Two sugar beet chitinase genes, BvSP2 and BvSE2, analysed with SNP Amplifluor-like markers, are highly expressed after Fusarium root rot inoculations and field susceptibility trial

Plant material

The 22 sugar beet breeding lines (accessions) used in this study are listed in Table S1 with their corresponding origins. All breeding material is self-incompatible, with population-based selection and propagation employed for their production. Seeds were supplied by the Taldykorgan branch of the Kazakh Research Institute of Agriculture and Plant Growing, Taldykorgan, Almaty district, Kazakhstan. Seeds of KWS2320 were kindly provided by KWS SAAT SE, Einbeck, Germany. The genome of KWS2320 as the Reference line has been completely sequenced (Dohm et al., 2014) and is publicly available on the NCBI database (accession number. PRJNA74567; www.ncbi.nlm.nih.gov).

Artificial inoculation with F. oxysporum fungal suspension in the laboratory and resistance scoring of infected plants

For the laboratory test, sugar beet seedlings were grown in 10-cm diameter pots filled with field-collected, autoclaved soil (120 °C, 20 min treatment) in a greenhouse at 28/22 °C with 16/8 h day/night to the 3–4 leaf growth stage. Fungal inoculation was carried out as described by Walker (1969) with the following steps. Cuts were made with a sterile scalpel measuring one cm in length, penetrating approximately 1/3 of the width in the top part of the root in the pot-grown plants. F. oxysporum Sch., strain No. 5, was isolated earlier from infected roots in a previous field trail among two other major strains No. 50 and 150 (Maui, 2002). The isolates of F. oxysporum were inoculated onto the surface of Czapek media and grown until they were undergoing rapid growth, as described earlier (Thom & Church, 1926). All types of conidia were collected by scalpel and mixed into 1 ml of sterile water. One drop (50 µl) of the resulting suspension was added into the incision and allowed to absorb into the root tissue. The inoculation was repeated 10 days later in the same razor cut and in the same plants. Fusarium RR symptoms were assessed after one month (on leaf blades and petioles) and two months (in roots) following the date of the first inoculation. In laboratory tests, 10 plants were scored in each accession.

The field trials were conducted in the Taldykorgan, Almaty district, on south-eastern part of Kazakhstan during two years (2016 and 2017). Plants of each breeding line were grown in a 16 m row, with four rows in each plot, 60 cm between rows, 18–20 cm between plants, and a planting density of 5–6 plants within 1 m of a row. A total of four replicates of plots in each breeding lines were completely randomized. Plants in the field trial received no artificial inoculations. A total of 10 plants growing in the same general area were randomly selected for scoring in each replicate in the field trial; scoring was repeated for four replicate plots in each accession. The total number of scored biological replicates (plants) in field was 40 for each breeding line.

An identical five-point scoring system was applied in both the laboratory tests and the field trial, as follows: 0—Highly resistant (disease development 0–15% of plants); 1—Stable (disease development 16–30% of plants); 2—Average resistance (disease development 31–50% of plants); 3—Susceptible (disease development 51–70% of plants); 4—Highly susceptible (disease development 71–100% of plants) (Khovanskaya et al., 1985; Urazaliev et al., 2013). Each score was based on two assessments (leaf blades/petioles and roots) in the same inoculated plants in the laboratory test and in plants without inoculation in the field trial. If scores in both assessments were identical, it was recorded as a common result, but both scoring results were recorded in the form of a range, if the two assessment scores differed. Non-inoculated plants in the laboratory test were used as controls with a 0 score.

DNA extraction, bioinformatics and sequencing for SNP identification

DNA was extracted from young plants at the 5–6 leaf stage, using the CTAB method (Dellaporta, Wood & Hicks, 1983). The concentration of DNA was measured using a NanoDrop (Thermo Fisher Scientific, Waltham, MA, USA) and was then adjusted with sterile water to 10 ng/μl.

DNA sequences were obtained from the annotated line KWS2320 on the NCBI database (www.ncbi.nlm.nih.gov); BvSE2 from LOC104888158 and BvSP2 from LOC104888888. Primers were designed to cover entire genes, based on the annotated sequences. Sequences of BvSE2 and BvSP2, as well as primers for sequencing and gene-specific SNP markers KIZ4 and KIZ3, as well as the positions of SNPs identified in this study, are provided in Figs. S1 and S2, respectively. Amplified PCR products for BvSE2 and BvSP2 genes (1,076 and 1,609 bp, respectively), were purified using a PCR Purification kit (Qiagen, Melbourne, Australia) and Sanger sequenced from both 5′- and 3′-ends using a service in the Australian Genome Research Facility, Adelaide, Australia. The comparison of BvSE2 and BvSP2 sequences in breeding lines from Kazakhstan and KWS2320 was used to test for the presence of SNPs. Two independent biological replicates were used to verify the accuracy and stability of the sequence. The SNP positions identified were then used for primer design and SNP Amplifluor-like analysis.

