QTL mapping and underlying genes for heat tolerance in grapevine (Rhine Riesling × Cabernet Sauvignon) under field conditions

Abstract

Key message

QTL analysis for key physiological traits assessed during hot days highlighted 26 genomic regions and promising candidate genes for thermotolerance and response to light stress under field conditions in grapevine.

Abstract

Grapevine is one of the most widely cultivated perennial fruit crops in the world, with its economic relevance mainly related to wine production. Climate change, with global warming and increased frequency of intense phenomena, is greatly affecting viticulture and the wine sector. Thus, studying the genetic factors involved in grapevine response to high temperatures can help to improve vineyard management strategies and support plant breeding innovations. In this experiment, a mapping population (Rhine Riesling × Cabernet Sauvignon) was used to perform a genetic dissection of the physiological response to increased temperatures under vineyard conditions. Photosynthetic activity and stomatal dynamics were evaluated for three seasons during hot days at different plant developmental stages. Results of quantitative trait loci (QTL) analysis highlighted 26 genomic regions that consistently contribute to the eight tested traits. Candidate genes with supporting evidence, underlying QTL clusters with explained variance above 10%, are those associated with signal perception and transduction, protein homeostasis, osmoprotection, photosynthesis and response to radiation which are relevant mechanisms for plant heat acclimation. Within the stable chromosomal intervals identified by this exploratory analysis, other gene predictions emerged that may be tested for their involvement in grapevine resilience to increasing temperatures. The genetic architecture of quantitative traits linked to grapevine heat tolerance investigated under real field conditions, helps to define key targets for adapting an important traditional crop to environmental changes.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00122-025-04972-2.

Introduction

Nowadays global warming has become an issue because of the increased frequency of intense phenomena (e.g., heat waves, heavy rains) that have adverse impacts on both ecosystems and human systems (IPCC 2023). One of the worst affected sectors is agriculture and consequently the food production chain. In fact, increased temperatures, often coupled with periods of water deficit, reduce yield and food quality with negative consequences on an economic level. Viticulture, with its 258 mhl of wine production and 37.6 bn EUR of global trade value (OIV (International Organisation of Vine and Wine) 2023), is one of the most relevant agricultural sectors at an economic level, and is highly impacted by climate change. Indeed, increased temperatures, compromise yield and influence phenological timings, anticipating berry ripening in the summer heat, leading for example to a higher alcohol content and lower acidity in wines with consequent modification of their sensory profile (Jones and Davis 2000; Van Leeuwen et al. 2019; Faralli et al. 2024). A future warmer climate will threaten the sustainability of viticulture in the current winemaking regions, especially in southern Europe, where temperatures are expected to increase above optimum values, accompanied by severe dryness (Santos et al. 2020). With the aim of improving the management of existing varieties and developing new cultivars with enhanced performances under changing environmental conditions, the study of the genetic factors involved in grapevine response to increased temperatures is essential to gain knowledge about the mechanisms underlying thermotolerance (Fahad et al. 2017). In this framework, quantitative trait loci (QTL) and genome-wide association studies (GWAS) are of utmost importance for an initial analysis of the complex phenotype, combining character investigation with genetic dissection. QTL mapping and GWAS have been successfully applied in the study of biotic stresses, with the discovery of many resistance loci for pests and diseases, also in grapevine (Delrot et al. 2020). Regarding the genetic control of abiotic stress tolerance, drought stress has received considerable attention and has been extensively studied in the past years in major annual crops, but also in walnut (Arab et al. 2022), apple (Wang et al. 2018) and grapevine (Marguerit et al. 2012; Coupel-Ledru et al. 2014, 2016, 2022; Brault et al. 2021; Trenti et al. 2021), which led to the discovery of various QTL, meta-QTL and candidate genes related to yield, root systems and stomatal conductance traits. In recent years, crop species have been intensively studied for thermotolerance too, with first QTL analysis on wheat (e.g., Paliwal et al. 2012) followed by tomato (e.g., Cappetta et al. 2021), maize (e.g., Sheoran et al. 2022) and rice (e.g.. Kilasi et al. 2018), sorghum (Chen et al. 2017), groundnut (Sharma et al. 2023), chickpea (Mohanty et al. 2024) and cowpea (Angira et al. 2022), while literature on perennial crops is very poor. All these works described a polygenic control for heat-responsive traits. Only a few quantitative genetics investigations are reported for the grapevine response to changes in temperature, including a QTL analysis on cold hardiness (Su et al. 2020) and two GWAS on cold tolerance (Wang et al. 2021b) and leaf firing (Coupel-Ledru et al. 2024). Therefore, the genetic architecture of the grapevine response to rising temperatures remains to be explored.

In thermotolerance studies the most frequently considered trait is plant photosynthetic activity. Indeed, together with genetic studies, indicators of photosynthetic performance were used to investigate varietal behavior in heat stress studies of several crop species, such as wheat (Monneveux et al. 2003; Sharma et al. 2012, 2015), maize (Sinsawat et al. 2004), raspberry (Molina-Bravo et al. 2011), tropical fruits (Yamada et al. 1996), tomato (Willits and Peet 2001; Zhou et al. 2015; Poudyal et al. 2019) and particularly in grapevine (Kadir 2006; Palliotti et al. 2009; Greer and Weston 2010; Luo et al. 2011; Greer and Weedon 2012; Xu et al. 2014; Hochberg et al. 2015; Sun et al. 2016, 2017; Xiao et al. 2016; Zha et al. 2016b, a, 2018; Greer 2017, 2019; Liu et al. 2020b; Gallo et al. 2021). Actually, photosynthesis is the first plant process influenced by thermal variation: high temperatures can cause an increase in thylakoid membranes fluidity, which favors the disruption of the oxygen evolving complex (OEC), and can reduce the activity of Rubisco (ribulose-1,5-bisphosphate carboxylase/oxygenase), impairing the chloroplasts major biochemical processes (Wahid et al. 2007; Fahad et al. 2017; Venios et al. 2020).

Several methods can be used for assessing key-steps of processes associated with leaf photosynthesis. Among them, the activity of Photosystem II (PSII) can be estimated via chlorophyll fluorescence emission in dark-adapted samples. Light absorbed by chlorophyll, in fact, can be employed for photosynthesis (photochemistry) or can be either dissipated through heat (non-photochemical quenching, NPQ) or re-emitted as light (fluorescence) (Murchie and Lawson 2013). Therefore, the light emitted from the Photosystem II (PSII) in the absence of NPQ can give information on the status of the photochemical process, providing a good indication of plant status. This technique is particularly attractive because it is non-invasive and non-destructive for plants, and measurements are simple and relatively fast. Moreover, the quantification of the light re-emitted by chlorophyll is usually done with portable instruments (fluorimeters) that can be easily used in field experiments (Maxwell and Johnson 2000).

A reduced leaf transpiration due to stomata closure to minimize water loss results in an increase in leaf temperature under high light conditions, which, combined with heat stress phenomena, can increase leaf sensitivity to sub-optimal environmental conditions. Since water conservation via high stomatal sensitivity to environmental cues could be relevant in the scenario of climate change (Faralli et al. 2019), investigating plant stomatal behavior holds significant importance. However, studies performing QTL analysis on stomatal dynamics in relation to abiotic stress factors are very few in literature and mainly related to salinity stress (Liu et al. 2017) or drought stress (Wang et al. 2003; Trenti et al. 2021).

In the present work, a biparental population obtained from the cross between Rhine Riesling, a German variety adapted to cool-moderate temperatures, and Cabernet Sauvignon,’ one of the world most widespread varieties cultivated in hot-temperate climates, was used to perform a genetic dissection of grapevine physiological response to increased temperatures under field conditions. Photosynthetic activity and stomatal dynamics were evaluated for three consecutive seasons during hot days at different plant developmental stages, in particular flowering and véraison, which are known to be sensitive to temperature increases (Yeh et al. 2012; Jagadish et al. 2021). QTL analysis for these traits highlighted several QTL stable across years, phenotyping sessions or parameters. The investigation of these genomic regions allowed the identification of genes putatively involved in the response to heat stress or solar radiation, contributing to a deeper understanding of the mechanisms involved in grapevine thermotolerance.

