Cucumber Team | 2023 Progress Report

View Figures 1 – 16 in the PDF version of the 2023 CucCAP Cucumber Team Report.

Team members:

  • Yiqun Weng (YW), USDA-ARS, University of Wisconsin Madison,
  • Rebecca Grumet (RG), Michigan State University
  • Kai-Shu Ling (KL), USDA-ARS Charleston, SC
  • Anthony Keinath (AK), Clemson University, SC

CucCAP Affiliated postdocs and graduate Students

  • Dr. Feifan Chen, Postdoc research associate, University of Wisconsin Madison (Weng)
  • Ying-Chen Lin, graduate students, Michigan State Univ (Grumet)
  • Stephanie Rett-Cadman, graduate student, Michigan State Univ (Grumet)
  • Junyi Tan, graduate student, University of Wisconsin Madison (Weng)
  • Daoliang Yu, graduate student, University of Wisconsin Madison (Weng)
  • Savanna Beyer, technical support, USDA-ARS (Weng)
  • Sierra Zardus, technical support, Clemson University (Keinath)
  • Anna Mothersbaugh, technical support, Clemson University (Keinath)
  • Bazgha Zia, postdoc research associate, USDA-ARS, Charleston, SC (Ling)

Seed multiplication of cucumber core collection (YW & industry collaborators)

Seed increase of 388 accessions from diverse taxonomic groups, geographic origins, and market groups was continued in 2022 by five industry collaborators. As of March 28, 2023, seed increase of 106 accessions has completed (>1000 seeds each). The goal under this sub-objective is expected to be accomplished by 2023.

Map and develop markers for disease resistance

QTL mapping of resistances | Downy mildew (YW & AK)

One goal under this objective is to conduct QTL mapping of DM resistance in WI7773 (an introgression line, PT108, with DM resistance derived presumably from C. hystrix) and WI7631. In 2022, 91 WI7769 F2:3 families from the cross between WI7773 and susceptible line ‘9930’ (two replications, eight plants per rep) were grown in open fields at Clemson Univ., SC to examine inoculation responses to natural infection of the DM pathogen. Phenotypic data for general impression (GI) of DM symptoms were recorded at two time points. Based on the two years’ DM data in two populations (WI7747 and WI7769), we performed bulked segregant analysis (BSA)-Seq aiming to identify sub-chromosomal regions harboring DM resistance QTL in the parental lines. Genome-wide delta SNP-index curves for the two populations are shown in Figure 1 which suggest a major-effect DM resistance QTL on chromosome 2 and two on chromosome 5 in the WI7769 and WI7747 populations, respectively.

Figure 1. Genomewide delta SNP-index curve for the WI7668 (left) and WI7747 (right) populations. The red and green lines in each graph indicate 95% and 99% confidence interval, respectively. Based on BSASeq, a major-effect DM resistance QTL on chromosome 2 and two on chromosome 5 seem to present in the two populations, respectively.

In 2022 field trials at Clemson Univ. we also screened many other materials for DM responses including 120 inbred lines (for GWAS), recombinant plants used for fine mapping of the dm4.1 (from WI7120) and dm5.3 (from PI 197088) major-effect QTL (see below), as well as multiple introgression lines for DM resistance QTL. In August 2022, there was an epidemic of downy mildew at the University of Wisconsin Hancock Agriculture Research Station where DM occurrence is usually very rare. We were able to observe DM responses for 204 cucumber accessions. Many known sources of DM resistance were confirmed, and some new sources of resistance were identified. The epidemic also confirmed the DM resistance of introgressed lines with varying numbers of DM resistance QTL.

QTL mapping of resistances | Phytophthora fruit rot (RG and Ying-Chen Lin)

1) QTL mapping of resistance derived from PI 109483. The QTL on chromosome 5 identified from the cross with pickling cucumber breeding line Gy14 and DH line A4-3, qPFR5.1, was tested in a second genetic background, the fresh market cucumber variety, Poinsett 76. F2 plants (Poinsett 76 x DH A4-3; n=768) were genotyped and individuals homozygous for either Poinsett 76 or DH A4-3 alleles at qPFR5.1 were self-pollinated. The resulting 25 F3 families were grown in the greenhouse and field. Fruit were harvested 2-3 times a week at the age of 5-7 dpp (7-10 cm long) to provide 10-50 fruits/plant for replication and brought to the lab for inoculation with P. capsici. Consistent with the Gy14 background, the DH A4-3 allele was associated with resistance in the Poinsett background in the greenhouse and field (Figure 2).

Figure 2. Allelic effect of qPFR5.1 in fresh market cucumber background. F3 families of Poinsett 76 x A4-3 possessing either the A4-3 or Poinsett allele at qPFR5.1. Each point is the mean of >20 fruits/family from the greenhouse and > 50 fruits/family from the field.