SNP Amplifluor-like genotyping assay

Amplifluor-like SNP analysis, as described earlier (Shavrukov et al., 2016; Jatayev et al., 2017), was carried out using a QuantStudio-7 Real-Time PCR Cycler (Thermo Fisher Scientific, Waltham, MA, USA). The PCR cocktail in each well contained 2xMaster-Mix with the following reagents in their final concentrations: 1.8 mM MgCl₂, 0.25 μM each fluorescent label probe, 0.2 mM each of dNTPs, 0.15 μM of each forward primer, 0.78 μM of reverse primer, 1× PCR Buffer and 0.5 units of Taq DNA polymerase (GenLab, Astana, Kazakhstan). Half of the total PCR volume (5 μl out of 10 μl) was genomic DNA, adjusted for 10 ng/μl, and 5.0 μl or 2.5 μl of each DNA sample was added in 96- or 384-well microplates, respectively. One μl of 1:100 diluted ROX was added as a passive reference label to the Master-Mix according to the manufacturer’s protocol for the PCR Cycler.

PCR cycling followed a program adjusted from those published earlier (Shavrukov et al., 2016; Jatayev et al., 2017), and included initial denaturation at 95 °C for 1 min; 26–28 cycles of 95 °C for 30 s, 55 °C for 30 s and 72 °C for 50 s; and final extension at 72 °C for 5 min. Genotyping with SNP calling was determined automatically by software accompanying the instrument, but each SNP result was checked manually using amplification curves and allele discrimination. Genotyping experiments were repeated twice (two technical replicates) over different days, where technical replicates confirmed the confidence of SNP calls.

Genotyping of eight individual plants in each breeding line was scored based on the majority of the identified alleles in SNP analysis for both studied genes, BvSP2-a₁a₁; -a₁b₁; -b₁b₁; and BvSE2-a₂a₂; -a₂b₂; b₂b₂. The same plants were used for phenotyping scores, where all studied breeding material was highly heterogeneous for both BvSE2 and BvSP2. Reference line KWS2320 was homozygous for both genes, BvSP2-a₁a₁ and BvSE2-a₂a₂.

RNA extraction, cDNA construction and qPCR analysis for gene expression

Plants were grown in the laboratory until 3–4 leaf growth stage, as described above. All leaves were collected from each plant immediately prior inoculation with F. oxysporum Sch., strain No. 5 (Day 0, used as Control), with following similar samplings of other plants from the same pots, 14 and 18 days after the inoculation, based on our preliminary tests. In the field trial, two spots were designated as non-infected (Controls, C), where plants had no symptoms of RR and others were designated as infected (I) areas where they showed severe RR symptoms after one month of growth. Two biological replicates for each treatment and control were used. RNA was extracted from leaf samples of individual plants both in the laboratory and in the field trial using TRIsol-like reagents, as published earlier (Shavrukov et al., 2013). RNA quality was checked by agarose gel electrophoresis with following treatments for 15 min, at room temperature, with 1 µl of DNase I (Thermo Fisher Scientific , Waltham, MA, USA) to remove any traces of DNA. cDNA was constructed from 2 µg of DNase-treated RNA using oligo-dT₍₁₈₎ and Reverse Transcriptase kit, M-MuLV-RH (Biolabmix, Novosibirsk, Russia; http://www.biolabmix.ru) following the manufacturer’s instructions.

Non-diluted cDNA samples were used for qPCR analyses in a Real-Time PCR system, Model CFX96 (BioRad, Gladesville, NSW, Australia) at Flinders University, Australia. The total volume of 10 µl qPCR reactions included 5 µl of 2× KAPA SYBR FAST (KAPA Biosystems, Wilmington, MA, USA), 4 µl of cDNA, and 1 µl of mixed forward and reverse gene-specific primers (3 µM of each). Expression data for the target genes, BvSP2 and BvSE2, were normalized using the reference gene, Glutamine synthetase (LOC104883503), as suggested and used for sugar beet (Taski-Ajdukovic et al., 2012) and repeated twice (two technical replicates). It was checked with infected and control plants as a suitable reference gene prior to the analysis. The Relative standard curve method is based on the ABI Guide to performing relative quantitation of gene expression using real-time quantitative PCR (http://www3.appliedbiosystems.com/cms/groups/mcb_support/documents/generaldocuments/cms_042380.pdf), where serial dilutions were used for each target and reference gene individually (Soole & Smith, 2015). Based on linear calibration of template cDNA dilution factor and Cq value, the threshold cycle values were determined. The coefficients of determination, R², were greater than 0.995 in all studied target and reference genes. The efficiencies of all qPCR primers were calculated based on slope of the corresponding calibration line, E = 10^slope (Borges, Tsai & Caldas, 2012), and they were in the suitable range, 1.8–2.0. Specificities of target and reference genes were confirmed by single distinct peaks on a melting curve and a single band of the expected size in 2% agarose gel electrophoresis. Details of primer sequences and product sizes are present in Table S2.