Materials and methods

Plant materials

The study was conducted for 3 years (2021–2023) on 122 grapevine individuals obtained from the cross ‘Rhine Riesling’ × ‘Cabernet Sauvignon’ (RRxCS). The cross was made in 2005 and the seedlings were planted in the field on their own roots in 2008. Plants were grown at the Giaroni experimental field of Fondazione Edmund Mach (San Michele All’Adige, Trentino Alto Adige, Italy, 46°18′N, 11°13′E) and were trained using the Guyot system. Analysis of the soil texture showed that Giaroni soil is typically a sandy loam soil (the percentage of sand was 52.5% with 41.9% silt and 5.6% clay, < 1% skeleton) with a soil depth of 0.70–1 m and an estimated available water content (AWC) of 164 mm. During the growing season plants were regularly treated against common grapevine pathogens and underwent two seasonal trimmings. Weather conditions during the study were monitored with an automated meteorological station situated near the vineyard (https://meteo.fmach.it, Fondazione Edmund Mach, Technology Transfer Center, Agrometeorology and Irrigation Unit).

Phenotypic characterization

Chlorophyll a fluorescence transient (O-J-I-P) was measured with a Handy Plant Efficiency Analyzer (Handy-PEA, Hansatech, Norfolk, UK) set up with a light intensity of the saturating pulse of 3500 µmol m⁻² s⁻¹. For parental lines, 5 biological replicates were studied, while 1 biological replicate was available for the progeny. For each plant, three measures were taken on the 5th leaf of three different shoots starting from the shoot base. Shoots were chosen from the head, middle, and end of the cane. Before measurement each leaf was dark-adapted for 30 min with suitable leaf clips. The same leaves were measured throughout the summer season.

In 2022 and 2023, stomatal conductance (g_s) and leaf vapor pressure (VPleaf), leaf temperature (Tleaf), light-adapted maximum quantum yield of the Photosystem II (PhiPS2), electron transport rate (ETR) and leaf apparent transpiration (E) were measured with a LI-600 Porometer/Fluorometer (LI-COR Environmental, Lincoln, Nebraska USA). “Auto gsw” was selected as protocol for LI-COR measurements.

Measures in non-stress conditions (control, C) were taken in the early morning (5:30–8:00), while measures in stress conditions (heat stress, HS) were collected in the afternoon during the hottest hours (13:00–15:30). To ensure the proper temperature and light conditions during control measurements, the population was divided into two subgroups and evaluated on two consecutive hot days. Measures in 2021 were taken at flowering (E-L21, beginning of June) and véraison (E-L35, beginning of August). In 2022 and 2023, in addition to these stages, two other datasets were collected at the end of June and in July (corresponding to berry pea-size, E-L31, and pre-véraison E-L33, respectively), while in 2023 an additional dataset was acquired at full véraison in August (E-L36), when the highest air temperature was reached.

QTL analysis

The RRxCS population was genotyped with an Illumina Vitis18K SNP chip and SSR markers in a previous work and a high-density linkage map was obtained (Vervalle et al. 2022). The integrated map contained 3459 representative markers mapped at 1815 unique positions along 19 linkage groups, with a length of 1413 centiMorgan (cM) and an average inter-locus gap of 0.78 cM. For each individual of the RRxCS progeny and for each phenotyping session the difference between stress and non-stress conditions of measured parameters was normalized for control values (formula: (C-HS) * C⁻¹ * 100). According to the literature (Talukder et al. 2014) and a preliminary QTL analysis, this approach was considered more robust than analyzing datasets obtained from simple subtraction between data at control and during stress. QTL detection was carried out on the integrated map (named ‘CP map’, from cross-population) and the two parental maps (named ‘RR map’, from Rhine Riesling, and ‘CS map’, from Cabernet Sauvignon) using the MapQTL 6.0 software (Van Ooijen 2009) with data from separate phenotyping sessions. Simple interval mapping (SIM) was performed for all three maps with the mixture model algorithm, while multiple QTL mapping (MQM) was performed only on the two parental maps due to its high computation cost. MQM analysis was based on the Forward Selection approach because it appeared less prone to false positives compared with Automatic Cofactor Selection. LOD (logarithm of odds) significance thresholds at a P value of 0.05 were established for each linkage group through 1000 permutations. A nonparametric Kruskal–Wallis test was carried out to provide support to marker-trait associations. The QTL confidence interval (in cM) was calculated as the genetic region where the LOD value was higher than the maximum LOD score of the QTL minus 1. For confirmed QTL in at least two datasets, potential candidate genes were investigated in the consensus region (i.e., QTL overlapping region or, in the case of two adjacent QTL, the entire region was considered). Markers’ physical positions (in base pair, bp) were obtained by blasting the probe sequences (for SNPs) or the repetitive sequences (for SSRs) on the PN40024.v4 reference (REF) genome assembly. For PN40024.v4 heterozygous regions, alternative (ALT) sequences were also considered (Velt et al. 2023). Structural and functional gene annotations (in the v4.3 and v4.1 version, respectively) were recovered from the Grape Genomics Encyclopedia portal (https://grapedia.org/) (Velt et al. 2023). Candidate genes were selected, among those included in the QTL confidence intervals, based on the knowledge of the main mechanisms that plants have evolved to counteract the effects of heat stress as well as on previous works exploring the transcriptomic response of grapevine to heat stress.

Among the traits measured and calculated by the Handy-PEA fluorometer in relation to chlorophyll fluorescence (Table S 1.1, Online Resource 1), in this study four parameters were selected for QTL analysis: F_v/F_m, which gives an estimation of the maximum quantum efficiency of PSII; Area, which is proportional to the pool size of the electron acceptors Q_A⁻ on the reducing side of PSII; Phi(Eo), which is the quantum yield of electron transport; and Phi(Ro), which is the average quantum yield of primary photochemistry. These independent parameters were chosen based on their physiological meaning. In fact, many parameters reported in Table S 1.1 are calculated and derived from others and therefore they are expected to display the same trend and results in the QTL analysis.

Traits related to stomatal dynamics and selected for QTL analysis were: g_s, which measures stomatal conductance in vivo; E, which gives an estimation of leaf transpiration; VPleaf, which measures leaf vapor pressure; Tleaf, which corresponds to leaf temperature; PhiPS2, which gives an estimation of the maximum quantum efficiency of PSII without dark adaptation; and ETR, which gives an estimation of the electron transport rate (complete list of measured parameters in Table S 1.2, Online Resource 1).

Statistical analysis and graphical representation

Statistical analysis and graphical representation were performed using the R software environment 4.2.1 (https://www.r-project.org/). The normality of each trait distribution was evaluated by the Shapiro–Wilk Normality Test. Correlations between physiological traits measured in different phenotyping sessions and years were calculated with the package ‘tidyverse’ using the nonparametric Spearman correlation coefficient (rho). QTL intervals were visualized in a single image with the R package ‘RIdeogram.’

Results

Environmental conditions

The period from bud burst to harvest (April-September) was characterized by moderate temperatures and approximately ~ 600 mm of cumulative precipitations (Fig. 1) well distributed over the growing season. Due to the near-hydromorphic nature of the soil located close to the Adige River and the maximum water-holding capacity reached at the beginning of each season, no significant level of soil moisture deficit occurred during the three vintages. Years 2022 and 2023 were characterized by mean temperatures, both in their minimum and maximum values, about ~ 1 °C higher compared to 2021, with 2022 having the highest maximum mean temperature of 27.8 °C, and a maximum air temperature of 36.6 °C for two days in July. In 2023, the maximum air temperature recorded during the study, 36.8 °C, was reached in August (Fig. 1).

Fig. 1.

Climograph reporting precipitations (mm) and air temperatures (mean in green, maximum in red and minimum in light-blue) in Giaroni experimental field (Trentino Alto Adige, Italy, 46°18′N, 11°13′E), for seasons 2021–2023 in the period between bud burst-harvest (April–September) (colour figure online)

Phenotyping for QTL analysis related to heat-responsive traits was performed at different phenological stages. Detailed temperature information for all the phenotyping sessions is reported in Table 1. In 2021 and 2023, temperatures during flowering phenotyping sessions were low; therefore, those datasets were not included in the present study.

Table 1.

