Molecular Breeding

, Volume 20, Issue 4, pp 359–374

Genetic mapping and localization of quantitative trait loci affecting fungal disease resistance and leaf morphology in grapevine (Vitis vinifera L)

Authors

  • Leocir J. Welter
    • Federal Centre for Breeding Research on Cultivated PlantsInstitute for Grapevine Breeding Geilweilerhof
  • Nilgün Göktürk-Baydar
    • Federal Centre for Breeding Research on Cultivated PlantsInstitute for Grapevine Breeding Geilweilerhof
    • Faculty of Agriculture, Department of HorticultureSüleyman Demirel University
  • Murat Akkurt
    • Federal Centre for Breeding Research on Cultivated PlantsInstitute for Grapevine Breeding Geilweilerhof
    • Faculty of Agriculture, Department of Horticulture Ankara University
  • Erika Maul
    • Federal Centre for Breeding Research on Cultivated PlantsInstitute for Grapevine Breeding Geilweilerhof
  • Rudolf Eibach
    • Federal Centre for Breeding Research on Cultivated PlantsInstitute for Grapevine Breeding Geilweilerhof
  • Reinhard Töpfer
    • Federal Centre for Breeding Research on Cultivated PlantsInstitute for Grapevine Breeding Geilweilerhof
    • Federal Centre for Breeding Research on Cultivated PlantsInstitute for Grapevine Breeding Geilweilerhof
Article

DOI: 10.1007/s11032-007-9097-7

Cite this article as:
Welter, L.J., Göktürk-Baydar, N., Akkurt, M. et al. Mol Breeding (2007) 20: 359. doi:10.1007/s11032-007-9097-7

Abstract

The aim of this study was the improvement of a genetic map of the F1 population from the cross between the fungus-resistant grapevine cv. “Regent” and the susceptible cv. “Lemberger” and its use to localize factors affecting pathogen resistance and leaf morphology. To construct an integrated map combining the information from both parental meiotic recombination frequencies co-dominant microsatellite markers were employed. Resistance gene analog (RGA)-derived and sequence characterized amplified regions (SCAR) markers correlated with powdery and downy mildew resistance were additionally mapped. The new integrated map contains 398 markers aligned along 19 linkage groups, covers a total length of 1,631 cM and shows an average distance between markers of 4.67 cM. One hundred and twenty-two microsatellite markers were newly mapped. This genetic map was used to localize QTLs (quantitative trait loci) conferring resistance to powdery and downy mildew pathogens transmitted from “Regent”. Factors influencing specific leaf morphology traits were identified in addition. A major QTL for powdery mildew resistance and one major and one minor QTL for downy mildew resistance were detected. Some RGA-derived markers are found co-located in the region covered by the major QTL for resistance to downy mildew hinting at their putative functional relevance. Furthermore, 27 QTLs affecting leaf morphology descriptors were identified. This map is an important tool for grapevine breeding and resistance research.

Keywords

Erysiphe (syn. Uncinula) necatorGenetic mappingLeaf morphologyPathogen resistancePlasmopora viticolaQTL analysisResistance gene analogs

Introduction

Development of genetic maps as tools for grapevine breeding was retarded as compared to other crops due to high heterozygosity levels and inbreeding depression. The first genetic map for grapevine was published in 1995 (Lodhi et al. 1995). Nine genetic mapping studies have been published meanwhile (Dalbó et al. 2000; Doligez et al. 2002; Grando et al. 2003; Fischer et al. 2004; Doucleff et al. 2004; Adam-Blondon et al. 2004; Riaz et al. 2004; Doligez et al. 2006; Lowe and Walker 2006). The first maps were mainly based on dominant random amplified polymorphic DNA (RAPD) and amplified fragment length polymorphism (AFLP) markers (Dalbó et al. 2000; Doligez et al. 2002; Grando et al. 2003; Fischer et al. 2004; Doucleff et al. 2004). These kinds of markers allow a rapid generation of maps but do not easily permit the transfer of information between different genotypes and comparison between maps (Adam-Blondon et al. 2004). To generate a set of codominant markers for grapevine genetics, 21 research groups formed an international consortium to identify microsatellite-based markers for grapevine (VMC, Vitis Microsatellite Consortium, coordinated by AGROGENE, Moissy Cremayel, France). More than 350 VMC-markers were developed and used for construction of the internationally approved reference linkage map from the cross of “Riesling” × “Cabernet Sauvignon” (Riaz et al. 2004). Two additional sets of microsatellite-based markers for grapes were recently developed in Italy (UDV, Di Gaspero et al. 2005) and in France (VVI, Merdinoglu et al. 2005). A second microsatellite-based linkage map using the VMC set together with the VVI markers yielded the most elaborate published microsatellite-based map for grapevine currently available (Adam-Blondon et al. 2004).

The use of PCR primers with 20–30 nucleotides in length flanking the microsatellite loci usually leads to amplification of a unique, specific locus. The conservation of microsatellite-flanking sequences within cultivars and closely related species frequently allows one to transfer results between different mapping studies. It enables one to compare the linear order of markers between maps obtained from different Vitis vinifera cultivars or Vitis species genotypes. It permits one to search for coinciding location of important traits evaluated in different mapping studies using genotypes with divergent genetic backgrounds. This approach can elucidate main factors involved in selection processes.

