD2B	CAGAACCATTAGCTTTG
1458	3' AhFAD2	CAGAACTTGTTCTTGTACCAATAAACACC
1459	5' AhFAD2B	TCAGAACCATTAGCTTTGTAGTAGTGC
1460	5' AhFAD2A	GATTACTGATTATTGACTTGCTTTGTAG

Another target for TILLING, the Δ¹²-fatty acid desaturase gene AhFAD2 has been characterized in studies by Jung et al. [12], López et al. [20], and Patel et al. [21]. This gene is also present in two copies, one in each subgenome of A. hypogaea. The gene sequences are highly conserved between the two, except for an insertion of 19 bp in AhFAD2A (or a deletion in AhFAD2B), 80 bp upstream of the start codon. Gene-specific primer sequences utilizing this indel produce amplicons nearly identical in size: 1,228 bp for AhFAD2A and 1,221 bp for AhFAD2B (Table 1).

Peanut TILLING Populations and Mutation Frequencies

Ara h 2 Mutations

In total, nine SNPs were identified in Ara h 2.01, and five in Ara h 2.02. The first two amino-acid changes identified were in Ara h 2.01 in lines 20-6 (L 49 F) and 37-4 (R 55 H; Table 3). Line 37-4 actually had two nucleotide changes in this gene, but one of them was silent. These two mutations were confirmed in the M₃ and M₄ generations using TILLING. DNA from M₃ or M₄ individuals was analyzed both alone and mixed with wild type DNA. Homozygotes were identified when SNPs were detected in mixed samples but not in the corresponding unmixed samples. Homozygous mutants allowed the testing of IgE binding on the altered proteins from seed extracts. Total protein extracts from homozygous M₄ lines of 20-6 and 37-4 were normalized for loading equal amounts of Ara h 2.01, as measured by anti-Ara h 2 chicken polyclonal antibody, and were tested for binding to serum from four patients with peanut hypersensitivity (HW, DAM, CM, and NF). The IgE-immunoblot showed no differences between the native Ara h 2.01 present in the peanut cultivar Georgia Green (GG) [22] and the Ara h 2.01 allelic variants detected by TILLING in lines 20-6 and 37-4 (Figure 3). Although the mutations were generated in cultivar Tifrunner [23] there is no difference between the Ara h 2.01 proteins of these two cultivars.

Four more silent mutations were found in Ara h 2.01, one of which is identical to the silent mutation in line 37-4. One other amino acid change (A 82 T) was also identified in Ara h 2.01. Three amino acid changes were identified in Ara h 2.02, but two of them (D 70 N) are identical (Table 3). This change occurs in the second DPYSPS motif, which is a known allergenic epitope [14, 15]. The third amino acid change (R 62 Q) also lies within an allergenic epitope, just before the first DPYSPS motif (Additional File 1). Because homozygous seed has not yet been recovered, the ability of these mutant proteins to bind IgE has not yet been tested, although these look to be promising candidates for reduced allergenicity of Ara h 2.02. A G to A mutation was also found 315 bp upstream of the start codon of Ara h 2.02; however, it does not appear to be located within any important promoter elements.

Lastly, a G to A transition was identified in the start codon of Ara h 2.02. A downstream ATG is out of frame, and so a protein knockout was expected. Two M₃ seeds were recovered, a small chip was taken from each for protein analysis, and the seeds were planted. Both seeds grew into phenotypically normal plants. SDS-PAGE analysis of the seed protein extracts confirmed that one of the seeds was indeed missing the 21 kD band which represents the Ara h 2.02 protein [9], and was thus homozygous for the mutation (Figure 4A). The other seed appeared to have a reduced quantity of Ara h 2.02; DNA sequence analysis (data not shown) confirmed that this plant was a heterozygote. Western blot analysis (Figure 4B) also confirmed the absence of Ara h 2.02 protein in the homozygous mutant. Further analysis with 2-D difference gel electrophoresis (2-D DIGE) confirmed that both of the Ara h 2.02 isoforms, shown to differ by a two amino acid truncation at the carboxy terminus [24], were missing in the homozygous mutant line (Figure 4C).