2) Association analysis for PFR resistance.
GWAS. The cucumber core collection was screened from 2019 to 2021, the number of accessions grown each year varied depending on seed availability. Phenotypic data was obtained from 378 accessions with 1-4 years of disease scores per accession; 70% of the accessions had at least two years of data. Disease scores for the population were normally distributed, consistent with a quantitative trait. The correlation between years ranged from 0.48-0.80 (Figure 3).
GWAS was performed using the resequencing data obtained for the cucumber core collection. SNP data was downloaded from CucGenDB and filtered using VCFtools (Danecek et al., 2011) with the following criteria: bi-allelic, GQ scores > 15, maximum read depth within one standard deviation of the mean read depth, and minor allele frequency > 0.1. GWAS marker-trait association analyses were carried out for the phenotypic data from each year, as well as the estimated genotypic best linear unbiased predictor (BLUPs) to correct effects from different environments. BLUPs were calculated using the R package lme4. Association analysis was performed using the R package GAPIT 3.0 with its implanted MLM, FarmCPU, and BLINK models (Wang & Zhang, 2021). The Manhattan plots and Quantile-Quantile (Q-Q) plots were graphed using R package gwaspr. The genome-wide significance threshold was determined using the Bonferroni correction at α = 0.05.

Figure 3. Disease score distribution and correlation from 2019-2021, and distribution of weighted average and BLUP values over the three years. The value for each accession in each year is the mean of 20-50 fruit. Darkly shaded bars in weighted average represent lines selected for XP-GWAS.

Though variations were observed across phenotypic data and models, significant SNPs on chromosome 1 at 21.11 and chromosome 2 at 10.22 Mb were detected consistently (Figure 4).

Figure 4. Manhattan and QQ plots of young fruit resistance in the cucumber core collection using combined (BLUP) data from 2019-2021. The GWAS analyses were performed using MLM, MLMM, FarmCPU, and BLINK models. The consensus SNPs identified from multiple models are indicated by vertical lines.

XP-GWAS: To enable additional replication of phenotypic data, we are also performing extreme phenotype (XP) GWAS. The weighted disease scores from 2019-2021 data were used to select the 30 most resistant and susceptible accessions (Figure 2). These genotypes were tested in an additional replicated trial in 2022 (Figure 5). XP-GWAS analysis is in progress.

Figure 5. Distribution of disease scores of selected resistant and susceptible lines in 2019, 2021, their performance in replicated trial in 2022 and weighted average
(2019-2022). Scoring for 20192021 is as described in Figure 2. Values for each line in 2022 are the mean of 50-100 fruits/line.

Several of the most resistant lines identified in 2019 were also tested in replicated trials in 2021 and 2022. Three lines showed reproducibly lower disease scores than Gy14 (Figure 6).

Figure 6. Young fruit response of selected lines to inoculation with P. capsici in 2019, 2021, and 2022.
AM #
Country of origin
AM032 PI 105340 Kuai Huang Kwa China
AM280 NSL 197095 Wautoma United States
AM185 PI 481614 Gagon Bhutan

QTL mapping of resistances | CGMMV (KL and YW)

Cucurbit green mottle mosaic virus (CGMMV) is a seed-borne virus that can be introduced to new areas via infected plant material. It is easily spread through mechanical means, agricultural practices, and plant to plant contact. Due to its contagious nature, it is important to devise permanent solution to manage CGMMV. CGMMV causes serious disease symptoms and losses in cucurbits particularly, cucumber and watermelon. As an effort to combat CGMMV in cucumber and watermelon, genetic resources were explored to develop genetically resistant/tolerant cucumber and watermelon lines in this study.
Initially 50 cucumber lines were screened to assess phenotypic reactions to the CGMMV infection. As a result, three lines were identified as tolerant with no phenotypic symptoms but intermediate serological reactions. The three tolerant lines all belong to the Chinese Long type. The selected tolerant lines were crossed with susceptible ones to develop segregating populations, which together with previously developed recombinant inbred line (RIL) population were subjected to screen for phenotypic reactions. With the systemic virus infection, the phenotype scoring was not that straightforward. We used 0, 1, 2, and 3 to rate and calculate disease severity index (DSI) (Figure 7). The resistance class with the rating 0 showed no visible mottle or mosaic symptom on leaves and the rating 1 had mild mosaic and plant recovery. The susceptible class included inoculated plants with the rating 2 with mosaic and leaf deformed and the rating 3 on plant stunting, leaf deformation or dying plant (Figure 8).
Results in the screening of two F2 populations (WI7182F2 and WI7814F2) developed between the resistant and susceptible parents showed segregation ratios that seem to be consistent with a simply inherited gene underlying the resistance in both populations (Table 1). We also tested 20 RILs (WI7326 population) derived from the cross of PI 197088 × WI7156 which showed segregation to CGMMV inoculation (Figure 9). Test of CGMMV inoculation responses on more RILs from this population is underway. Bulked segregant analysis (BSA)-Seq will be performed with data from the two F2 populations aiming to map the resistance loci to a sub-chromosomal region.

