QTL-Seq Identifies QTL for Relative Electrical Conductivity Associated with Heat Tolerance in Bottle Gourd (Lagenaria Siceraria)

Heat is a major abiotic stressor seriously affects watermelon produce; however, its effects may be mitigated through grafting onto heat tolerant bottle gourd rootstock. To understand the genetic basis and select reliable markers are necessary in the variety breeding process. In this paper, the relative electric conductivity (REC) was used as a visual indicator for heat stress to study the genetics and SNP marker of heat tolerance in bottle gourd via QTL-seq approach. The results showed that recessive inheritance and a major QTL locus controlled REC related to heat tolerance. Seven heat-tolerant loci (qHT1.1, qHT2.1, qHT2.2, qHT5.1, qHT6.1, qHT7.1, and qHT8.1) exhibited high Δ(SNP-index) values (ranging from 0.19-0.38), and of these, the greatest value (0.32, P < 0.01) and greatest number of detected SNPs (9052) were found in chromosome 2 at qHT2.1 region (11.03 - 19.25 Mb); qHT2.1was taken as a promising major QTL for heat tolerance in bottle gourd. A total of 34 putative candidate genes related to heat stress were detected within the qHT2.1 region. Three polymorphic SNPs (BG_GLEAN_10022642, BG_GLEAN_10022727, and BG_GLEAN_10022589) that are involved in pollen sterility, intracellular transport, and signal recognition were validated and exhibited significant marker-trait association that could be used in heat tolerant molecular breeding for bottle gourd. The qHT2.1region is an important finding can be used for fine mapping and discovering new genes associated with heat tolerance in bottle gourd.


Introduction
Heat stress negatively affected physiological processes, reproduction, and adaptation in plants [1], and changes in the climate have exacerbated such effects [2]. Watermelon (Citrullus lanatus var. lanatus) dominates the fruit market in the world [3], its growth frequently restricted by high temperatures reach more than 35℃ during summer months [4]. Bottle gourd (Lagenaria siceraria (Mol.) Standl.) tended to be the rootstock for watermelon to reduce heat stress and improve performance [5,6]. Phylogenetic tree analysis has demonstrated the close genetic relatedness of bottle gourd and watermelon [7,8], explains the high grafting survival rate and reduction in taste of bottle gourd in grafted watermelon.
Bottle gourd originated from Africa; it is widely distributed across the tropics [9], where long-term exposure to high temperatures has enhanced its tolerance to heat stress, albeit with wide variation among genotypes [10,11]. Understanding the genetics of heat tolerance in bottle gourd and identification of reliable markers are necessary for the development of novel bottle gourd rootstocks for heat adaptation improvement of scion; however, underlying tolerance characteristics remain unclear.
Heat tolerance is believed to be a polygenic trait that renders the introgression of multiple favorable alleles into cultivars problematic [12]. There have been attempts to deconstruct stress tolerance into measurable components for accurate phenotyping, with the aim that QTLs associated with heat tolerance may then be identified and suitable alleles may be introgressed into elite genetic backgrounds [13]. Cell membrane stability, which has been used as an indicator of heat stress [14], may be quantified by relative electric conductivity (REC). REC is highly sensitive to abiotic stress [15] and has been used in studies of abiotic stress tolerance in a range of crops, including salinity-alkalinity tolerance in muskmelon [16], drought tolerance in Perennial ryegrass [17], and cold tolerance in alfalfa [18].
Xu et al. [19] first reported partial sequencing of the bottle gourd genome using the 454 GS-FLX Titanium sequencing platform, from which 400 SSR markers were developed. RAD-Seq [20] technology has also been applied to an F 2 bottle gourd population for SNP and insertion-deletions marker development [21,22]. Wu et al. [8] reported a high-quality bottle gourd genome sequence was constructed using de novo genome assembly that reconstructed the most recent common Cucurbitaceae ancestor genome through comparison with available extant modern cucurbit species genome resources [23][24][25].These studies provide an available reference genome and useful molecular markers for the study of bottle gourd agronomic traits.
Many crop plant traits, such as heat stress tolerance, are quantitative, so identification of quantitative trait loci (QTL) may allow efficient crop breeding through marker-assisted selection (MAS) [26]. However, deriving the molecular markers for genetic linkage analysis, QTL analysis is labor-intensive and time-consuming [27]. The bulked-segregant analysis (BSA) provides an effective and simple way to identify molecular markers tightly related to target genes or major QTLs and recent advancements in next-generation sequencing (NGS) technologies provide rapid and efficient QTL identification. For example, Takagi et al. [27] proposed QTL-sequencing (QTL-seq) to detect QTLs that control partial resistance to rice blast using recombinant inbred lines and seedling vigor in F 2 populations; candidate QTLs detected using the QTL-seq method were confirmed using traditional QTL analysis. The QTL-seq approach has been widely used in a range of crops, such as sunflower [28], rice [29], sorghum [30], potato [31], and cucumber [32,33], for the efficient detection of QTLs for complex quantitative traits. The QTL detection for heat tolerance has been studied in rice [34], chickpea [35],wheat [36], and tomato [37]. In this study, we used QTL-seq to identify the major QTL for heat tolerance in bottle gourd, based on REC, from which polymorphic SNPs and putative candidate genes for heat tolerance were identified using available bottle gourd genome sequence information [21].