Ara h 1 Mutations

In the longest amplicon, Ara h 1.01 (2,211 bp), signals from both IRDye channels sometimes were not visible on Li-Cor gels due to background and fragment length, but SNPs identified from single-channel signals were later verified by sequencing. Four mutations have been confirmed in Ara h 1.01 (Table 1). One of these, a C to T transition at bp position 593, is silent, but the other three are predicted to induce amino acid changes: R 333 W, P 405 L, and E 437 K (Table 3; Additional File 2). The arginine to tryptophan change at position 333 lies within epitope 12 [25]. Only one mutation was confirmed in Ara h 1.02; a premature stop codon is produced at bp position 304 by a C to T mutation. This is expected to result in a truncated protein of 102 amino acids (Line 133; Additional File 2). All four of these non-silent mutations have been confirmed in the M₃ generation by TILLING. A CAPS (cleaved amplified polymorphic sequence) marker was developed to detect the Ara h 1.02 truncation mutant in succeeding generations. The wild-type amplicon contains six BslI sites, one of which is deleted in the mutant. This marker was used to identify a homozygous mutant in the M₄ generation (Figure 5).

Both Ara h 1 proteins appear as a thick band of approximately 63.5 kD on SDS-PAGE [26]. Although the two genes encode proteins of slightly different sizes, we were unable to resolve both of them with one-dimensional electrophoresis. Thus, 2D SDS-PAGE and 2D-DIGE were attempted to confirm the absence of the protein in seeds of the homozygous Ara h 1.02 truncation mutant. From the 2-D PAGE and 2-D Western blot (Additional File 3) it was not possible to resolve only two distinct Ara h 1 isoforms, an expected result based on published 2-D gel analyses for Ara h 1 [19]. Multiple post-translational protein modifications (i.e. various cleavage products or glycosylation) are produced from the two isoforms of Ara h 1. However, there was a definite difference in the relative Cy3 (wild-type) and Cy5 (mutant) signal intensities for the group of spots in the pI range of 5.9-6.4 representing Ara h 1. From these data it is not possible to conclude that the Ara h 1.02 isoform has been completely eliminated. However, quantitative analysis of the 2-D DIGE mutant and wild-type gels showed that the intensities of three pI 5.9-6.0 spots representing Ara h 1 (Figure 6A, spots 474, 482, 485) were reduced 2.4-2.6-fold in the mutant, but others with a higher pI appeared to increase (1.5-3.5-foldTable 4), although these isoforms were less abundant than the lower pI isoforms in both wild-type and mutant. Also, spots 482 and 485/491 which appear as doublets in the wild-type (Figure 6B) appear as single spots in the mutant (Figure 6C), suggesting that several protein products have indeed been eliminated in the mutant.

AhFAD2 Mutations

One silent mutation was found in each of AhFAD2A and AhFAD2B, and one predicted amino acid change (P 254 L) was found in AhFAD2A. All three of these mutations were C to T transitions, typical for EMS-induced mutations. Several mutations were also identified in these genes which were not typical: an A-insertion, observed twice in AhFAD2B, and three identical A to G mutations in AhFAD2A (Table 3). These are unusual for EMS-induced mutations, but it is perhaps the location and frequency of these mutations which is most intriguing. The A-insertion in AhFAD2B occurs 442 bp after the start codon, causing a frameshift, and likely resulting in a truncated protein due to a premature stop codon (line 81-4; Additional File 4). This mutation was identified in two different M₂ plants in our TILLING populations. Using a CAPS marker [27], this mutation has been shown to be stably inherited in the M₃ generation derived from one of our TILLING mutants (data not shown). In AhFAD2A, three different M₂ plants were found to contain the same mutation, an A to G transition at 448 bp after the start codon. This is predicted to change the amino acid at position 150 from asparagine to aspartic acid (line 4-3; Additional File 4).