Considering the grapevine genome sequencing programs operating currently in France and Italy genetic maps saturated with transferable markers will be instrumental to utilize the genomic information from model genotypes to target genomic regions for comparative analysis in non-model genotypes contributing important traits for breeding such as resistance to pathogens.

In highly heterozygous crops like grapevine a strategy termed “double pseudo-testcross” is employed to generate genetic maps (Grattapaglia and Sederoff 1994). This approach yields two separated genetic maps, one for each parent. The implementation of codominant markers like microsatellites allows their combination into one single integrated map. Furthermore, a consensus map can be generated by integrating information from different mapping populations (Doligez et al. 2006). The use of microsatellite-based markers is an important strategy to achieve these aims.

The principal aim of the present study was the construction of an integrated grapevine map by introduction of microsatellite markers into the previously elaborated map (Fischer et al. 2004) derived from the cross between “Regent” and “Lemberger”. Additionally, 14 RGA (resistance gene analogs)-based markers and three SCAR (sequence characterized amplified regions)-markers correlated with resistance to powdery and downy mildew (Akkurt et al. 2006; Akkurt et al., in prep.) were analyzed. The resulting map was employed for QTL analysis of powdery and downy mildew resistance and leaf morphology traits. The latter have been investigated as a first example of localizing morphogenetic regulatory factors in grapevine.

Material and methods

Mapping population

The mapping population consisted of 144 F1 plants obtained by the cross of the red wine cultivars “Regent” and “Lemberger.” “Regent” was bred at the Institute for Grapevine Breeding Geilweilerhof and shows resistance to both powdery (Erysiphenecator) and downy mildew (Plasmopora viticola) (Anonymous 2000). “Lemberger” is a traditional fungus-susceptible Vitis vinifera cultivar. The population segregates also for other agronomical and morphological traits. Young leaves of the population (second and third insertion from the apices) were collected at the start of the vegetation cycle and stored after shock freezing with liquid nitrogen at −70°C until DNA extraction. DNA extraction was performed according to the protocol described by Thomas et al. (1993).

Phenotypic evaluation of resistance traits

The mapping population was scored for resistance to powdery and downy mildew in five (1999, 2000, 2003, 2004 and 2005), respectively, four (1999, 2000, 2003, 2004) growing seasons. The evaluations were performed at the Institute for Grapevine Breeding Geilweilerhof, omitting any fungicide protection, under natural field infection conditions. Additionally, a replicate of the mapping population was evaluated in 1999 for downy mildew resistance after experimental infection under greenhouse conditions at INRA Colmar. The degree of resistance to powdery and downy mildew was evaluated independently on leaves and berries or clusters, as described by Fischer et al. (2004).

Phenotypic determination of leaf morphology characteristics

From 129 genotypes of the mapping population 10 adult leaves from the middle third of several shoots were collected in summer 2005, pressed and dried. Eighteen ampelometric leaf characteristics (Genres 081 2001; OIV 2007) (Table 1) were recorded by using a digitiser tablet (SummaSketch II Professional Plus) (Fig. 1). Mean values were calculated for each leaf characteristic. The following ratios were calculated: length of vein N3/length of vein N1 (OIV 603/OIV 601), length petiole sinus to upper leaf sinus/length of vein N2 (OIV 605/OIV 602), length of petiole sinus to lower leaf sinus/length of vein N3 (OIV 606/OIV 603), length of tooth of N2/width of tooth of N2 (OIV 612/OIV 613), length of tooth of N4 / width of tooth of N4 (OIV 614/OIV 615) and length of vein N5/length of vein N1 (OIV 611/OIV 601).
Table 1

List of leaf morphology characteristics scored as OIV (Organization Internationale de la Vigne et du Vin, International Vine and Wine Organization, Paris; OIV 2007) descriptors

OIV code

Morphological traits of mature leaves

601

Length of vein N1

602

Length of vein N2

603

Length of vein N3

604

Length of vein N4

605

Length petiole sinus to upper leaf sinus

606

Length petiole sinus to lower leaf sinus

607

Angle between N1 and N2 measured at the first ramification

608

Angle between N2 and N3 measured at the first ramification

609

Angle between N3 and N4

610

Angle between N3 and the tangent between petiole point and the tooth tip of N5

611

Length of vein N5

612

Length of tooth N2

613

Width of tooth N2

614

Length of tooth N4

615

Width of tooth N4

617

Length between the tooth tip of N2 and the tooth tip of the first secondary vein of N2

618

Opening/overlapping of petiole sinus

665*

Vein N3, length petiole sinus to vein N4

* Descriptor 665 is not an official OIV descriptor but has been developed by Dettweiler (1987)

https://static-content.springer.com/image/art%3A10.1007%2Fs11032-007-9097-7/MediaObjects/11032_2007_9097_Fig1_HTML.gif
Fig. 1

Schematic representation of leaf ampelometric measures according to OIV 2007 as described in Table 1

Genotyping

The earlier genetic maps of “Regent” and “Lemberger” were based mostly on randomly amplified polymorphism DNA (RAPD) and amplified fragment-length polymorphism (AFLP) markers as described by Fischer et al. (2004). These maps were improved and combined by newly mapping 122 microsatellite loci, three sequence characterized amplified regions (SCARs) and 12 resistance gene analogs (RGA)-derived markers.