For each phenotyping session performed, dataset code, phenotyping date, day mean (Day mean T), maximum (Air T_max) and minimum (Air T_min) air temperature, temperature difference between morning and afternoon phenotyping sessions (ΔT) and monthly total precipitation are reported

Dataset code	Phenotyping dates	Day mean T (°C)	Air Tmax (°C)	Air Tmin (°C)	ΔT (°C)	Monthly total rainfall (mm)
Véraison 2021	11/08/2021	24.0	32.3	19.0	13.3	127.8
	12/08/2021	24.9	33.0	18.5	14.5
	Mean	24.5	32.7	18.8	13.9
Flowering 2022	02/06/2022	22.4	28.8	15.3	13.5	133.8
	04/06/2022	24.0	30.3	18.7	11.6
	Mean	23.2	29.6	17.0	12.6
Berry pea-size 2022	20/06/2022	24.0	32.0	15.8	16.2	133.8
	21/06/2022	23.9	31.1	18.5	12.6
	Mean	24.0	31.6	17.2	14.4
Pre-Véraison 2022	19/07/2022	27.3	34.8	17.5	17.3	50.4
	20/07/2022	27.7	35.9	17.4	18.5
	Mean	27.5	35.4	17.5	17.9
Véraison 2022	03/08/2022	25.0	32.8	17.5	15.3	106
	05/08/2022	27.0	35.0	18.5	16.5
	Mean	26.0	33.9	18.0	15.9
Berry pea-size 2023	20/06/2023	25.1	30.5	19.0	11.5	27.8
	21/06/2023	25.1	31.7	18.2	13.5
	Mean	25.1	31.1	18.6	12.5
Pre-Véraison 2023	11/07/2023	28.2	34.2	21.8	12.4	217.6
	17/07/2023	27.3	33.4	21.6	11.8
	Mean	27.7	33.8	21.7	12.1
Véraison 2023	09/08/2023	19.2	25.7	14.7	11.0	127
	10/08/2023	22.7	28.9	17.0	11.9
	Mean	20.9	27.3	15.9	11.5
Véraison Full 2023	22/08/2023	27.4	36.8	20.4	16.4	127

Datasets with the highest average maximum air temperature were Pre-Véraison 2022 and Véraison Full 2023, with values of 35.4 °C and 36.8 °C, respectively. The corresponding average ΔT values were 17.9 and 16.4 °C. For the other collected datasets, the average maximum air temperature ranged from 27.3 to 33.9 °C and the average ΔT ranged between 11.5 and 15.9 °C.

Phenotypic characterization

Mean values and ranges of variation for parameters related to chlorophyll fluorescence analyzed in this study in all the datasets collected between 2021 and 2023 are reported in Table S 1.3 (Online Resource 1). For all parameters, there was a decrease in the afternoon compared to control values. The four parameters considered in the study (F_v/F_m, Area, Phi(Ro), Phi(Eo)) were, in general, positively correlated within all datasets collected, with rho values ranging from 0.21 (F_v/F_m and Phi(Ro) in Berry pea-size 2023) to 0.90 (F_v/F_m and Phi(Eo) in Pre-Véraison 2023) (Table S 1.4, Online Resource 1). On the contrary, correlations registered for the same parameter across different phenotyping sessions were in general small and often not significant, especially when comparing datasets belonging to different years (Table S 1.5, Online Resource 1). For example, for the F_v/F_m parameter, it is evident that individuals’ behavior changed not only between different years, but also along the season (Fig. 2).

Fig. 2.

Reduction in F_v/F_m (% relative to control) in different phenotyping sessions for all the individuals of the mapping population (‘Rhine Riesling’ × ’Cabernet Sauvignon’) reported with their numerical codes. Phenotyping sessions are indicated as follows: ▲ flowering, ■ berry pea-size, ● pre-véraison, × véraison, * véraison full; while dots are colored based on the year: 2021 (blue), 2022 (red), 2023 (green) (colour figure online)

Significant correlations of F_v/F_m across years ranged between 0.21 and 0.36, while correlations ranged between 0.21 and 0.73 when considering datasets from the same year (Table S 1.5- a, Online Resource 1). Parental lines showed a more consistent behavior throughout different years and seasons also because the reduction in F_v/F_m reported here is the mean of 5 biological replicates. Frequency distribution plots of the difference between stress and non-stress conditions of measured parameters (normalized for control values, formula: (C-HS) * C⁻¹ * 100) were approximately normal and showed continuous variation (Figure S 1.1, Online Resource 1). Rhine Riesling and Cabernet Sauvignon displayed a similar phenotype and transgressive segregants among the population were observed.

Regarding transpirational traits, we observed a general trend of increasing stomatal conductance and leaf transpiration in the afternoon (Table S 1.6, Online Resource 1). This trend was also observed for both parental lines, Rhine Riesling and Cabernet Sauvignon (data not shown). High correlations (0.93–0.999) were registered for Tleaf and VPleaf in all datasets acquired, while g_s and E in general did not correlate with Tleaf (Table S 1.7, Online Resource 1). All stomatal and transpirational parameters negatively correlated with PhiPS2, but the degree of correlation varied with the dataset considered. As for parameters measured with Handy-PEA, correlations of the same parameter between different datasets were higher when considering datasets from the same year (Table S 1.8, Online Resource 1). Correlations between traits measured with Handy-PEA and LI-600 porometer/fluorometer in the same phenotyping session were additionally calculated (Table S 1.9, Online Resource 1). In this case, correlations depended on the considered dataset, making it difficult to highlight common trends between datasets. However, there were significant and positive correlations between F_v/F_m and PhiPS2, with rho values ranging between 0.28 and 0.52, while Tleaf negatively correlated with Area, F_v/F_m and in some cases also with Phi(Eo). Frequency distribution plots for parameters measured with LI-600 Porometer/Fluorometer are available in Figure S 1.2, Online Resource 1. These traits appear to have a non-normal distribution; the progeny tends to display values which are similar to the parental lines, but transgressive segregation can still be observed.

QTL analysis

Chlorophyll fluorescence

SIM analysis performed on the integrated map with data from a total of 9 phenotyping sessions over three years resulted in the identification of 56 QTL: 19 for the maximum quantum yield of Photosystem II (F_v/F_m), 10 for the pool size of electron transport (Area), 15 for the quantum yield of electron transport (Phi(Eo)), and 12 for the average quantum yield of primary photochemistry (Phi(Ro)) (Table S 2.1, Online Resource 2). QTL identified in the integrated map have a LOD in the range 1.9–5.5 and a phenotypic explained variance (EV%) between 7.6 and 21.5%. On the other hand, MQM analysis performed on parental maps highlighted 96 QTL: 27 for F_v/F_m, 17 for Area, 30 for Phi(Eo), and 22 for Phi(Ro), respectively (Table S2.2, Online Resource 2). QTL identified in parental maps have a LOD in the range 1.6–5.4 and their EV is between 4 and 17.4%. QTL related to chlorophyll fluorescence appear to be distributed in almost all 19 chromosomes (Chr) of the grapevine genome, with Chr 3, 4, 7, 10, 12, 13, 17, and 19 being the most populated (see Figure S 2.1 in Online Resources 2 as a summary of QTL obtained from SIM analysis on the integrated map and MQM analysis on parental maps). As shown in Fig. 3, some of the QTL obtained from the MQM analysis co-locate, highlighting genomic regions worthy of investigation for candidate genes.

Fig. 3.

Results of MQM analysis (2021–2023) on chlorophyll fluorescence parameters. Parameters are represented with different colors (F_v/F_m in green, Area in red, Phi(Eo) in blue and Phi(Ro) in yellow) while years are reported with different shapes (● 2021, ▲ 2022, ■ 2023). Chromosomes are colored following the increasing frequency with which a genomic region harbors significant QTL according to MQM analysis: from 0 (gray) to 5 repetitions (red) across different datasets and/or parameters. The size of genomic regions highlighted in the figure is reported in basepairs (colour figure online)

Oppositely, SIM and MQM analysis performed with data from a total of 5 datasets for chlorophyll fluorescence parameters measured with the LI-600 porometer, PhiPS2 and ETR, did not highlight genomic regions repeated across phenotyping sessions or years (data not shown). Moreover, no common regions were found in QTL analyses performed for F_v/F_m and PhiPS2, parameters measured with (Handy-PEA) and without (LI-600 Porometer/Fluorometer) leaf dark adaptation, respectively. Therefore, those regions were not further investigated for candidate genes.

Stomatal conductance

SIM analysis performed on the integrated map with data from a total of 5 different datasets collected in 2 years highlighted 26 QTL: 5 related to stomatal conductance (g_s), 6 related to leaf transpiration (E), 7 for leaf vapor pressure (VPleaf), and 8 for leaf temperature (Tleaf) (Table S 2.3 Online Resource 2). QTL identified in the integrated map have a LOD in the range 1.5–6.1 and a phenotypic explained variance between 6 and 22.2%. On the other hand, MQM analysis performed on parental maps resulted in the identification of 41 QTL: 14 for g_s, 14 for E, 5 for VPleaf, 8 for Tleaf (Table S 2.4, Online Resource 2). QTL identified in parental maps have a LOD in the range 0.7–3.8 and their EV is between 0.1 and 12.5%.