Southern Blot Analysis of Ara h 1

DNA for Southern blot analysis was isolated from young leaves of peanut (Arachis hypogaea L.) cv. Georgia Green [22] using the DEAE-cellulose-based technique of Sharma et al. [44]. Twenty micrograms of purified genomic DNA was digested overnight with AseI, EcoRI, or HindIII, and was then loaded on a 0.7% agarose gel and electrophoresed in TBE buffer at 45 V for approximately nine hours. EcoRI-digested pCR-4 TOPO plasmids (Invitrogen, Carlsbad, CA) carrying either Ara h 1.01 or Ara h 1.02 (clones derived from PCR products using primer pairs 1306/1308 and 1306/1309, respectively; Table 2) were also loaded in adjacent lanes as positive controls. The DNA was transferred to Genescreen Plus nylon membrane (Perkin-Elmer, Boston, MA) overnight using the alkaline transfer method [45]. The membrane was probed with a full-length genomic fragment of Ara h 1.01, which was PCR-amplified from a plasmid carrying the fragment. The probe was labeled with α³²P-dCTP using the Random Primed DNA Labelling Kit (Roche, Indianapolis, IN). Unincorporated label was removed using Sephadex G-50 (Sigma, Saint Louis, MO). Hybridization and washing conditions were as described by Sambrook and Russell [45]. The final wash was carried out at 65°C for 15 min. in 0.5 × SSC buffer (75 mM NaCl, 7.5 mM sodium citrate, pH 7.0) with 0.1% SDS. The blot was visualized by exposure to a Storage Phosphor Screen (Amersham Biosciences, Piscataway, NJ) which was then scanned using a Storm 840 imaging system (Amersham Biosciences).

Mutant Peanut Populations

Ethyl methanesulfonate (EMS) or diethylsulfate (DES) treatments were used to induce mutations in the peanut cultivar 'Tifrunner' [23]. Seeds were imbibed in tap water for 10-12 hours. The tap water was then replaced with aqueous solution of mutagen. Three mutagen treatments were tested: 0.4% EMS for 12 h, 1.2% EMS for 4.5 h, or 0.25% DES for 4.5 h. Seeds were soaked in the mutagen solution in 2L Fernbach flasks on a rotary shaker, and were then washed three times in deionized water. (Washes were collected for disposal). The seeds were then rinsed in running water overnight. The M₁ seeds were planted in the field, and one pod was harvested from each plant to generate an M₂ population. M₂ seeds were planted in either the field or greenhouse, and M₃ seed was harvested from them to create permanent TILLING populations.

The entire population will not be distributed because of limited seed availability, although screening for specific mutant genes and distribution of individual lines is possible.

DNA Isolation and Quantification for TILLING

Young leaf tissue was collected from individual M₂ plants, frozen using liquid nitrogen, and either stored at -80°C or lyophilized directly in 96-well collection plates. It was then ground into powder by vortexing with three to four 3-mm stainless-steel grinding balls in 2-ml flat-bottom microcentrifuge tubes, or using a GenoGrinder 2000 (OPS Diagnostics LLC, Bridgeview, NJ), set at 500 strokes/min for 20 sec (liquid nitrogen-frozen tissue), or 1 min (lyophilized tissue). Genomic DNA was extracted using the DNeasy 96 Plant Kit (Qiagen Inc. USA, Valencia, CA) according to the manufacturer's instructions. The DNA was quantified by fluorometry using either PicoGreen (Invitrogen, Carlsbad, CA) or Hoechst 33258 dye in a FluoroCount (Packard/Perkin-Elmer, Waltham, MA) microplate reader. Samples of purified DNA were also run on agarose gel to verify quality. Individual DNA samples were diluted to a working concentration of 5 ng/μl. Individual DNA samples were then four-fold pooled in 96-well format. For verification of individual mutants, genomic DNA from 'Tifrunner' was used as the control.

Primer Design and PCR

Since Ara h 2 genes are small and without introns, differences in the upstream regions of these two genes were used to design gene-specific primers for TILLING (Primers 815 and 816). Based on the available sequence information in GenBank, primers 1306 and 1307 were designed to amplify both copies of Ara h 1. Indels near the 3' end of the open reading frame allowed us to design gene-specific primers 1308 (Ara h 1.01) and 1309 (Ara h 1.02). Primer sequences 1055 (AhFAD2A) and 1101 (AhFAD2B) utilize the indel 80 bp upstream of the start codon to amplify one specific gene copy. These primers are identical to primers aF19 and bF19 used by Patel et al. [21]. For amplification with IRDye-labeled primers, longer oligos are preferred, so primers 1458, 1459, and 1460 were designed. All primer sequences used in this study are shown in Table 2.