Materials and Methods Plant Material and Phenotyping
We selected the heat tolerant inbred line L1 (P17) and sensitive inbred line L6 (P23) as parent material, based on previous work [38], from which F 1 and F 2 segregating populations were created. We germinated three replicates of seed from147, 56 and 60 F 2 individuals of L1, L6 and F 1 plants, respectively. The seedlings were planted in 32-hole plastic plugs (230cm 3 /hole) filled with nursery substrate (2:1 mix of turf: vermiculite) and grown in a phytotron growth chamber set to a 16 h light/8 h dark cycle (30,000lux) of 25/18℃, with 80% relative humidity. The first true leaf of the seedlings was collected for DNA extraction, and when the third true leaf began to expand, the plants were moved into a high temperature controlled phytotron (40℃, with 80% relative humidity and 30,000lux) for 6h continuous heat exposure.
The electric conductivity (EC) of leaves was measured before and after the heat treatment to assess cell membrane damage, as described by Zhou and Leul [39] and He et al. [40] with modifications. In brief, sampled leaves were washed using deionized water, cut into 0.5-cm pieces, and immersed in deionized water for 30min; then, conductivity of the solution was measured using a conductivity meter (PHSJ-3F, Jingmi Instruments Co., Ltd., Shanghai, China) and recorded as S1. After boiling the leaf samples for 15 min, EC of the solution at room temperature was measured again and recorded as S2. REC was calculated as S1/ S2×100%.
We used analysis of variance (ANOVA) and t-tests to determine differences in heat tolerance, as indicated by mean (±SE) LRI, among P 1 , P 2 and F 1 plants at P<0.05 using SPSS 13.0 software [16]. For the F 2 population, the frequency distribution of LRI was plotted as a histogram and the normal distribution was fitted using the GaussAmp function in ORIGIN 9.1 software [44].

DNA Sequencing and Analysis
Genome DNA was isolated from young leaves of parental lines L1 and L6, and 147 F 2 plants using the CATB method [45]. For QTL-seq, two DNA pools (heat tolerance: T-pool; heat sensitive: S-pool) were constructed by mixing an equal amount of DNA from nine heat tolerant (LRI = 1-5%) and nine heat sensitive (LRI = 55-77%) individuals from the F 2 population.
A library of around 500 bp insert size was constructed and pair-end sequenced (parent average 15×coverage; progeny average was 20×coverage; read length: 150 bp) on an Illumina Hi-Seq4000 at Shanghai Biozeron Co., Ltd.
To identify candidate regions for heat tolerance QTLs, we calculated and plotted two parameters (SNP-index and Δ(SNP-index) for T-pools and S-pools [32,49]. For the polymorphic SNPs located in the genomic regions harboring the major heat tolerance QTL, further functional annotation was completed using available bottle gourd genome data (http://cucurbitgenomics.org/organism/13) and Uniprot database (www.uniprot.org) using ANNOVAR analysis (http://www.openbioinformatics.org/annovar/) to detect putative candidate genes [21].