QTL related to stomatal conductance and leaf transpiration are fewer in number and less distributed across chromosomes compared to QTL for chlorophyll fluorescence, with Chr 3 and 5 being the most populated (see Figure S 2.2 in Online Resources 2 as a summary of QTL obtained from SIM analysis on the integrated map and MQM analysis on parental maps). Despite the frequency of their repetition throughout different datasets being quite low, 10 genomic regions with co-localized QTL and underlying candidate genes were identified (Fig. 4).

Fig. 4.

Results of MQM analysis (2021–2023) on leaf transpirational parameters. Parameters are represented with different colors (g_s in green, E in red, VPleaf in blue and Tleaf, in yellow) while years are highlighted with different shapes (● 2021, ▲ 2022, ■ 2023). Chromosomes are colored following the increasing frequency with which a genomic region harbors significant QTL according to MQM analysis: from 0 (gray) to 5 repetitions (red) across different datasets and/or parameters. The size of genomic regions highlighted in the figure is reported in basepairs (colour figure online)

Consensus genomic regions, with stable QTL identified throughout years and/or phenological stages, were considered for candidate genes exploration (Tables 2, 3). Functional candidate genes that have been associated with the transcriptomic response of plants to heat stress, are summarized in Table 4. The complete list of positional candidate genes within each region is reported in Table S 3.1 and Table S 3.2, Online Resource 3, for QTL related to chlorophyll fluorescence and stomatal dynamics, respectively.

Table 2.

Consensus genomic regions (indicated with a QTL code and reported in base pairs on the reference genome assembly PN40024.v4) in which QTL for chlorophyll fluorescence (cf) kinetics related to heat stress co-locate. The number of QTL detected, the genomic regions corresponding to their confidence intervals, the associated traits and datasets, the explained variance (EV%) and the number of genes in the interval are also indicated. QTL clusters originated by QTL with EV > 10% are highlighted in bold

Consensus genomic region				QTL co-locating						n genes
QTL code	Chr	Start (bp)	End (bp)	n QTL	Trait	Dataset	Start (bp)	End (bp)	EV%
cfQTL_3.1*	3	19,392,699	19,751,147	4	Area	Véraison 2021	19,392,699	19,751,147	14.2	19
					Area	Flowering 2022	19,392,699	19,813,686	9.4
					Area	Véraison Full 2023	19,392,699	19,751,147	8.1
					Phi(Ro)	Véraison 2022	19,363,439	19,392,699	16.3
cfQTL_3.2	3	19,929,243	20,575,870	2	Phi(Ro)	Véraison 2021	19,929,243	20,575,870	7.2	43
					Phi(Ro)	Flowering 2022	19,929,243	20,575,870	9.5
cfQTL_4.1	4	1,295,066	1,419,855	3	Phi(Eo)	Berry pea-size 2022	1,295,066	1,419,855	11.6	14
					Phi(Eo)	Pre-Véraison 2022	1,295,066	1,419,855	12.2
					Phi(Ro)	Véraison Full 2023	1,295,066	1,419,855	8.3
cfQTL_4.2	4	7,561,727	8,842,722	2	Fv/Fm	Pre-Véraison 2022	7,526,331	8,842,722	8.1	56
					Fv/Fm	Véraison 2022	7,561,727	8,842,722	8.3
cfQTL_7.1	7	21,817,724	23,247,875	2	Fv/Fm	Pre-Véraison 2022	21,817,724	23,247,875	8.5	147
					Phi(Eo)	Pre-Véraison 2022	21,805,347	21,817,724	10.8
					Phi(Eo)	Berry pea-size 2022	21,817,724	22,449,811	7.1
					Phi(Eo)	Véraison 2022	22,988,546	23,533,273	9
cfQTL_10.1*	10	4,562,576	5,568,625	2	Fv/Fm	Véraison 2021	4,562,576	5,441,269	7	94
					Phi(Eo)	Véraison Full 2023	5,441,269	5,568,625	6.6
cfQTL_11.1	11	493,591	664,729	2	Phi(Eo)	Véraison Full 2023	459,975	664,729	6.6	27
					Area	Véraison Full 2023	493,591	664,729	7.6
cfQTL_12.1	12	2,814,814	2,886,340	5	Fv/Fm	Berry pea-size 2023	2,814,814	2,886,340	13.1	1
					Fv/Fm	Pre-Véraison 2023	2,814,814	2,886,340	8.3
					Area	Berry pea-size 2023	2,814,814	2,886,340	8.2
					Phi(Eo)	Berry pea-size 2023	2,814,814	2,886,340	9.1
					Phi(Eo)	Véraison 2023	2,814,814	2,886,340	10.7
cfQTL_12.2	12	14,585,617	14,910,568	2	Area	Véraison 2022	14,585,617	14,910,568	5.6	13
					Phi(Ro)	Berry pea-size 2022	14,585,617	14,910,568	5.5
cfQTL_12.3*	12	17,505,010	17,528,465	2	Fv/Fm	Pre-Véraison 2022	17,505,010	17,528,465	6	2
					Phi(Ro)	Véraison 2021	17,505,010	17,542,846	6.7
cfQTL_13.1	13	10,082,268	10,844,992	2	Fv/Fm	Flowering 2022	10,082,268	10,844,992	7	43
					Phi(Eo)	Flowering 2023	10,082,268	10,844,992	7
cfQTL_13.2	13	20,532,927	20,532,927	2	Fv/Fm	Véraison Full 2023	20,532,927	20,532,927	6.5	126
					Area	Véraison 2022	18,926,455	21,543,135	6.2
cfQTL_15.1	15	20,559,862	20,653,974	2	Phi(Ro)	Flowering 2022	20,559,862	20,653,974	7.5	22
					Area	Flowering 2022	20,498,981	20,653,974	7.5
cfQTL_17.1	17	2,362,104	2,419,076	2	Phi(Eo)	Pre-Véraison 2023	2,362,104	2,629,802	7.1	39
					Fv/Fm	Berry pea-size 2023	2,203,470	2,419,076	9.8
cfQTL_17.2*	17	4,591,415	4,678,056	2	Area	Berry-pea size 2023	4,591,415	4,678,056	9.8	22
					Fv/Fm	Véraison 2022	4,565,731	4,733,756	6.1
cfQTL_19.1	19	15,845,364	17,044,447	3	Fv/Fm	Berry pea-size 2022	14,538,060	17,044,447	11.5	87
					Phi(Eo)	Berry pea-size 2022	14,538,060	17,044,447	15.9
					Phi(Ro)	Véraison 2022	15,845,364	17,660,947	6.7

Table 3.

Consensus genomic regions (indicated with a QTL code and reported in base pairs on the reference genome assembly PN40024.v4) in which QTL for stomatal dynamics (tr) related to heat stress co-locate. The number of QTL detected, the genomic regions corresponding to their confidence intervals, the associated traits and datasets, the explained variance (EV%) and the number of genes in the interval are also indicated. QTL clusters originated by QTL with EV > 10% are highlighted in bold

Consensus genomic region				QTL co-locating						n genes
QTL code	Chr	Start (bp)	End (bp)	n QTL	Trait	Dataset	Start (bp)	End (bp)	EV%
trQTL_3.1	3	697,606	738,212	2	Tleaf	Pre-Véraison 2023	697,606	738,212	7.2	8
					Vpleaf	Pre-Véraison 2023	697,606	738,212	7
trQTL_3.2	3	3,783,778	3,803,994	2	Tleaf	Flowering 2022	3,783,778	3,803,994	8.3	2
					Vpleaf	Flowering 2022	3,783,778	3,803,994	12
trQTL_3.3	3	6,881,348	7,027,379		gs	Berry pea-size 2023	6,881,348	7,027,379	7.2	7
					E	Berry pea-size 2023	6,881,348	7,027,379	7.1
trQTL_4.1	4	578,155	622,483	2	gs	Véraison 2023	578,155	622,483	0.1	8
					E	Véraison 2023	578,155	622,483	2.8
trQTL_4.2	4	16,249,818	16,854,397	2	gs	Berry pea-size 2023	16,249,818	16,854,397	8.2	34
					E	Berry pea-size 2023	16,249,818	16,854,397	6.8
trQTL_5.1	5	5,502,441	5,593,619	2	gs	Berry pea-size 2023	5,502,441	5,563,075	8.5	10
					gs	Berry pea-size 2022	5,563,075	5,593,619	12.5
					Tleaf	Flowering 2022	5,563,075	5,593,619	8.1
trQTL_5.2	5	6,546,178	6,717,310	2	gs	Véraison 2023	6,546,178	6,717,310	0.2	23
					E	Véraison 2023	6,546,178	6,717,310	3.7
trQTL_9.1	9	3,585,046	4,406,275	2	gs	Berry pea-size 2023	3,585,046	4,406,275	10.3	99
					E	Berry pea-size 2023	3,585,046	4,406,275	10.1
trQTL_15.1	15	16,825,206	17,149,995	2	gs	Berry pea-size 2022	16,825,206	17,149,995	5.8	26
					E	Berry pea-size 2022	16,825,206	17,149,995	5.5
trQTL_19.1	19	14,538,060	17,044,447	2	Tleaf	Pre-Véraison 2023	14,538,060	17,044,447	8.6	65
					Vpleaf	Pre-Véraison 2023	14,538,060	17,044,447	9

Table 4.