The primer pairs flanking microsatellite loci originated from different marker sets: VVS (Thomas and Scott 1993), VVMD (Bowers et al. 1996; 1999), VrZAG (Sefc et al. 1999), VMC (Vitis Microsatellite Consortium), UDV (Di Gaspero et al. 2005) and VVI (Merdinoglu et al. 2005). Primers developed by the Institute for Grapevine Breeding Geilweilerhof within the VMC are listed in Table 2. All “forward” primers were labeled at their 5′-ends with fluorescent dyes (Ned, Hex or Fam) and the PCR products were analyzed by capillary electrophoresis using the ABI 3100 Genetic Analyzer (Applied Biosystems, Foster City, California/USA). Two hundred and twelve primer pairs were first tested for segregating polymorphism using the genitors and 14 randomly picked genotypes from the progeny. After this screening, informative primer pairs were combined into multiplex applications for PCR considering the fluorescence labels, annealing temperatures and lengths of the amplified products. The markers were amplified in standard reactions of 10 μl final volume containing 8.0 ng template DNA, 1× NH4Taq buffer (Invitek, Berlin/Germany), 1.5 mM or 2.5 mM MgCl2, 0.2 mM of each dNTP (Sigma, Taufkirchen/Germany), 0.2 μM of each primer and 0.3 U Taq DNA Polymerase (Invitek, Berlin/Germany). Amplification was performed in GeneAmp PCR System 9700 thermocyclers (Applied Biosystems, Foster City, California/USA), using the following program: 94°C for 3 min; 30 cycles of 94°C for 1 min, 48–65°C annealing for 1 min (depending on the primer pair sequences) and 72°C for 2 min and finally 72°C for 20 min.
Table 2

Primer sequences flanking microsatellite loci developed by the Institute for Grapevine Breeding Geilweilerhof within the VMC consortium

Primer name

 

Primer sequence (5′ > 3′)

Accna

VMC1a7

Forward

ACGACCGGCAGAACAACAGT

BV681752

Reverse

GGGCCAAACCTCTAAAAGCA

 

VMC1a12

Forward

ATGTAATTACCGGTCATGAGTT

BV681753

Reverse

TTCTTGTTTTGCCTATCTATCC

 

VMC1b11

Forward

CTTTGAAAATTCCTTCCGGGTT0

BV681754

Reverse

TATTCAAAGCCACCCGTTCTCT

 

VMC1b12

Forward

AGGTGCTCCAGCCAGTCAG

Reverse

CCCCTAATGCTCCGTGTTC

 

VMC1c10

Forward

CACAGCTGTTCCAAGTCCCA

BV681755

Reverse

ACAAGCCTTCCGCCACTCTC

 

VMC1d10

Forward

CAGGTGTCCAGGACATATAAGG

BV681756

Reverse

TTGGTTGGAATCTTGTAGAGGG

 

VMC1d11

Forward

CTGCATGCTCATTGTACTATCA

BV681757

Reverse

AGTGTCTTCTCGTCTTAAAACCT

 

VMC1e8

Forward

CAGCGAGCTCTTGATTTATTGT

BV681758

Reverse

GATCATAGCTTCAACGGCTTTT

 

VMCe11

Forward

GGGGTCCAATGTGGACTTTATC

BV681759

Reverse

CCATGAACAACAAACATGGCTT

 

VMC1e12

Forward

GTGTGACCTTATGCAACACCAA

BV681760

Reverse

GCTACCACATGCAGACAGGTTAGT

 

VMC1f10

Forward

CATACAAGGAATTTACCCCCA

BV681761

Reverse

ACCTCTTGTGCTGTCTAACCA

 

VMC1f12

Forward

AAACCTTTCTGATGGTATCTAA

BV681762

Reverse

GCTCATTGTAACATCAAAACTT

 

VMC1g7

Forward

GGGTCCACATAGGTAGGAGATT

BV681763

Reverse

AGCCCATAAAGGCCTTAAAAAC

 

VMC1h9.1

Forward

ACAAGCTCCTACCGGTTCCAA

BV681764

Reverse

TTCTGTGGCAATGGGGTAGTTC

 

VMC1h11.1

Forward

TGGGTTACTTCAGGAGACAAAA

Reverse

ACAACATAATTGGCCTCCACAT

 

VMC_NG4b9

Forward

CTGGGGAGCATATACACATACCAG

BV681765

Reverse

CTCTCTCTTCCCGATAGCCACC

 

VMC_NG4c8

Forward

CGAGAATCACCGGCGAA

BV681766

Reverse

TGCAGCGCGGAGCA

 

VMC_NG4c10

Forward

AAGCAATGAACACAACATTCTCC

BV681767

Reverse

CTAAGTTTCTATGACACTTTCCTCCA

 

VMC_NG4d10.1

Forward

AGGGGGAGACGCACGAA

BV681768

Reverse

GCGCAGCCTTTGCCAGA

 

VMC_NG4e9

Forward

AGAGACAGGGAGAGGAGAGT

BV681769

Reverse

TGGGAAATGCAAACAGAG

 

VMC_NG4e10.1

Forward

AATGCAGCAGCGCCAGATG

BV681770

Reverse

GCAGGCTGCTGCTGTTTTG

 

VMC_NG4f9.2

Forward

GGGGAGAGTGGAGTGGAGT

BV681770

Reverse

TCCTCCATGTCCCTCTGCT

 

VMC_NG4h9

Forward

GATCTGCCCGCAAATACCG

BV681772

Reverse

GGGACGAGGACGTGAGTGT

 

a NCBI GenBank accession number

A set of RGA-derived primers developed by Di Gaspero and Cipriani (2002, 2003) was tested for polymorphisms segregating in the present cross according to the original PCR protocol. Four of them showed clearly scorable polymorphic bands and were used to genotype the mapping population. The SSCP (single-strand conformational polymorphism) analysis was performed as described (Schneider et al. 1999). The three SCAR-markers employed are converted RAPD markers correlated with resistance traits (Akkurt et al. 2007; Akkurt et al. in prep.).