Figure 7. Rating classes of cucumber infected by CGMMV, rating 0: no symptom, rating 1: mild mosaic symptom, plant recovery; rating 3: severe mottle mosaic and rating 4: severe mottling and plant stunting.

Table 1. Segregating of CGMMV inoculation responses in two F2 populations of cucumber.

# Plants under each rating
Rating scale 0 1 2 3 Total
WI7182F2 0 22 88 1 111
WI7814F2 0 18 94 0 112
Figure 8. Distribution of individuals two F2 populations (WI7182F2 and WI7814F2) segregating for resistance to CGMMV based on disease severity index rating (R in rations 0 and 1, S in ratings 2 and 3).
Figure 9. Disease index (DSI, Y-axis) of CGMMV inoculation responses among 20 RILs from PI 197088 × WI7156 cross.

Marker development and verification | Downy mildew (YW & AK)

We proposed to conduct fine mapping of the major-effect DM QTL, dm4.1, and dm5.3, and introgress them into different genetic backgrounds through marker-assisted QTL pyramiding. For fine mapping, we have developed near isogenic lines (NILs) for dm4.1 and dm5.3 in two backgrounds: the Chinese Long inbred line 9930 and the US pickling cucumber line Gy14 that also carries dm1 (CsSGR). Genotyping and extensive phenotyping of these NILs and recombinants from NIL-derived F2 and backcross progeny revealed four sub-QTL at the dm4.1 locus that are present in both WI7120 and PI 197088 including dm4.1.1, dm4.1.2A, dm4.1.2B, and dm4.1.3. DM resistance of NILs with different combinations of sub-QTL was evaluated in both growth chamber (Figure 10) and field conditions.

Figure 10. Performance of inoculation responses by DM pathogen (P. cubensis) on NILs carrying combinations of different sub-QTL (dm4.1.1, dm4.1.2A, dm4.1.2B, and dm4.1.3) at the dm4.1 locus of WI7120. DM resistance was scored by four criteria: anti-chlorosis (Yel), anti-necrosis (Nec), antisporulation (Spor), and general impression (GI), which is a composite trait including Ye, Nec and Spor.

The candidate genes for dm4.1.2A and dm4.1.3 in PI 197088 have been identified by another group (Berg et al. 2020, 2021). We examined allelic diversity in the dm4.1 region between WI7120 and PI 197088 and did not find any DNA sequence variation indicating both lines have the same resistance alleles at the four sub-QTL. We focused on fine mapping and cloning of dm4.1.2B using NIL-derived segregating populations. Phenotypic characterization of DM resistance of the NILs in both field and growth chambers revealed anti-chlorosis nature of dm4.1.2B (Figure 11).

Figure 11. Anti-chlorosis effect conferred by dm4.1.2B in response to P. cubensis infection. At both 7 and 9 days post infection (dpi), the resistant RIL (NIL-R, A) shows less chlorosis than NIL-S (B) with significantly lower mean yellowing (Yel) scores (C and D).

We conducted fine genetic mapping of dm4.1.2B locus from WI7120 which delimited dm4.1.2B into a 36.2 kb region on Chromosome 4 (Figure 12) with four predicted genes. Confirmation of the candidate gene for dm4.1.2B is underway.

Figure 12. Map-based cloning of the dm4.1.2B sub-QTL. Eight recombinants in NIL-F2 defined by 11 markers loci which delimit dm4.1.2B into a 36.3 kb interval on Chr 4. R (red), S (blue), and H (gray) indicates resistant, susceptible and heterozygous (segregating) alleles, respectively.

For map-based cloning of dm5.3 in PI 197088, NILs were developed for this major-effect QTL (Figure 13). GBS of the NILs suggests uniform genetic background (9930, the susceptible recipient of resistance allele from PI 197088, orange color in Figure 13A) except the dm5.3 region that has ~5.24 Mb introgression from the donor line (WI7088D, green color in Figure 13A). Disease screening of NILs and relevant lines revealed moderate DM resistance contributed by dm5.3 with both anti-chlorosis and anti-necrosis effects.

Figure 13. Development and characterization of DM resistance of NILs for dm5.3. The NIL-R carries ~5.24Mbp fragment from PI 197088 (A), which confers both anti-chlorosis and anti-necrosis in 9930 background (B and C).