Candidate genes for the response to heat stress and UV-light radiation underlying consensus genomic regions in which QTL related to chlorophyll fluorescence or stomatal dynamics co-locate. For each candidate gene, the QTL code of the region where it is located, the Gene ID, the corresponding functional annotation and supporting evidence are reported. Genes predicted in consensus regions with EV > 10% are highlighted in bold. Abbreviations: HS = heat stress, WDS = water deficit stress

QTL code	Gene ID	Annotation	Supporting evidence
Signal perception, transduction and transcriptional control
trQTL_3.3	Vitvi03g00614	Somatic embryogenesis receptor kinase 1	Down-regulated by recovery from heat stress (HS) (Liu et al. 2012)
cfQTL_4.2	Vitvi04g00728	Histidine Kinase 5	Involved in response to different stresses (Pham and Desikan 2012)
trQTL_4.2	Vitvi04g01151	Auxin efflux carrier component 5	Down-regulated by HS (Zha et al. 2020)
trQTL_5.1	Vitvi05g00540	Serine/threonine-protein kinase HT1	Down-regulated by HS (Zha et al. 2020)
trQTL_5.1	Vitvi05g00542	Protein NETWORKED 1A	Down-regulated by HS (Liu et al. 2012)
trQTL_5.2	Vitvi05g00640	Transcription factor bHLH92	Up-regulated by HS (Rocheta et al. 2014; Liu et al. 2020a); up-regulated by combined HS-WDS (Ju et al. 2021)
cfQTL_7.1	Vitvi07g01408	Probable 1-deoxy-D-xylulose-5-phosphate synthase 2, chloroplastic	Up-regulated by HS (Liu et al. 2012)
cfQTL_7.1	Vitvi07g01425	Protein ecdysoneless-like	Up-regulated by HS (Rocheta et al. 2014)
cfQTL_7.1	Vitvi07g01457	14-3-3-like protein GF14 kappa	Up-regulated by HS (Cheng et al. 2018)
cfQTL_7.1	Vitvi07g01488	Homeobox-leucine zipper protein ATHB-6	Down-regulated by HS (Rocheta et al. 2014)
cfQTL_7.1	Vitvi07g01498	14-3-3-like protein	Down-regulated by HS (Liu et al. 2012)
trQTL_9.1	Vitvi09g00321	Ankyrin repeat-containing protein ITN1-like	Down-regulated by HS (Zha et al. 2020)
trQTL_9.1	Vitvi09g00323	Dehydration-responsive element-binding protein 3	Down-regulated by HS, WDS and combined HS-WDS (Rocheta et al. 2014; Ju et al. 2021)
trQTL_9.1	Vitvi09g00355	Dynein light chain LC6, flagellar outer arm	Down-regulated by HS (Liu et al. 2012; Zha et al. 2020)
trQTL_9.1	Vitvi09g00393	Trihelix transcription factor ASIL1	Candidate gene for chlorophyll content (Tafesse et al. 2020)
cfQTL_10.1	Vitvi10g00398	PAP-specific phosphatase HAL2-like	Down-regulated by HS (Zha et al. 2020)
cfQTL_10.1	Vitvi10g00417	Protein STRUBBELIG-RECEPTOR FAMILY 1-like isoform X2	Up-regulated by HS (Liu et al. 2012; Rocheta et al. 2014)
cfQTL_11.1	Vitvi11g00030	Sucrose synthase	Down-regulated by HS (Liu et al. 2012, 2020a; Rocheta et al. 2014)
cfQTL_11.1	Vitvi11g00045	Ethylene-responsive transcription factor ERF017	Up-regulated by HS, WDS and combined HS-WDS (Liu et al. 2012, 2020a; Rocheta et al. 2014; Zhao et al. 2014)
cfQTL_11.1	Vitvi11g00046	Ethylene-responsive transcription factor ERF016	Up-regulated by HS (Zhao et al. 2014)
cfQTL_12.1	Vitvi12g00181	Serine/threonine-protein kinase EDR1	Down-regulated by HS (Rocheta et al. 2014)
cfQTL_13.1	Vitvi13g00852	GDSL esterase/lipase At5g03610	Down-regulated by HS (Zha et al. 2020)
cfQTL_13.2	Vitvi13g01358	Probable leucine-rich repeat receptor-like protein kinase At1g35710	Down-regulated by HS (Zha et al. 2020)
cfQTL_13.2	Vitvi13g01396	65-kDa microtubule-associated protein 5	Up-regulated by HS (Zha et al. 2020)
cfQTL_15.1	Vitvi15g01077	Mitogen-activated protein kinase homolog MMK2	Involved in ROS metabolism (Nakagami et al. 2006)
cfQTL_17.1	Vitvi17g00157	Protein SHOOT GRAVITROPISM 5	Down-regulated by HS (Rocheta et al. 2014)
cfQTL_19.1, trQTL_19.1	Vitvi19g01284	C2 domain-containing protein At1g53590	Up-regulated by HS (Rocheta et al. 2014)
cfQTL_19.1	Vitvi19g01356, Vitvi19g04443	9-cis-epoxycarotenoid dioxygenase 1	Up-regulated during WDS and combined HS-WDS (Lehr et al. 2022); up-regulated by HS (Guo et al. 2022)
cfQTL_19.1, trQTL_19.1	Vitvi19g02230	Gibberellin 2-beta-dioxygenase 1	Up-regulated by HS (Zha et al. 2020; Upadhyay and Upadhyay 2021)
ROS homeostasis
cfQTL_7.1	Vitvi07g01502	Peroxidase 17	Down-regulated by HS (Zha et al. 2020)
trQTL_9.1	Vitvi09g00375	Pyridoxal 5′-phosphate synthase-like subunit PDX1.2	Vitvi09g00375 up-regulated by HS (Liu et al. 2012; Rocheta et al. 2014; Jiang et al. 2017; Lecourieux et al. 2020; Zha et al. 2020)
	Vitvi09g04108
	Vitvi09g04109
trQTL_9.1	Vitvi09g01604	Putative Peroxidase 48	Down-regulated by HS (Upadhyay and Upadhyay 2021)
cfQTL_11.1	Vitvi11g00038	Peroxiredoxin Q	Involved in protecting photosynthesis (Carvalho et al. 2015)
cfQTL_12.2	Vitvi12g01624	Peroxidase 43	Down-regulated by HS (Rocheta et al. 2014)
Protein homeostasis
cfQTL_4.1	Vitvi04g00148	Thermospermine synthase ACAULIS5	Involved in protection from HS-induced damage (Sagor et al. 2013)
trQTL_5.1	Vitvi05g00545	Stress-induced-phosphoprotein 1	Co-chaperone involved in response to stress (Toribio et al. 2020)
cfQTL_7.1	Vitvi07g01393	Chaperone protein Dnaj A6, chloroplastic	Up-regulated by HS (Rocheta et al. 2014; Zhen et al. 2023); involved in the response to various stresses, including HS (Chen et al. 2021); playing a positive role in heat stress resistance (Zhen et al. 2023)
cfQTL_7.1	Vitvi07g01524	Heat shock 70 Kda protein, mitochondrial	Up-regulated by HS or combined HS-WDS (Rocheta et al. 2014; Zha et al. 2020; Ju et al. 2021; Guo et al. 2022; Liu et al. 2022)
cfQTL_7.1	Vitvi07g02588	Protein CHAPERONE-LIKE PROTEIN OF POR1, chloroplastic	Up-regulated by HS (Zha et al. 2020)
cfQTL_7.1	Vitvi07g03046	F-box protein FBW2	Down-regulated by HS (Rocheta et al. 2014)
trQTL_9.1	Vitvi09g00392	Heat shock 70 kDa protein 14	Up-regulated by HS and combined HS-WDS (Liu et al. 2014; Rocheta et al. 2014; Zha et al. 2020; Ju et al. 2021; Guo et al. 2022)
cfQTL_10.1	Vitvi10g00370	F-box/LRR-repeat protein 25-like	Vitvi10g01748 up-regulated by HS (Rocheta et al. 2014)
	Vitvi10g01748	F-box protein At1g61340-like
	Vitvi10g01749	F-box protein At5g03100
	Vitvi10g01751	Putative F-box protein At1g49610
cfQTL_11.1	Vitvi11g00044	20 kDa chaperonin, chloroplastic	Up-regulated by HS (Rocheta et al. 2014; Zha et al. 2020)
cfQTL_15.1	Vitvi15g01082	RING-H2 finger protein ATL68	Down-regulated by HS (Upadhyay and Upadhyay 2021)
cfQTL_15.1	Vitvi15g01083,	Late embryogenesis abundant protein At1g64065	Vitvi15g01084 up-regulated by HS (Rocheta et al. 2014)
	Vitvi15g01084*	Late embryogenesis abundant protein Lea14-A
cfQTL_17.1	Vitvi17g00171	Chaperone protein DnaJ	–
cfQTL_19.1	Vitvi19g01339	F-box/LRR-repeat protein At4g14103	–
	Vitvi19g01351	F-box/LRR-repeat protein At3g59190
	Vitvi19g02200	F-box/LRR-repeat protein At4g14096-like
Osmoprotection
cfQTL_7.1	Vitvi07g01389	Putative mannitol dehydrogenase	Vitvi07g02555 up-regulated by osmotic stress (Conde et al. 2015) and by HS (Lecourieux et al. 2020)
	Vitvi07g02553
	Vitvi07g02554
	Vitvi07g02555
	Vitvi07g02556
cfQTL_7.1	Vitvi07g04623	Probable mannitol dehydrogenase	Up-regulated by HS (Rocheta et al. 2014; Lecourieux et al. 2020)
Water and ion movement
cfQTL_3.2	Vitvi03g01288	zinc transporter 8	Down-regulated by HS (Liu et al. 2012)
trQTL_5.1	Vitvi05g00546	Sulfate transporter 3.1	Up-regulated by HS (Upadhyay and Upadhyay 2021)
trQTL_9.1	Vitvi09g00350	Sulfite exporter TauE/SafE family protein 3	Down-regulated by HS (Rocheta et al. 2014; Zha et al. 2020)
trQTL_9.1	Vitvi09g00399	Probable sulfate transporter 3.4	Up-regulated by HS (Rocheta et al. 2014)
cfQTL_10.1	Vitvi10g00385	Syntaxin-124	Up-regulated by HS (Zha et al. 2020)
cfQTL_15.1	Vitvi15g01072	Mechanosensitive ion channel protein 1, mitochondrial	Involved in maintaining redox homeostasis under HS (Lee et al. 2016)
cfQTL_17.1	Vitvi17g00162	Potassium transporter 1-like isoform X1/NF-X1-type zinc finger protein NFXL1	Down-regulated by HS (Liu et al. 2012)
Photosynthesis and respiration
cfQTL_4.1	Vitvi04g00151	NAD(P)H-quinone oxidoreductase subunit U, chloroplastic	Involved in HS response (Liu et al. 2014)
trQTL_9.1	Vitvi09g00361	Photosystem I reaction center subunit VI, chloroplastic	Down-regulated by HS (Guo et al. 2022)
cfQTL_10.1	Vitvi10g00372	Chlorophyll a-b binding protein 13, chloroplastic	Involved in HS response (Liu et al. 2014)
cfQTL_10.1	Vitvi10g00399	Internal alternative NAD(P)H-ubiquinone oxidoreductase A1, mitochondrial	Involved in HS response (Liu et al. 2014)
cfQTL_10.1	Vitvi10g04257	External alternative NAD(P)H-ubiquinone oxidoreductase B2, mitochondrial	Involved in HS response (Liu et al. 2014)
	Vitvi10g04258
cfQTL_13.2	Vitvi13g01328	Ribulose bisphosphate carboxylase/oxygenase activase, chloroplastic	Driver of heat tolerance (Kurek et al. 2007; Phillips et al. 2022)
cfQTL_13.2	Vitvi13g02302	Membrane-associated protein VIPP1, chloroplastic	Function in maintaining chloroplast membranes (Zhang and Sakamoto 2015)
Response to radiation
cfQTL_3.1	Vitvi03g01243	Protein FAR-RED IMPAIRED RESPONSE 1	Involved in the response to UV-radiation and temperature stress (Dai et al. 2022)
cfQTL_3.1	Vitvi03g01244	Protein FAR1-RELATED SEQUENCE 2	Involved in the response to UV-radiation and temperature stress (Dai et al. 2022)
cfQTL_3.2	Vitvi03g01304	Ultraviolet-B Receptor UVR8	Involved in UV-radiation protection (Loyola et al. 2016); up-regulated by HS (Zha et al. 2020)
cfQTL_4.2	Vitvi04g00693	Protein SUPPRESSOR OF PHYA-105 1	Involved in light signaling (Baumgardt et al. 2002)
cfQTL_13.2	Vitvi13g01385	Protein MAINTENANCE OF PSII UNDER HIGH LIGHT 1	Involved in protecting photosystem II from photodamage (Liu and Last 2015)
Other
trQTL_5.1	Vitvi05g00538	Pentatricopeptide repeat-containing protein At3g22690	Down-regulated by HS (Upadhyay and Upadhyay 2021)
cfQTL_7.1	Vitvi07g01482	Bifunctional riboflavin biosynthesis protein RIBA 1, chloroplastic	Candidate gene for photochemical reflective index (Tafesse et al. 2020)
trQTL_9.1	Vitvi09g01610	Pentatricopeptide repeat-containing protein At3g16010	Candidate gene for chlorophyll content (Tafesse et al. 2020); up-regulated by HS (Zha et al. 2020)
cfQTL_10.1	Vitvi10g02336	Polyphenol oxidase, chloroplastic	Possibly involved in abiotic stress responses (Boeckx et al. 2015); Vitvi10g04289 modulated by HS (Liu et al. 2012)
	Vitvi10g04287
	Vitvi10g04289
	Vitvi10g04292
cfQTL_12.2	Vitvi12g01622	Serine carboxypeptidase-like 42	Involved in response to different stresses (Wang et al. 2021a)
cfQTL_12.2	Vitvi12g01636	Pentatricopeptide repeat-containing protein At2g36240	Down-regulated by HS (Zha et al. 2020)
cfQTL_13.1	Vitvi13g04266	GDSL esterase/lipase	Involved in response to different stresses (Ding et al. 2019)