Statistical analysis

The non-parametric Mann–Whitney U-test was applied for ranking scores in pair comparisons, as suggested for plant disease severity (Shah & Madden, 2004), and the online calculator was used at http://www.socscistatistics.com/tests/mannwhitney. The Mann–Whitney U-test in this case operates with a Nominal variable or Categorical (ordinal) variable data (Harris et al., 2008), and also with a small number of analyses in each studied accession (Fagerland, 2012), making it the most suitable statistical treatment for ranking score comparisons in pairs. The paired Sign test was used to analyze if two sets of scores were correlated with each other using the online calculator http://www.socscistatistics.com/tests/signtest/Default.aspx. For gene expression analysis, average and standard errors were calculated using standard Excel software. Probabilities for significance, P < 0.05 and P < 0.01, were calculated using Student’s t-test.

Structure of the chitinase genes BvSE2 and BvSP2, and their localisation on chromosome 3S

Sequences of BvSE2 and BvSP2 were annotated using sequence obtained from the sugar beet genome database (NCBI). The full sequence of sugar beet BvSE2 chitinase, Class III was identical in all annotated accessions (S66038, AAB28479, XM_010673039 and LOC104888158) including accession number 567433940 for the doubled haploid (DH) line KWS2320 using a Whole genome shotgun sequencing approach (WGSGS). The ORF was 882 base pairs (bp) with 293 predicted amino acids (aa) making up the corresponding protein BvSE2 (accession XP_010671341). Similarly, sugar beet annotated accessions, L25826, AAA32916, XM_010673997 and LOC104888888 with accession 730225012 for WGSGS, were identified for the sequence of BvSP2 chitinase, Class IV. In the comparison of the two chitinases, both ORF of the BvSP2 gene and the BvSP2 protein sequence were shorter, comprising 867 bp and 288 aa, respectively (XP_010671341). No SNP or other polymorphism was found in the databases for either BvSE2 or BvSP2 genes.

The chromosome locations of BvSE2 and BvSP2 on sugar beet chromosome 3 were indicated in the annotated LOC accessions. The positions of these genes on the physical map of chromosome 3S were identified using published information with flanking SNP markers, EBS0085 and BB02714 (Holtgräwe et al., 2014). Using WGSGS, the genes were located in the interval between 97,069 and 122,718 bp on the short arm of chromosome 3; the position of BvSE2 was determined as distal and close to SNP marker EBS0085, while BvSP2 was proximal and located slightly further from the SNP marker BB02714 (Fig. 1).

Sequence analysis revealed SNPs in BvSE2 and BvSP2

Two sugar beet breeding lines (2217 and 2263), with different germplasm origin (Table S1), were selected for sequencing. Two biological replicates were included for each accession. The sequenced genetic regions in the BvSE2 and BvSP2 genes comprised 1,076 and 1,609 bp, respectively. The fragments encompass both 5′- and 3′-UTR, and cover entire genes, including introns. Sequencing of BvSE2 and BvSP2 fragments from both ends of the DNA fragments revealed the full sequence identity when compared to KWS2320. A single SNP was identified in the BvSE2 and BvSP2 genes, with substitutions [W = A/T] and [Y = C/T] at positions 377 and 345 from the Start-codon, in BvSE2 and BvSP2, respectively. Both identified SNPs were heterozygotes in the two biological replicates studied. This means that one allele at each SNP position was identical to the sequence in KWS2320 but the other differed. It is important to note that the SNP [W = A/T] corresponded to a substitution of N₁₂₆ with I₁₂₆ in the predicted amino acid sequence of the BvSE2 protein, but the SNP [Y = C/T] of R₁₁₅ was synonymous.

Phenotyping of the response to F. oxysporum Sch. inoculation, strain No. 5 and genotyping with an SNP Amplifluor-like marker for BvSP2 chitinase gene, Class IV

A total of 10 individual plants for each of 22 accessions plus KWS2320 were examined for response to F. oxysporum Sch., strain No. 5, as described in the Materials and Methods. Laboratory phenotype scores assessing the apparent resistance levels one and two months after inoculation were consistent. An overview of scoring results is presented in Table 1. Breeding lines 1017, 2282 and 2296 showed the highest resistance (score of 0–1) to F. oxysporum, strain No. 5, identical to those of KWS2320. More than half of all studied accessions were classified as resistant, with scores ranging between 1 and 2. Three lines had an equal number of resistant and susceptible plants. Two lines (2125 and 2210) after inoculation with F. oxysporum strain No. 5, showed the strongest symptoms of susceptibility to the phytopathogen (Table 1A).