Because peanut DNA is highly complex, a first round of unlabeled PCR was used to increase the concentration of target sequences for subsequent labeled PCR. Based on available sequence information and suitability of priming sites, primers for the first round of PCR were designed to amplify both copies of Ara h 2, both copies of Ara h 1, or one specific copy of AhFAD2. The first PCR was carried out in a 25 μl final volume containing 10 ng gDNA, 0.5 U JumpStart Taq DNA Polymerase in 1 × PCR Buffer (Sigma, Saint Louis, MO), 0.2 mM each dATP, dCTP, dGTP and dTTP, and 0.2 μM each forward and reverse primers, under the following conditions: 94°C for 1 min; followed by 8 cycles at 94°C for 35 sec, 58°C for 35 sec (-1°C/cycle), 72°C for 100 sec. The touchdown cycles were followed by 30 cycles of 94°C for 35 sec, 50°C for 35 sec, 72°C for 100 sec, with a final extension of 72°C for 7 min. Reactions were conducted using either a Gene Amp 9700 (Applied Biosystems, Carlsbad, CA) or a PTC-200 (MJ Research, Waltham, MA) thermal cycler.

An aliquot (2 μl) from a 1:40 dilution of the first PCR product was used as input for a second round of PCR, carried out in 10 μl final volume with 0.2 mM each dNTP, 0.25 U ExTaq HS DNA Polymerase (TaKaRa Bio Inc, Shiga, Japan) with IRDye-labeled primers (MWG Biotech, Huntsville, AL), designed to specifically amplify one gene copy. Labeled and unlabeled primers (100 μM stocks) were mixed into a cocktail in a ratio of 3 parts IRD-700-labeled 5' primer: 2 parts unlabeled 5' primer: 4 parts IRD-800-labeled 3' primer: 1 part unlabeled 3' primer. Concentrations of primer cocktail, PCR buffer, and MgCl₂ were optimized for each individual gene. Touchdown PCR was conducted in a PTC-200 thermal cycler (MJ Research, Waltham, MA) as follows: denaturation at 95°C for 2 min followed by 6 cycles of 94°C for 30 sec, 58°C for 30 sec (-1°C/cycle), temperature ramp +0.5°C/sec to 72°C for 80 sec; then 45 cycles of 94°C for 30 sec, 52°C for 30 sec with a temperature ramp +0.5°C/sec to 72°C for 80 sec. This was followed by a final extension at 72°C for 7 min. PCR was immediately followed by the heteroduplex formation step: denaturation at 99°C for 10 min, 70 cycles of reannealing at 70°C for 20 sec, decreasing 0.3°C/cycle, with a final hold at 4°C.

Preparation of Celery Juice Extract (CEL1 Nuclease)

Celery juice extract (CJE), containing CEL1 nuclease, was prepared following the purification protocol from Till et al. [46] with minor modifications. The endonuclease activity and the concentration were tested using a plasmid nicking assay as follows: 200 ng of circular plasmid were incubated with 10 μl of CJE dilution in 1 × CELI Buffer (10 mM MgSO₄, 10 mM HEPES, 10 mM KCl, 0.02% Triton X-100, 0.002% bovine serum albumin) in 20 μl final volume. After incubation at 45°C for 15 min, the sample was placed on ice and 10 μl of 0.15 M EDTA was added to stop the reaction. The digestion products were analyzed on 1% agarose gel. The activity of the CJE was compared with that of Surveyor Nuclease (Transgenomic, Omaha, NE) on a known mutant, detected previously by EcoTILLING [38, 47].

Mutation Screening

After PCR amplification, samples (5 μl from the second PCR) were digested in 1 × CEL1 Buffer with 0.03-0.06 μl CJE in 10 μl total volume, incubated for 15 min at 45°C as described by Till et al. [46]. To stop the reaction 5 μl of 0.15 M EDTA was added per sample, while keeping the samples on ice. The samples were cleaned using Sephadex G-50 (Sigma, Saint Louis, MO), uniformly loaded in 96-well MultiScreen-HV filter plates using a 45-μl MultiScreen Column Loader (Millipore, Billerica, MA) following the manufacturer's instructions. The samples were collected in a catch plate, transferred to a 96-well PCR plate, and dried in an ISS110 Speed Vac centrifugal evaporator (Thermo Savant, Milford, MA). The dried samples were resuspended in 8 μl of formamide loading buffer (37% formamide, 3.75 mM EDTA pH 8, 0.0075% bromophenol blue), and then heated to 80°C for 7 min, and then to 92°C for 2 min [30]. Samples could then be stored in the dark at 4°C for several days until analysis. Samples (0.8 μl) were loaded on 6.5% polyacrylamide gel in 1 × TBE and electrophoresed at 1500 V, 40 mA, 30 W, at 45°C on a Li-Cor 4300 DNA Analyzer (Li-Cor Biosciences, Lincoln, NE). Images were visually analyzed for the presence of cleavage products using Adobe Photoshop (Adobe Systems, Inc, San Jose, CA) and GelBuddy [48]. Putative mutations were identified by fragments appearing in both the 700 and 800 channels, with sizes adding up to that of the full-length PCR product. Because the DNA was pooled four-fold for initial screening, each of the four individuals was then screened against wild type (Tifrunner) to identify the mutant.