Genetic mapping and QTL analysis

The “double pseudo-testcross strategy” (Grattapaglia and Sederoff 1994) was used for the construction of the genetic map. The implementation of codominant microsatellite-derived markers allowed the combination of both parental maps into one map.

Marker segregation was tested with regard to the goodness-of-fit to the expected ratio using the x2 test. Markers showing distorted segregation were originally included in the map calculation. After a preliminary mapping, individual distorted markers located in regions surrounded by non-distorted markers were excluded and the map was re-calculated.

The genotypic information was subjected to genetic mapping through linkage and recombination analysis with JoinMap 3.0 software (Van Ooijen and Voorrips 2001), applying the Kosambi function for the estimation of map distances (Kosambi 1944). LOD (logarithm of the odds) score thresholds equal or greater than 6.0 were used to determine linkage groups. The maximal recombination fraction permitted was 0.4.

Putative QTLs were primarily identified by interval mapping (Lander and Botstein 1989; Young 1996). Subsequently, molecular markers coinciding or closely flanking the LOD maxima of QTLs were used as co-factors in multiple QTL analysis (restricted MQM and full MQM mapping). The linkage group specific and genome wide significance thresholds of QTL LOD scores were determined by permutation tests (1.000 permutations, P ≥ 0.05) of the quantitative trait data (Churchill and Doerge, 1994). All these calculations employed MapQTL 4.0 software (Van Ooijen et al. 2000).

Results

The integrated map

The combination of new microsatellite marker information with previously generated, mostly dominant marker data, allowed the construction of an integrated map for the cultivars “Regent” and “Lemberger”. In total, 398 markers were aligned along 19 linkage groups (LG), covering 1,631 cM with an average distance between markers of 4.67 cM (Table 3; Fig. 2). The LGs were numbered following the nomenclature of the International Grape Genome Program (IGGP) (Riaz et al. 2004; Adam-Blondon et al. 2004). In this presentation 122 microsatellites markers were mapped. Seventy-six (60%) of these were heterozygous in both “Regent” and “Lemberger,” and 67 (53%) of them could be mapped as co-dominant markers. The nine remaining markers were analyzed as dominant markers due to the presence of “null” alleles. “Regent” and “Lemberger” were heterozygous for 89% and 71 % of the mapped microsatellite markers, respectively (Fig. 3)
Table 3

Summary of the information generated for the integrated map “Regent” × “Lemberger”

LGsa

Length (cM)

No. of markers

Microsatellite markers

RGA/SCAR markers

Average distance

1

94

26

7

stkVr001

3.62

2

90

14

6

6.43

3

58

13

7

4.46

4

79

15

7

5.27

5

88

28

11

3.14

6

93

13

6

7.15

7

84

20

5

rgVrip158

4.20

8

80

25

9

3.20

9

42

5

1

8.40

10

66

8

6

8.25

11

100

22

7

4.55

12

118

33

7

stkVa011

3.58

13

77

21

7

3.67

14

119

18

6

6.61

15

76

25

4

ScORA7, ScORN3R

3.04

16

86

31

5

2.77

17

90

23

4

3.91

18

102

33

9

rgVamu137, ScPRA14

3.10

19

89

25

8

3.56

Total

1631

398

122

4.67

a Linkage groups numbered according to the International Grape Genome Program (IGGP) nomenclature

https://static-content.springer.com/image/art%3A10.1007%2Fs11032-007-9097-7/MediaObjects/11032_2007_9097_Fig2a_HTML.gif
https://static-content.springer.com/image/art%3A10.1007%2Fs11032-007-9097-7/MediaObjects/11032_2007_9097_Fig2b_HTML.gif
Fig. 2

Integrated “Regent” × “Lemberger” genetic map. Linkage groups are numbered according to the International Grape Genome Program (IGGP) nomenclature. The genetic map was constructed employing JoinMap 3.0 software with LOD ≥ 6.0 for linkage. The cumulative distance between markers in cM, calculated with the Kosambi function, is indicated on the left of the linkage groups. Microsatellite loci and RGA-derived markers or ScOR markers are represented with light grey and grey boxes, respectively. Distorted microsatellite markers are indicated by asterisk(s) according the level of distortion (*P ≤ 0.05; ** P ≤ 0.01; *** P ≤ 0.001)

https://static-content.springer.com/image/art%3A10.1007%2Fs11032-007-9097-7/MediaObjects/11032_2007_9097_Fig3_HTML.gif
Fig. 3

Frequency distribution (%) of the mapping population plants based on the level of resistance expressed to Plasmopara viticola and Erysiphe (syn. Uncinula) necator in different years. The level of resistance was evaluated under field conditions employing the OIV classification

Distortion of segregation

Distortion of segregation is a commonly observed phenomenon in genetic mapping studies. Markers showing distorted segregation were hence included in the mapping as described in the methods section. Clusters of distorted markers (two or more) were identified in ten linkage groups as indicated in Fig. 2. Microsatellite markers located in five of these regions have also been found to be distorted in the map constructed from “Syrah” × “Grenache” (Adam-Blondon et al. 2004). Interestingly, one of these regions co-locates with the major QTL for resistance to powdery mildew (LG-15).