SNP Marker Development and Selection
Five candidate SNPs were selected, based on the annotation and classification of putative candidate genes located in the QTL of qHT2.1 interval; the flank sequences were used for PCR primer development using Primer 6.0 soft (http://www.PromerBiosoft.com). The five SNPs were genotyped for four heat tolerant and four sensitive F 2 individuals using PCR and Sanger sequencing, and sequence results were visualized and checked using SnapGene software. We used trait-marker association to select the promising SNP markers.

LRI Indicator of Heat Tolerance
There were differences in mean LRI between L1 (P17: 9.88±5.00%) and L6 (P23: 47.26±10.78%) under heat stress conditions (Table 1). Mean LRI of the F 1 was greater (31.98±6.64%) than the mid-parent value (28.32±7.89%), and there was a difference in LRI between F 1 and L1; there was no difference in LRI between F 1 and L6. The LRI of the F 2 progeny ranged from 0.34 to 97.49% (mean: 29.68±8.52) ( Table 1). Transgressive segregation was observed on the both sides of the distribution, because distribution of the three F 2 populations was slightly skewed towards L1 (P17) and showed Gaussian segregation (Fig 1A, B, C).

Sequencing Data
Sequencing yielded 31.9, 32.7, 40.9, and 40.9 M 150-bp paired-end raw reads from L1, L6, T-pool, and S-pool, respectively. After trimming and filtering, more than 80% of reads were mapped to the bottle gourd reference genome (

QTL Detection for Heat Tolerance by QTL-seq
The SNP-index graphs were generated for the T-pool (Fig 2A) and S-pool (Fig 2B), and Δ(SNP-index) was plotted against the genome positions ( Fig 2C). By examining the Δ(SNP-index) plot, we identified seven genomic positions that exhibited high Δ(SNP-index) values (Table 3) Fig 2). Thus, a promising major QTL controlling heat tolerance appeared to be harbored in the 11.03 to 19.25 Mb regions on chromosome 2 and was designated as qHT2.1.

Identification of Heat Tolerance Candidate Genes in qHT2.1 Intervals
The full intervals of the regions covered by the seven major Δ(SNP-index) peaks were wide and ranged 8.22 Mb for qHT2.1 ( Table 3). The 9052 SNPs in this region comprised 527 genes, of which 279 were annotated by BLAST analysis against the non-redundant protein database (S1Table, [50]) to leave 62 nonsynonymous and stoploss type of SNPs with Δ(SNP-index) ≥0.5 (S2 Table). Gene ontology classification revealed that these 62 genes were mainly associated with biological processes (Fig 3), such as metabolism, cellular components, such as the cell membrane, and molecular function, such as catalytic activity. The biological processes of the 62 candidate genes were retrieved from the available bottle gourd genome data and Uniprot database, 34 genes related to heath stress were determined, along with their corresponding biological processes, comprising pollen and flower sterility, oxidative stress response, autophagy, and abiotic stress response (