Genotyping results were based on self-designed Amplifluor-like SNP markers, and amplification with either FAM or VIC produced fluorescence signals linked with the corresponding SNP. This quantitative fluorescence was determined by the qPCR instrument and used for allele discriminations with automatic SNP calls among individual plants in different breeding lines. Examples of allele discrimination for the SNP in BvSP2 and BvSE2 are presented in Fig. 2. The genotypes formed clearly separate two groups with homozygotes and relatively small portion of heterozygote SNPs. All genotypes of KWS2320 plants fell into the homozygote classes, BvSP2-a₁a₁ and BvSE2-a₂a₂.

The genotyping results of plants used for Fusarium RR phenotyping, when tested with the SNP Amplifluor-like marker KIZ3 for the BvSP2 gene, revealed similar but not identical results (Table 1A). Only KWS2320 had identical genotyping scores for all studied plants, while the majority of the genotypes, BvSP2-a₁a₁; -a₁b₁; and -b₁b₁, were estimated in the other breeding lines. Genotyping results showed a high level of association (20 out of 22 lines in total) with the phenotypic estimation of the resistance or susceptibility to F. oxysporum, strain No. 5 (P < 0.01, Sign test). Two cases of non-opposed mismatches for BvSP2 gene were found between phenotyping and genotyping scores (in lines 1042 and 2154) with an intermediate phenotyping score (2) and genotyping of BvSP2-a₁a₁ alleles (Table 1A).

Comparison of inoculation and field trial scores with genotyping using an SNP Amplifluor-like marker for the BvSE2 chitinase gene, Class III

A comparison of disease severity ratings to Fusarium RR in sugar beet accessions using inoculation with F. oxysporum isolate, strain No. 5, and to RR in plants grown in a field trial is presented in Tables 1A and 1B. Most of the breeding lines (16 in the laboratory test and 14 in the field trial) out of 22 studied were resistant to RR. However, a number of accessions were classified as susceptible (score ratings 3 or 4) and differed in the phenotyping score categories, accounting for two and five lines in the laboratory test and field trial, respectively.

Results of the comparison between scores for the laboratory (Table 1A) and field tests (Table 1B) revealed that more than half (12 out of 22) of the studied breeding lines showed identical or very similar score results. Seven lines had overlapping RR scores, with a mostly intermediate phenotype between resistant and susceptible or a mixture of both. This indicates a high degree of similarity between symptom scores of RR in the two different tests. Finally, three conflicting cases (Breeding lines 1002, 2210 and 2236) were found with opposing RR scores in the laboratory test and the field trial (Table 1, indicated in Bold).

Genotyping results using the SNP Amplifluor-like marker KIZ4 for BvSE2 gene, BvSE2-a₂a₂; -a₂b₂; and b₂b₂, showed a strong association with the RR scores in plants grown in the field trial (Table 1B). Four out of five breeding lines susceptible to RR in the field trial (1042, 2154, 2236 and 2262) had a perfect match with the genotyping results. The remaining accessions susceptible to RR (1002 for field trial, and 2172 and 2263 for BvSE2-a₂b₂ genotypes) showed intermediate scores in both the field trial and genotyping. Eleven breeding lines were classified as recording very similar RR scores; all other lines had overlapping RR scores. No conflicting results with opposing RR scores for phenotyping in the field trial and genotyping with SNP marker for BvSE2 were found. KWS2023 was resistant to RR in both laboratory inoculation and in the field trial (Table 1).

Based on results presented in Table 1, three breeding lines with contrasting reactions to inoculation with F. oxysporum, strain 5 in the laboratory test and RR score in field trials, were selected for gene expression analysis. Line 2182 was identified as the most resistant in both tests; line 2210 was sensitive to strain 5 inoculation but resistant to RR in field trials; and, in contrast, line 2236 was resistant to strain 5 inoculation but sensitive to RR in field trials. There was no identified breeding line sensitive in both laboratory and field tests. KWS was used as comprehensive reference line.

Gene expression using qPCR

Expression analysis of BvSP2 and BvSE2 genes after inoculation with F. oxysporum, strain 5, showed that three selected lines (2182, 2236 and KWS2320) had very high expression on Day 14 after inoculation, which returned to levels similar to Controls at Day 18. In contrast, in RR sensitive breeding line 2210, the expression of BvSP2 was 4–7.5-fold smaller at Day 14 compared to three RR resistance lines but it was 200-fold higher on Day 18 (Fig. 3A). In leaf samples from the infected field trial, only accession 2236 showed significantly increased expression of BvSE2 compared to other samples from infected field trials as well as to all Controls (Fig. 3B).