To promote a better understanding of the events involved in the genetically distorted regions, the x2 test was applied to test the goodness-of-fit of the observed gametic (test 1) and zygotic (test 2) segregation with regard to the their expected segregation (methodology adapted from Lorieux et al. 1995). For this analysis only fully informative microsatellite markers (heterozygous in both parents) mapped to distorted regions were used. Distorted markers located in seven linkage groups were analyzed. The gametic segregation was tested independently for each parent, considering that an F1 population was used for the map construction. A region was considered to be under the gametic selection if only test 1 was significant, while the zygotic selection was assumed when tests 1 and 2 or only test 2 were significant (P ≤ 0.05).

The x2 tests showed that different phenomena were involved in the deviation of segregation ratios (data not shown). The microsatellite markers located on the distorted region of LG-01, 05 and 10 were significant only for the test 1 in “Lemberger,” suggesting a gametic selection in favor of one of the paternal alleles (pollen). Otherwise, marker VVIn73, located on LG-17, showed a gametic selection in favor of one of the maternal alleles (“Regent”). The microsatellite markers located in the distorted region of LG-15 and LG-19 present gametic selection in both parents. Finally, marker VMC_6 g1, located in the distorted region of LG-11, was significantly distorted in both tests, suggesting zygotic selection.

Fungal disease resistance QTLs

Use of interval mapping allowed the identification of QTLs conferring resistance to powdery mildew, respectively, downy mildew. The molecular markers with the highest LOD-value present in these regions were used in the second step in either restricted MQM (for downy mildew) or full MQM (for powdery mildew) mapping to increase the resolution of QTL analysis. Nevertheless only one major QTL conferring resistance to powdery mildew could be detected. This QTL, located on LG 15, shows significant effects on both the resistance of leaves (1999, 2000 and 2005) and berries (1999, 2000) and explains up to 56.8% resp. 64.5% of the phenotypic variance. In the years 2003 and 2004 no significant QTL could be found.

In contrast, one QTL showing major and one showing minor effects on the resistance to downy mildew were identified. The major QTL is located on LG 18 and was detected in three different scores of the trait: leaf resistance, berry/cluster resistance and lesion size. This QTL was very stable over the years, showing significant effects in all years evaluated, except for berry resistance data scored in 2004 (Table 4). The QTL seems to be spread over a large region in this linkage group. The markers with the maximal LOD value explains up to 37.3% of the phenotypic variance, detected in 2000 for leaf resistance (Table 4). The position of the QTL peak (maximal LOD value) identified by interval mapping fluctuated over the years. The selection of the molecular marker with the greatest LOD value as co-factor during the restricted MQM mapping did not reduce the extension covered by the QTL. More than one resistance factor seems to be located in this genomic region. The minor QTL is located on LG 4. Despite its small effect, the QTL could be detected in the 4 years evaluated, explaining up to 22.6% of the phenotypic variance (Table 4). The QTL located on LG 5 identified previously by Fischer et al. (2004) was significant only in 1999 in both field data and greenhouse assays (data not shown).
Table 4

Description of the quantitative trait loci (QTLs) detected for resistance to Plasmopara viticola and Erysiphe (syn. Uncinula) necator in different years

Trait

Year

LG a

QTL LOD max b

LOD threshold specific LGc

LOD threshold genome wide c

Map position LOD max

Confidence Interval d

Flanking marker e

% Var. expl.