Discussion

Chlorophyll fluorescence and stomatal dynamics were evaluated in relation to heat stress in a grapevine mapping population grown under field conditions, with the aim of adding evidence to abiotic stress tolerance studies conducted so far on potted plants under semi-controlled conditions (Marguerit et al. 2012; Coupel-Ledru et al. 2014, 2016, 2024; Su et al. 2020; Brault et al. 2021; Trenti et al. 2021). Given the strong influence of environmental conditions on these physiological traits, accurate phenotyping required multiple evaluation sessions conducted over three consecutive years, utilizing an optimized protocol for both measurement and data processing. For QTL analysis, the difference between stress and non-stress conditions was normalized for control values to compensate for observed variations, as reported in a similar study (Talukder et al. 2014). Differences in plant chlorophyll fluorescence between datasets of the same year can be ascribed to differences in phenological stages. Namely, flowering and berry pea-size are the most susceptible stages to changing temperature (Venios et al. 2020). Differences between datasets collected in different years depend, instead, mainly on seasonal environmental conditions and on plants’ vigor. Maximum temperatures experienced by vines in various sampling sessions over the 3 years were comprised between 27.3 and 36.8 °C, with differences between morning and afternoon sessions comprised between 11.5 and 17.9 °C. Those temperatures should not be considered as severe heat stress for grapevine, because photosynthesis starts to be highly affected when temperatures exceed 40 °C. However, at 35 °C, heat may trigger the onset of several acclimation mechanisms within the photosynthetic apparatus (Venios et al. 2020). Therefore, climatic conditions experienced by plants in the field, despite not being high enough for severe stress, were suitable for the study’s purpose.