Mutations were confirmed by cloning PCR-amplified genes or gene fragments into the pCR4-TOPO cloning vector using the TOPO-TA Cloning Kit for Sequencing (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions, followed by sequencing with M13 forward and reverse primers. Mutation frequency was estimated as the total number of confirmed mutations divided by the total number of base pairs screened [(amplicon size - 200 bp) × individuals screened)]. For each amplicon, 200 bp is subtracted to adjust for the 100-bp regions at the top and bottom of TILLING gel images that are difficult to analyze [5].

SDS-PAGE, Western Blots, and IgE-Binding Analysis

Seed protein extraction was carried out as described by Koppelman et al. [3]. Protein quantification was performed using the Bio-Rad Protein Assay kit (Bio-Rad, Hercules, CA) and bovine serum albumin (BSA) as the standard. Ten micrograms of total protein were loaded per lane and separated in 12% SDS-PAGE gels. The proteins were then electroblotted at 30 V overnight to PVDF membranes, 0.2 μm (Bio-Rad). The membranes were blocked with 5% non-fat dry milk in PBS-T (137 mM NaCl, 2.7 mM KCl, 1.4 mM KH₂PO₄, 4.3 mM Na₂HPO₄, 0.1% Tween-20) at room temperature for 1 h with gentle agitation. For Western blot analysis, after blocking the membrane was probed with chicken anti-Ara h 2 or anti-Ara h 1 (1:8,000 dilution) primary antibodies for 1 h followed by secondary anti-chicken antibody-horseradish peroxidase (HRP; 1:100,000; Sigma, Saint Louis, MO) in PBS-T with 2% non-fat dry milk for 1 h. For IgE binding analysis, sera were obtained from four individuals (HW, DAM, CM, and NF) documented as allergic to peanuts with either a positive food challenge or a convincing history of peanut allergy. Sera from allergic individuals were collected at the University of Arkansas for Medical Sciences, Little Rock, AR, approved and in accordance with the rules and regulations of the institutional review board. Blots incubated with a 1:20 dilution of the four sera were allowed to hybridize overnight at 4°C. All membranes were washed three times with PBS-T for 5 min after each antibody incubation. ECL Plus detection reagents (Amersham Biosciences) were used for chemifluorescent detection, followed by scanning on a Storm 840 imaging system (Amersham Biosciences). For IgE binding analysis, relative quantification of Ara h 2.01 protein was performed by densitometry with Quantity One software (Bio-Rad) and all samples were normalized to the amount of Ara h 2.01/10 μg of total protein in wild-type. Further IgE immunoblot analyses were performed using 100 ng of Ara h 2.01-normalized total protein.

2-D PAGE and DIGE

Suspected homozygous knockout mutants of Ara h 2.02 and Ara h 1.02 were subjected to two-dimensional protein analyses using difference gel electrophoresis (2-D DIGE) essentially according to Chu et al. [4]. Labeled protein samples for analysis of 17-19 kD Ara h 2 (wild-type - Cy3; mutant - Cy2) were resolved using pI 3-10 immobilized pH gradient (IPG) strips. For analysis of ~65 kD Ara h 1, 50 μg each of Cy3-labeled wild-type, Cy5-labeled mutant, and Cy2-labeled mixture of wild-type and mutant protein extracts were mixed and focused in a 24-cm pH 5.3-6.5 IPG strip at 10,000 V for 121 kVhr before running the second dimension in a 10% Tris-glycine SDS-PAGE gel. Labeled spots were detected with a Typhoon 9400 Variable Mode Imager (GE Healthcare) and digital image analysis was conducted with SameSpot Progenesis (Nonlinear Dynamics).