P. viticola leaf resistance (OIV-452)

1999

18

15.27

3.3

4.9

71.2 cM

66.8-93.5

M21300

34.9

4

5.80

2.8

4.9

25.1 cM

22.3-30.8

VMC7h3

15.2

1999*

18

10.62

3.3

5.4

71.2 cM

66.8–101.8

M21300

31.3

4

4.75

2.9

5.4

25.1 cM

22.3–35.8

VMC7h3

9.0

2000

18

17.56

3.3

5

86.8 cM

66.8–101.8

UDV112

37.3

2004

18

6.47

3.4

4.6

74.0 cM

66.4–88.2

M19940

15.6

4

5.38

3

4.6

30.8 cM

22.3–38.8

VMCNg2e1

12.5

P. viticola leaf resistance (size of necrose)

1999

18

12.04

3.4

5.6

71.2 cM

66.8–101.8

M21300

21.9

4

6.82

2.9

5.6

25.1 cM

22.3–35.8

VMC7h3

10.4

1999*

18

11.23

3.4

5.6

71.2 cM

66.8–84.2

M21300

36.1

4

5.61

2.9

5.6

30.8 cM

22.3–44.4

VMCNg2e1

22.6

2000

18

14.53

3.4

4.6

84.2 cM

66.8–101.8

A14500

37.2

2004

18

5.22

3.3

84.2 cM

66.8–96.8

A14500

17.7

P. viticola berry resistance (OIV-453)

1999

18

9.04

3.2

5.1

93.5 cM

66.8–101.8

M131220

29.8

1999*

18

10.0

3.3

5

93.5 cM

75.7–101.8

M131220

33.4

2000

18

10.86

3.3

7.7

84.2 cM

66.8–101.8

A14500

30.2

E. necator leaf resistance (OIV-455)

1999

15

20.1

3.1

6.5

47.1 cM

44.5–47.1

M121020

56.8

2000

15

16.98

43.5 cM

38.6–43.5

R1070

42.1

2005

15

3.84

3.1

4.5

66.7 cM

58.1–71.7

ScORA7

13.7

E. necator berry resistance (OIV-456)

1999

15

13.92

3

6.4

51.7 cM

48.1–51.7

N61770

64.5

2000

15

7.44

3

5

43.5 cM

43.5

R1070

21.9

a Linkage groups named as the International Grape Genome Program (IGGP) nomenclature

b Maximum LOD score obtained by restricted MQM (P. viticola) and full MQM (E. necator) mapping

c Calculated by permutation test at P ≤ 0.05

d The confidence interval was determined by dividing the highest LOD score of the QTL by two

e Nearest molecular markers to the peaks of the QTLs

* Phenotypical evaluation performed under greenhouse conditions at INRA Colmar

Position of RGA-derived markers

The RGA-STSs primer pairs used to screen the mapping population revealed rather complex patterns of bands in SSCP-analysis. Each segregating band was scored individually as a dominant marker. In total, 12 polymorphic bands amplified with four RGA-STSs primer pairs were recorded and mapped. The polymorphic bands obtained from one primer pair mapped in the same linkage group, generally closely linked to each other (Fig. 2). The RGA-derived markers were coded according to the original description by Di Gaspero and Cipriani (2003), followed by a letter specifying the scored band. Interestingly, two markers amplified with primer pair “rgVamu137” mapped on LG 18, in the region covered by the major QTL for resistance to downy mildew. Other RGA-derived markers were found on LGs 1, 7 and 12.

Leaf morphology QTLs

Leaf morphology traits are discriminant parameters in ampelography contributing to the identification of unknown cultivars (Genres 081 2001). Their variation is most likely caused by subtle differences in morphogenetic factors shaping the developing leaves. Selected leaf traits serve as internationally approved grapevine descriptors (OIV 2007) and their variation is considered to be cultivar-specific (Dettweiler 1987). Hence they must rely predominantly on genetic factors determining their phenotypic expression.

The “Regent” × “Lemberger” progeny shows considerable variation in leaf morphology. This population was chosen as a model to study the genetics of morphogenetic factors operating in grapevine. A selection of 18 internationally acknowledged and well defined leaf characteristics (Table 1) scored in summer 2005 and six different calculated trait ratios were analyzed in QTL mapping. For 13 out of 18 traits and four out of six calculated trait ratios QTLs exceeding the genome wide LOD significance thresholds could be identified (Table 5). The QTLs were found dispersed all around the genome on 12 of the 19 LGs corresponding to the grapevine chromosomes. Most LGs carried one or two different leaf trait QTLs with the noticeable exception of LG-01 that appears to carry factors affecting eight different morphological traits. Two of these, OIV 605 and 606, as well as the related trait ratio OIV 605/OIV 602 even exhibit their LOD maximum in the very same position at 13.6 cM. Most of these traits concern leaf sinus formation and thus may be regulated by a common factor located at this genetic position. The leaf trait with the most dispersed QTLs was OIV 607 (the angle between leaf veins N1 and N2 at first ramification) affected by QTLs on five different linkage groups (LG-01, 06, 12, 13 and 15) with maximal LOD scores of five to eight.
Table 5

Description of the quantitative trait loci (QTLs) detected for morphological leaf traits

OIV code

LGa

QTL LOD maxb

LOD threshold threshold specific LGc

LOD threshold genome LGc

Map position LOD max

2- LOD support interval

Flanking markerd

% Var. expl.