In this experiment, 16 stable QTL clusters were significantly associated with quantum efficiency of photosystem II (F_v/F_m) and other chlorophyll fluorescence parameters and explained 5.5–16.3% of the phenotypic variation (Fig. 3, Table 2, Table S 3.1, Online Resource 3). In addition, 10 stable QTL clusters were identified in relation to stomatal dynamics with an explained variance between 8.1 and 12.5% (Fig. 4, Table 3, Table S 3.2, Online Resource 3), one of which corresponds to a QTL cluster for chlorophyll fluorescence (cfQTL_19.1).

The identification of several QTL explaining a low percentage of total phenotypic variance for both chlorophyll fluorescence and transpirational parameters highlights the complex genetic architecture of the evaluated traits, similarly to what was reported for major crops.

Candidate genes by position in QTL clusters with explained variance above 10%, for which an involvement in heat stress response in grapevine has been reported (Liu et al. 2012, 2020a; Rocheta et al. 2014, 2016; Jiang et al. 2017; Zha et al. 2020; Ju et al. 2021; Upadhyay and Upadhyay 2021; Guo et al. 2022; Tan et al. 2023), allowed us to outline biological mechanisms implicated in heat stress response relevant in our experiment (Table 4).

Signal perception, transduction and transcriptional control

The initial stress signals (e.g., variations in membrane fluidity, osmotic or ionic effects, oxidative stress due to ROS (reactive oxygen species) generation, and protein denaturation) trigger downstream signaling cascades and transcriptional controls that activate stress-responsive mechanisms to re-establish homeostasis and protect/repair damaged proteins and membranes. Some broad groups of signal transduction molecules are the calcium-dependent protein kinases (CDPKs), mitogen-activated protein kinases (MAPK/MPKs), sugars (as signaling molecules), phytohormones, as well as nitric oxide (NO) and hydrogen peroxide (H₂O₂). Evidence of the activation of these mechanisms was provided by candidate genes within two of the most robust QTL clusters, cfQTL_12.1 and cfQTL_19.1. The gene Vitvi19g01284 codes for a C2 domain-containing protein belonging to the calcium-dependent lipid-binding (CaLB domain) family protein. Calcium signaling has been widely reported to play a crucial role in raising plant thermotolerance (Kang et al. 2024). While the serine/threonine-protein kinase EDR 1 (MAPK) gene (Vitvi12g00181) is known to be involved in the regulation of plant disease resistance but also in abiotic stress responses (Tang et al. 2005), including changing temperatures (Kim et al. 2003). Other protein kinases with supporting evidence from previous grapevine transcriptomic studies (see Table 4) are a somatic embryogenesis receptor kinase (Vitvi03g00614), a serine/threonine-protein kinase HT1 (Vitvi05g00540), a STRUBBELIG-RECEPTOR family member (Vitvi10g00417), a leucine-rich repeat receptor-like protein kinase (Vitvi13g01358), and a kinase-interacting protein (Vitvi05g00542); while Vitvi04g00728 and Vitvi15g01077 code for a histidine kinase 5 and a mitogen-activated kinase, respectively, for which an important role in ROS homeostasis is reported also in response to abiotic stimuli (Nakagami et al. 2006; Pham and Desikan 2012).

All major phytohormones, such as ABA, auxin, brassinosteroids, cytokinins, ethylene, gibberellins, jasmonic acid and salicylic acid, have been reported to play critical roles in plants’ response to heat stress as signaling molecules (Ahammed and Yu 2016). Studies on grapevine under heat stress, drought stress and combined stress, highlighted the involvement of ABA and of the enzyme NCED1 (9-cis-epoxycarotenoid dioxygenase), which catalyzes a very initial step in ABA biosynthetic pathway, in drought and combined stresses (Lehr et al. 2022). Moreover, in the experiment done by (Rocheta et al. 2014), both heat stress and water deficit stress led to moderate up-regulation of some ABA-responsive genes and ABA-related transcription factors, which makes ABA signaling genes important targets for abiotic stress resistance in a broad sense, not just for drought resistance. Based on these findings and on the observation that OsNCED1 confers heat stress tolerance in rice seedling plants (Zhang et al. 2022), we propose that the genes Vitvi19g01356 and Vitvi19g04443 may also contribute to heat stress response in grapevine. Other genes involved in ABA signaling are a homeobox-leucine zipper protein (Vitvi07g01488) and an ankyrin repeat-containing protein ITN1 (increased tolerance to NaCl1)-like (Vitvi09g00321) that may affect the ABA-mediated production of ROS (Sakamoto et al. 2008). A dehydration-responsive element-binding protein (Vitvi09g00323) as well as two ethylene-responsive transcription factors (Vitvi11g00045 and Vitvi11g00046) identified in this work, all belonging to the AP2/EREBP (APETALA2/ethylene-responsive element binding proteins) family of transcription factors, were reported to exhibit a strong differential expression upon heat stress, drought and combined stress exposure (see Table 4). Therefore, also ethylene seems to have a central role in both heat and drought stress pointing to a stress-mediated cross-talk in the ethylene signaling pathway, evidenced by similar expression of ethylene-responsive genes (Rocheta et al. 2014). Finally, genes coding for an auxin efflux carrier component (Vitvi04g01151), for a protein involved in auxin-mediated signaling (Vitvi07g01425) and for a gibberellin 2-beta-dioxygenase (Vitvi19g02230) support the role of these plant hormones in grapevine response to increasing temperatures. Other genes with a role in signal perception and transduction and with supporting evidence in literature are a dynein light chain LC6 (Vitvi09g00355), a 14–3-3-like protein (Vitvi07g01457 and Vitvi07g01498) and a trihelix transcription factor gene (Vitvi09g00393) (see Table 4).

In addition to the above signal transduction molecules, volatile terpenes have been reported to play a role in relieving oxidative and thermal stresses in plants (Loreto and Schnitzler 2010). In particular, Bertamini et al. showed how leaf monoterpene emission reduced photosynthetic downregulation under natural high temperature conditions in field-grown grapevine. This stress resilience seems to be mediated by the increased ability to counteract the negative effect of ROS either directly via scavenging, or indirectly via thylakoid membrane stabilization. However, the most likely hypothesis is that monoterpenes act as a signaling molecule (Bertamini et al. 2021). Interestingly, in this work we found a gene (Vitvi07g01408) coding for DXS (1-deoxy-D-xylulose-5-phosphate synthase), the first enzyme of the plastidial isoprenoid biosynthetic pathway. This gene is strongly expressed in young leaves and up-regulated by heat stress (see Table 4).

ROS homeostasis

Evidence of the involvement of mechanisms related to ROS homeostasis was found in different QTL clusters, with genes coding for three peroxidases (Vitvi07g01502, Vitvi09g01604, Vitvi12g01624), which were all reported to be down-regulated under heat stress, together with a peroxiredoxin Q (Vitvi11g00038) involved in protecting photosynthesis (Carvalho et al. 2015). Moreover, three genes coding for the positive regulator of vitamin B₆ biosynthesis pyridoxal 5′-phosphate synthase PDX1.2 (Vitvi09g00375, Vitvi09g04108 and Vitvi09g04109) were also found, the former of which is strongly up-regulated in different grapevine transcriptomic studies (see Table 4). Vitamin B₆ has an important role as antioxidant and its increase upon heat stress has been reported in literature (Moccand et al. 2014).