Fine genetic mapping with recombinants from NIL-derived segregating populations allowed to narrow down the dm5.3 locus into 144 kb region on chromosome 5 (Figure 14). Multiple lines of evidence support CsSIB1 (cucumber sigla factor binding protein1) as the most possible candidate gene.

Figure 14 Fine mapping of dm5.3 locus. A. Haplotypes of critical recombinants and DM ratings. B. Bar graph showing mean DM scores of recombinants. C. dm5.3 was narrowed down to a 144kb region containing 12 predicted genes. D. Multiple lines of evidence suggests CsSIB1 as the most possible candidate for dm5.3. The SNP between 7088D and 9930/Coolgreen is in the exon. E. Relative expression of CsSIB1 in NIL-R and NIL-S at 0, 3 dpi with qPCR.

Marker development and verification | Phytophthora fruit rot (RG and Ying-Chen Lin)

The qPFR5.1 initially identified by QTL-seq spanned ~5 Mb on chromosome 5. To facilitate fine mapping, we used a recombinant inbred line (RIL) population and F3 families selected for enriching recombination within the region for fine mapping. The prior tested RIL populations refined the region to 1.31 Mb, located between markers M26 and M5 (27.08-28.39 Mb). An additional set of F3 families was tested in 2022; 99 individuals from 33 families that were homozygous recombinant in the qPFR5.1 region (A4-3–Gy14 or Gy14 – A43) were grown in the greenhouse. Cuttings also were taken from each plant and transplanted to the field. The results from F3 families suggested an overlapping but shifted region that varied somewhat between the greenhouse and field results [M26 – M5 (27.08Mb – 28.39Mb) in the greenhouse and M2-M26 (25.17- 27.08) in the field)] (Figure 15). Additional markers between M2 and M3 and M28 and M5 are being tested.

Figure 15. Fine mapping of the qPFR5.1 QTL in recombinant F3 families. Color bars refer to different genotypes: dark grey – DH A4-3 allele; light grey – Gy 14 allele. Number on the left indicates the name of the lines/families sharing the same genotypes, and the letters on the right indicates phenotypes of each genotype: r – resistant lines with score < 4.5; s – susceptible lines with score > 6.0; and i – intermediate lines with scores between 4.5 to 6.0.

QTL introgression into breeding or advanced lines, and release to breeders (YW, RG and AK)

During fine mapping of dm4.1 and dm5.3, plants carrying different combinations of dm4.1, dm5.2 (both from WI7120) and dm5.3 (from PI 197088) resistance alleles were identified and backcrossed with Gy14 aiming to develop Gy14 carrying all permutations of the three QTL. In 2022, homozygous introgression lines carrying all three QTL (Gy14Q3) were obtained. DM resistance of Gy14Q3 was tested in both growth chamber and field trials at multiple locations (South Carolina, North Carolina, Michigan, Wisconsin). The three QTL were also introgressed into the Chinese Long (9930) and beit alpha (WI7204) backgrounds, which revealed some background effects on DM resistance conferred by the three QTL.

We also aim to develop inbred lines with both DM and PFR resistances through markerassisted QTL pyramiding. In 2021, a plant carrying homozygous qPFR5.1 QTL for PFR resistance (from PI 109483) was crossed with Gy14Q3. The resulting F1 plant carrying all four QTL (dm4.1, dm5.2, dm5.3, and qPFR5.1) was further backcrossed with Gy14 to advance to BC1, which were subjected to marker-assisted selection. Since dm5.2-qPFR5.1-dm5.3 were located in a ~9 Mbp block in repulsive phase on cucumber Chromosome 5, ideal recombinants combined with expected alleles at three loci were not identified so far. In 2022, we revised the strategy by identifying recombinants between dm5.1 and 1PFR5.1 first, then recombinants between qPFR5.1 and dm5.3 in the next BC generation. With this strategy, we successfully identified recombinants carrying resistance alleles for all four resistance loci (dm4.1, dm5.2, qPFR5.1, and dm5.3). Notably, the dm4.1 locus contains all four sub-QTL from WI7120. In 2022 summer filed trail at Hancock, Wisconsin, the Gy14 plants carrying three homozygous DM QTL but heterozygous at the qPFR5.1 locus (Gy14Q4) exhibited better resistance than Gy14 for natural epidemic of DM. Plants in the F2 population that was segregating for the qPFR5.1 locus were tested in the field in Michigan in 2022 and young fruit examined for PFR inoculation responses. Plants with the qPRF5.1 allele showed intermediate resistance to the resistant and susceptible parents, respectively (Figure 16).

Figure 16. Values shown are P. capsici disease ratings for young fruit from: A4-3; F2 family containing PFR5.1, DM5.2, and DM5.3 in Gy14 background; and Gy14.