601

LG-05

4.63

3.3

4.4

27.2 cM

21.9–28.5 cM

R06341

16.6

603/601

LG-11

5.51

3.0

4.6

7.1 cM

2.9–14.8 cM

N32354

20.7

605

LG-01

16.4

3.1

4.8

13.6 cM

8.6–20.8 cM

N9670

60.3

605/602

LG-01

22.5

3.2

7.5

13.6 cM

8.6–20.8 cM

N9670

71.0

606

LG-01

10.65

3.2

4.6

13.6 cM

8.6–20.8 cM

N9670

39.4

606

LG-15

5.2

3.0

4.6

71.7 cM

63.1–76.3 cM

ScORN3R

7.3

606/603

LG-01

14.16

3.4

5.0

56.2 cM

51.5–56.9 cM

VVIp60

36.8

607

LG-01

7.79

3.0

4.4

20.8 cM

5.0–26.5 cM

N9670

22.2

607

LG-06

8.04

2.8

4.4

38.6 cM

28.8–39.4 cM

VMC 3f12

17.1

607

LG-12

5.05

3.1

4.4

85.1 cM

81.2–86.3 cM

A92500

12.0

607

LG-13

6.62

3.0

4.4

41.0 cM

36.0–42.7 cM

UDV 124

14.5

607

LG-15

5.08

3.1

4.4

24.0 cM

18.9–28.0 cM

A11343

12.1

608

LG-05

5.23

3.2

4.4

54.7 cM

50.1–56.7 cM

VVMD 14

23.1

608

LG-12

4.54

3.1

4.4

85.1 cM

78.9–97.4 cM

A92500

10.9

608

LG-16

5.13

3.1

4.4

31.8 cM

28.1–35.3 cM

L14174

18.1

609

LG-01

6.58

3.1

4.6

33.0 cM

31.6–37.0 cM

VMC 8a7

20.5

610

LG-01

6.22

3.2

4.7

35.9 cM

31.6–37.9 cM

VMC 8a7

16.7

612

LG-02

7.87

3.0

4.6

58.1 cM

54.8–62.9 cM

M20940

33.3

613

LG-06

5.77

3.0

4.5

54.2 cM

46.6–69.0 cM

UDV 85

22.6

613

LG-07

4.67

3.0

4.5

33.2 cM

28.4–36.1 cM

RL18108

32.4

614

LG-02

7.05

2.8

5.5

58.1 cM

54.8–62.9 cM

M20940

12.9

614

LG-11

7.95

5.4

5.5

51.8 cM

51.4–54.1 cM

A14900

43.9

614/615

LG-02

8.13

3.0

4.6

89.8 cM

84.3–89.8 cM

M32200

23.2

614/615

LG-08

7.67

3.1

4.6

54.5 cM

46.1–57.4 cM

L014277

22.6

617

LG-01

5.27

3.2

4.5

35.9 cM

31.6–37.9 cM

VMC 8a7

13.6

618

LG-10

4.51

2.6

4.5

19.2 cM

5.0–32.5 cM

VMC RegVR1a

13.2

665

LG-07

6.8

3.1

4.6

36.6 cM

33.2–41.0 cM

L2200

25.1

a Linkage groups named as the International Grape Genome Program (IGGP) nomenclature

b Maximum LOD score obtained by MQM mapping

c Calculated by permutation test at P ≤ 0.05

d Nearest molecular markers to the peaks of the QTLs

Discussion

The Integrated “Regent” × “Lemberger” map

The mapping population used in the present investigation has been previously employed for mapping studies (Fischer et al. 2004). In contrast to the present study, two separate maps of the parental genotypes had been obtained and could only partially be integrated, due to the predominant application of dominant molecular markers (RAPD and AFLP) at that time. To render this map more informative, microsatellite markers and resistance-related markers (RGA-derived markers and SCAR markers) were additionally mapped in this new investigation. The codominant microsatellite markers easily permitted an integration of the parental maps (Fig. 2). The combination of various types of molecular markers (Microsatellites, AFLP, RAPD, SCAR, CAPS, RGA-based markers) detailed the genetic map with a high density of markers, improving its usefulness for QTL detection and future map-based cloning approaches.

Several of the microsatellite loci used in this investigation have also been mapped in other studies (Riaz et al. 2004; Adam-Blondon et al. 2004; Di Gaspero et al. submitted). A congruence of the linear order of the commonly mapped microsatellite markers was generally observed when comparing the present map to the others. Few inversions occurred in microsatellite markers located close to each other. The map length observed is in the range of other genetic maps developed for grapevine.

Distorted marker segregation has been observed in genetic maps constructed from grapevine crosses (e.g., Adam-Blondon et al. 2004). In the case of “Regent” × “Lemberger,” one distorted region (LG-15) coincides with the major QTL for resistance to powdery mildew (see below). This fact may hint at the possibility that these regions were introgressed from Vitis wild species genotypes serving as resistance donors in the complex pedigree of “Regent” (Akkurt et al. 2007). The sequences of the two homologous chromosomal regions of the diploid grapes may differ considerably in these introgressed areas affecting meiotic synapsis and recombination.

QTL analysis

Powdery and downy mildew resistance factors

The integrated “Regent” × “Lemberger” map was employed for QTL analysis by interval mapping and subsequent MQM mapping. This allowed the identification of QTLs affecting the resistance to powdery and downy mildew in grapevine. For all the QTLs detected the positively linked marker allele is one from “Regent”. “Regent” is a recently bred red grapevine cultivar that combines high wine quality and resistance to the two worldwide most important pathogens (powdery and downy mildew). Its resistance traits were introgressed by complex cross-breeding between susceptible traditional Vitis vinifera cultivars (high wine quality) and specific genotypes of wild Vitis species: These show inferior wine quality, but are the only known source of resistance to both pathogens. Different wild species could have been the donor of the resistances to both diseases presents in cv. “Regent” (Akkurt et al. 2007).

Both for powdery and downy mildew one QTL with major effect was detected and is in agreement with the earlier description of Fischer et al. (2004). The present investigation provides three additional years of phenotypic evaluations for both diseases as compared to the previous study. The resistance to powdery and downy mildew has now been scored in the population for at least 5 years. However, the data from 2003 did not allow to detect any QTL, probably due to the low natural infection pressure resulting from the exceptionally warm and dry climate Europe experienced in that summer. Disregarding the year 2003, the major resistance QTL to downy mildew located on LG-18 was detected reproducibly in all years evaluated, despite the variation on climatic conditions. Contrarily, the major QTL for powdery mildew on LG-15 could not be detected in 2004 and showed only a small effect in 2005. This is probably due to effects of varying climatic conditions.