Protein homeostasis

Genes coding for heat shock proteins (HSPs) were found in the confidence intervals of four consensus genomic regions: two 70 kDa heat shock proteins (Vitvi07g01524 and Vitvi09g00392), a 20 kDa chaperonin (Vitvi11g00044), two chaperone proteins DnaJ (Vitvi07g01393 and Vitvi17g00171) and a chaperone-like (Vitvi07g02588). Heat shock proteins, like HSP70 and low molecular weight proteins, are widely known to be involved in plant heat stress response and acquisition of thermotolerance (Wahid et al. 2007), playing a central role in preventing the misfolding and aggregation of proteins and helping their proper refolding (Venios et al. 2020). They are also engaged in facilitating the proteolytic degradation of unstable proteins by targeting them to lysosomes or proteasomes. Our results are in line with findings from the literature reporting an overexpression of HSPs in grapevine in response to heat stress (Liu et al. 2014). Bernfur et al. also demonstrated the association of HSP21 to thylakoid membranes in heat stressed plants, suggesting a role in stabilization of the membrane (Bernfur et al. 2017). DnaJ proteins serve as co-chaperones of HSP70, playing crucial roles in the response to various stresses, including heat stress (Chen et al. 2021). For example, tobacco plants overexpressing VvDNAJA6 (Vitvi07g01393) and VvHSP70 showed a high-temperature tolerance phenotype, thanks to the accumulation of ROS-scavenging enzymes and osmotic substances like proline (Zhen et al. 2023). An increasing number of studies suggest indeed that the HSPs/chaperones can play a role in stress signal transduction/gene activation and interact with other stress-response mechanisms such as the production of antioxidants and osmolytes (Wang et al. 2004). Furthermore, we found a stress-induced phosphoprotein 1 (STIP1, Vitvi05g00545), also known as HOP, which is a co-chaperone of HSP70/HSP90 and is involved in plant heat stress response (Toribio et al. 2020). The synthesis of heat-shock proteins can be influenced by polyamines that are known to provide protection to plants from heat stress in different ways (Hasanuzzaman et al. 2013). In line with this, a gene coding for thermospermine synthase (Vitvi04g00148) has a potential role in protection from heat stress-induced damage (Sagor et al. 2013). With respect to protein fate, we also identified some genes implicated in ubiquitination and proteolysis (Vitvi07g03046, Vitvi10g00370, Vitvi10g01748, Vitvi10g01749, Vitvi10g01751, Vitvi15g01082, Vitvi19g01339, Vitvi19g01351 and Vitvi19g02200). Indeed, F-box-like proteins are known to play a major role in plant responses to various developmental and stress conditions, including heat (Guérin et al. 2021).

Osmoprotection

Six genes coding for mannitol dehydrogenase (Vitvi07g01389, Vitvi07g02553, Vitvi07g02554, Vitvi07g02555, Vitvi07g02556, Vitvi07g04623) were found in cfQTL_7.1. Together with amino acids and quaternary amines, polyol and sugars are known to play a role as osmolytes (Wang et al. 2003). Their primary function is to maintain cell turgor, but they can also act as ROS scavengers or chemical chaperones maintaining protein structures and functions. In particular, mannitol has been recognized to play a role in coping with heat stress-induced oxidative damage and excessive solar irradiance (Lecourieux et al. 2020).

Photosynthesis and respiration

Genes involved in the regulation of the photosynthetic process were found in several stable QTL clusters identified in this study. Among them are genes for a Photosystem I reaction center subunit (Vitvi09g00361), a ribulose bisphosphate carboxylase/oxygenase activase (Rubisco activase, Vitvi13g01328), a membrane-associated protein VIPP1 (Vitvi13g02302), some NAD(P)H-quinone oxidoreductase (Vitvi04g00151; Vitvi10g00399, Vitvi10g04257 and Vitvi10g04258), and a chlorophyll a-b binding protein (Vitvi10g00372). It is already known from the literature that photoinhibition caused by heat stress is related to Rubisco inactivation and to the thermal instability of Rubisco activase, which is involved in the step of Rubisco re-activation (Salvucci and Crafts-Brandner 2004). Moreover, it has been found that an enhanced thermal stability of Rubisco activase can confer higher thermotolerance and an improved photosynthetic activity under heat stress (Kurek et al. 2007; Phillips et al. 2022). VIPP1 is a membrane-associated protein with multiple roles, including the maintenance of the photosynthetic complex in thylakoid membranes (Zhang and Sakamoto 2015). Its involvement has been proposed in relation to membrane stress, such as for example osmotic stress, but heat stress is also known to cause damage to thylakoid membranes by increasing their fluidity, with consequences also for PSII (Wahid et al. 2007; Venios et al. 2020). Hence both Rubisco activase and VIPP1 could play a role in grapevine photosynthetic response to heat stress. Conversely, NAD(P)H-quinone oxidoreductase and chlorophyll a-b binding protein role in grapevine heat stress response is supported by a proteomic study (Liu et al. 2014).

Response to radiation

Two genes coding for a Far-Red Impaired Response 1 (FAR1, Vitvi03g01243) and a protein FAR1-Related Sequence (FRS, Vitvi03g01244) were found in cfQTL_3.1, a genomic region conserved in the 3 years of analysis and with explained variance 8.1–16.3%. FAR1 belongs to the wider FRS protein family and together with Far-Red Elongated Hypocotyls 3 (FHY3) it is known to be responsible for the translocation of phytochrome A in response to far red light (Wang and Wang 2015). Furthermore, FAR1/FHY3 are involved in UV-B signaling, in the circadian rhythm regulation, and in ROS homeostasis (Wang and Wang 2015). Moreover, Dai et al. reported their involvement in salt and temperature stress responses (Dai et al. 2022), which suggests a potential role in plant response not only to UV-light radiation stress but also to heat stress. In other QTL regions, additional genes involved in the response to UV light were found coding for an ultraviolet-B receptor UVR8 (Vitvi03g01304), a protein suppressor of Phytochrome A (PhyA) (SPA1, Vitvi04g00693), and a Maintenance of PSII Under High Light 1 (MPH1, Vitvi13g01385). The protein UVR8 has been studied also in grapevine, and its fundamental role in the UV-B signaling pathway was demonstrated, as it regulates a huge number of genes involved in UV-radiation protection (Brown et al. 2005; Loyola et al. 2016). Conversely, SPA1 is known to play a role in the light-signal transduction (Baumgardt et al. 2002), while MPH1 codes for a proline-rich protein that helps the maintenance of normal PSII activity under photoinhibitory stress and protects PSII from photooxidative damages (Liu and Last 2015).

Conclusions

For perennial fruit crops, this study on grapevine is the first attempt to identify QTL and candidate genes related to heat stress response in the vineyard by analyzing photosynthetic and transpirational phenotypes. Among the biological processes implicated in plant heat tolerance, signal perception and transduction emerge as central mechanisms in our experiment, considering that two of the four most conserved and significant genomic regions harbor genes encoding key signaling components. Notably, these include 9-cis-epoxycarotenoid dioxygenase (NCED) and serine/threonine-protein kinase EDR1. Though part of broad classes of signaling molecules, these proteins have documented roles in plant responses to high temperatures, supporting their potential as functional candidates for heat stress adaptation. Results obtained in this study further support proteostasis, together with osmoprotection, as key adaptive strategies for plant survival under high-temperature conditions. Several genes encoding heat shock proteins (HSPs) and molecular chaperones, in fact, were identified as functional candidates within the most stable genomic regions. These proteins are well-documented for their roles in protecting cellular proteins from heat-induced damage by facilitating proper folding, preventing aggregation, and assisting in refolding or degradation of misfolded proteins. Finally, the identification of genes involved in the maintenance of photosynthetic efficiency and the response to high light and radiation stress supports the hypothesis that preserving photosynthetic function is a critical adaptive response to prolonged heat exposure. These include genes encoding a Rubisco activase, essential for the regeneration of Rubisco activity under thermal stress; NAD(P)H-quinone oxidoreductases, which play a role in chloroplast redox homeostasis and electron transport; Far-Red Impaired Response 1 (FAR1), involved in light signaling and circadian regulation; and Maintenance of PSII Under High Light 1 (MPH1), which supports the stability and repair of photosystem II under high irradiance conditions. In addition to the candidate genes discussed, from this exploratory analysis other gene predictions in the stable QTL clusters emerged, whose role in grapevine adaptation to and recovery capacity from heat stress may be tested. Overall, this work paves the way for future research on grapevine resilience to rising temperatures, providing a robust workflow for the analysis of physiological traits in the field and suggesting thermotolerance candidate genes for advanced investigations.

Supplementary Information

Below is the link to the electronic supplementary material.

Authors contribution

SP: experimental design, experimental work, data analysis, manuscript drafting; LuC: conceptualization, funding acquisition, manuscript revision; MF: experimental design, manuscript revision; MSG: conceptualization, resources, funding acquisition, manuscript revision; LaC: conceptualization, experimental design, experimental work, manuscript revision.

Funding

Open access funding provided by Università degli Studi di Trento within the CRUI-CARE Agreement. This work was supported by the PhD project ‘Dissecting the genetic and physiological mechanisms of grapevine resilience to heat stress’ (C3A-UniTN, FEM and UniPD) and the Accordo di Programma from the Autonomous Province of Trento, and by the Italian Ministry of University and Research in the framework of the PRIN 2022 call – Prot. 20227EXKME.

Declarations

Competing interests

The authors declare no competing interests.