For downy mildew resistance, one supplementary QTL with minor effects was identified on LG-04. The presence of a major QTL accompanied by minor QTLs appears to be a common phenomenon in plant genetics of resistance (e.g., Nair et al. 2005; Calenge et al. 2005a; George et al. 2003). In contrast, no minor QTLs were identified for powdery mildew, although the resistance response revealed a continues phenotypical distribution. Hypothetically, diverse factors such as the limited size of the population, epistatic interactions and environmentally caused variation could limit the power of QTL analysis, preventing the identification of additional regions with a small effect on the phenotype. A unique major QTL responsible for the resistance to powdery mildew was also identified in some other crops such as wheat (Jakobson et al 2006) and mungbean (Humphry et al. 2003).

The identification of (a) genomic region(s) affecting the resistance to pathogens through QTL analysis is only the first step to isolate and characterize the resistance factors. QTL analysis allows only a rough localization of the resistance factors, requiring further efforts of physical mapping and genetic fine analysis for map-based cloning approaches. Several studies demonstrated the high efficiency of RGA-based markers for the identification of markers closely linked to resistance loci (e.g., Xu et al. 2005; Calenge et al. 2005b; Donald et al. 2002). Therefore, RGA-STS (sequence-tagged sites) primers developed for grapevine (Di Gaspero and Cipriani 2002; 2003) were tested. Two RGA-based markers amplified by primer pair “rgVamu137” are found located within the support interval of the major QTL for resistance to downy mildew. These primers had been designed to amplify a Toll-interleukin-type receptor nucleotide binding site (TIR-NBS-LRR) RGA class, which is thought to play an important role in pathogen recognition and signal transduction. Their co-location with the major resistance QTL could hint at a putative functional role to be studied further.

A greater number of RGA-derived markers were mapped in the two mapping populations “Chardonnay” × “Bianca” and “Cabernet Sauvignon” × hybrid 20/3 (Di Gaspero et al. submitted). Several RGA-derived markers used in that study were mapped in a cluster on LG-18. The RGA marker “rgVrip064” was found associated to downy mildew resistance in some resistant genotypes (Di Gaspero and Cipriani 2002) and is located in the same region (Di Gaspero et al. submitted). Based on the position of commonly mapped microsatellite markers it was confirmed that this is the same region covered by the major QTL for downy mildew resistance as identified here in “Regent.”

This study permitted the identification of genomic regions associated with powdery and downy mildew resistance. RGA-derived markers within the support interval of the QTL with major effect on the resistance to downy mildew were detected as well as microsatellite loci in close linkage to resistance QTLs. These markers can now be tested for their correlation to resistance in different genetic backgrounds (e.g., Akkurt et al. 2007). Such validated markers will serve in marker-assisted selection procedures to accelerate breeding as well as in germplasm characterization. In addition, they can be employed for positional cloning of the genomic regions of interest to study possible candidate genes for resistance and understand the cellular pathways involved.

Leaf morphology traits

Grapevine leaf morphology traits have high discriminant power for the differentiation of the comprehensive variety of cultivars present (estimated ca. 8,000–10,000 worldwide). They have proven very useful for ampelographic examination of cultivars. These traits should be controlled by genetic determinants influencing the developmental patterns of leaf morphogenesis. As a first approach to address morphogenetic regulation in grapevine, the segregation of leaf characteristics was studied in the “Regent” by “Lemberger” progeny. In total, 18 different ampelometric traits clearly defined as OIV descriptors were measured (OIV 2007). Ratios of OIV 603 to OIV 601, 605/602, 606/603, 611/601, 612/613 and 614/615 were calculated in addition, as they are environmentally more stable scores than the individual ampelometric measures. All leaf characteristics were processed through QTL analysis.

In total, 27 statistically significant QTLs affecting leaf morphology were identified in the “Regent” × “Lemberger” map (Table 5). The morphology of leaf teeth and the angles of leaf veins seem to be determined by many different loci dispersed around the genome. Leaf angles are strongly correlated with the opening or overlapping of the leaf sinus (Dettweiler 1987). There seems to be an accumulation of morphogenetic factors, particularly those affecting the depth of the leaf sinus, on LG-01.

To our knowledge, this is the first report on the analysis of morphogenetic factors in the genome of grapevine.

Acknowledgments

We wish to thank Gabriele di Gaspero (University of Udine, Italy) for exchange of information prior to publication. Sabine Wiedemann-Merdinoglu (INRA Colmar) contributed with downy mildew resistance data from greenhouse experiments in order to validate them with field data. Johan W. Van Ooijen was helpful for QTL analysis. Charlotte Gleich provided expert technical assistance. This work was supported by the Coordination for the Improvement of Higher Education Personnel (CAPES)-Brasilia/Brazil (Ph.D. fellowship provided to L.J.W.), the German Academic Exchange Service (Deutscher Akademischer Austauschdienst, DAAD, visiting scientist fellowship for N. G.-B.), Ankara University (Ph.D fellowship provided to M. A.) and funds from the German Federal Ministry of Nutrition, Agriculture and Consumers Protection (Bundesministerium für Ernährung, Landwirtschaft und Verbraucherschutz, BMELV).

Copyright information

© Springer Science+Business Media B.V. 2007