Genetics of Heat Tolerance in Bottle Gourd
Plants have evolved a range of metabolic responses, such as antioxidant activity, membrane lipid unsaturation, protein stability, gene expression and translation, and accumulation of compatible solutes [68], to cope with heat stress through the activation of stress-response genes [69]. Heat tolerance, which is characterized by a series of phenotypic and physiological traits, may be summarized as the quantitative trait [70]. Poly genes are known to play a critical role in the control of heat tolerance in crop plants, such as wheat [71,72], rice [34], and broccoli [73,74], although there are reports monogenic and oligogenic responses to heat tolerance [75,76]. Previous understanding of inheritance patterns of heat tolerance in bottle gourd was limited, and our study is the first to report on the recessive inheritance and a major QTL locus for heat tolerance in this crop plant, based on REC using a QTL-seq approach. The plants in the F 1 progeny of the L1×L6 cross were heat sensitive, indicating that heat tolerance is a recessive trait in bottle gourd, as confirmed through genotyping of the candidate marker SNPs 2 (BG_GLEAN_10022642) and 31 (BG_GLEAN_10022727) as homozygous CC or TT, respectively, that indicate heat tolerance, and A_ (AA, AC or AT) that indicate heat sensitivity. These results also infer the presence of a dominant locus of AA in heat sensitive parent material.
The distribution of the F 2 population was skewed towards L1, indicating that a major QTL locus controlled heat tolerance-related LRI. According to the principle of QTL-seq, described by Takagi et al. [27], the SNP-index graphs of the T-pool and S-pool were expected to be identical for the genomic regions, without phenotypic differences (heat tolerance), whereas the genomic region(s) harboring heat tolerance QTL(s) were expected to exhibit unequal contributions from P 1 and P 2 parental genomes. Furthermore, the SNP-indexes of these regions for T-and S-pools should appear as a mirror image around SNP-index = 0.5. Although seven QTLs were identified using QTL-seq, only qHT2.1 was identified as the most promising major QTL for heat tolerance, because of the greatest Δ(SNP-index) value (0.32, P <0.01) and number of detected SNPs (9052), and clear mirroring of SNP-index plots. Similar results have been reported for African rice (Oryza glaberrima Steud.), in which a major QTL that controlled survival rate of seedlings exposed to high temperatures was governed by the Thermo tolerance 1 gene encoded in the 23S proteasome subunit, which degrades ubiquitinated cytotoxic denatured proteins that are formed due to high temperature stress [77]. Major QTL control of complex traits, such as heat tolerance, may be explained by inheritance patterns of the trait [13] and action of major regulator genes that may switch off subordinate genes, if a key gene is mutated [78].

Annotation of Heat Tolerance Candidate Genes
Understanding of functional genomics in plants facilitates the elucidation of candidate genes and their relationship with traits [35]. Heat stress affected pollen and flower sterility [79]. Of the 34 candidate genes, four genes were identified to have key roles in gametogenesis [51], pollen development [52,53] or Ogura cytoplasmic male sterility [54]. Heat-stress also caused cellular damage by oxidative stress and toxicity due to the reactive oxygen species (ROS) formation [80,81]. Antioxidants scavenge reactive oxygen species to mitigate oxidative. Four putative candidate genes were found to have a function in defying oxidative stress and recovering plants from heat-stress damage. Many transcription factors, binding factors, intracellular transporters and enzyme responsible for abiotic tolerance in different crops were reported earlier [35]. In addition, signalling molecule [59], autophagy [60] play important roles in plant responses to stress.

Annotation of Candidate SNP
In this study, three SNPs with contrasting function annotation were found to be associated with heat tolerance in bottle gourd, including SNP2 (BG_GLEAN_10022642; annexin 5) associated with pollen and flower sterility; SNP 31 (BG_GLEAN_10022727; AP-3 adaptor complex) related to intracellular transport; and, SNP 16 (BG_GLEAN_10022589; serine/threonine-protein kinase) involved in signal recognition; which exhibited significant marker-trait association and could be attractive option to identify the heat tolerance bottle gourd genotypes.
A common response of crop plants to temperature stress is significant yield loss, as a result of effects on spikelet fertility in rice [82], pot set in lentil [78], and pollen sterility in canola [79]. Annexins belong to a ubiquitous family of proteins present in eukaryotic organisms [83] localized in various subcellular compartments [84], and are known to be involved in a variety of cellular processes, such as maintenance of vesicular trafficking, cellular redox homeostasis, actin binding, and ion transport [85], due to their calcium-and membrane-binding capacity. Annexin 5 is involved in pollen grain development and germination, and pollen tube growth through the promotion of calcium-modulated endo-membrane trafficking [86]. For example, down-regulation of Arabidopsis annexin5 in transgenic Ann5-RNAi lines has been shown to cause sterility in pollen grains [52].
Adaptor proteins that are involved in protein trafficking and sorting [87][88][89] may recognize cargo and coat proteins during vesicle formation [90]. AP-3 was first identified in mammalian cells and likely functions as a clathrin adaptor [91]; losses in AP-3 function reduce seed germination potential [92,93], due to mistargeted protein S-ACYL transferase10 that is critical for pollen tube growth during dynamic vacuolar organization in Arabidopsis [94].
Serine/threonine-protein kinases (STKs) are involved in signal transduction networks to coordinate growth and differentiation of cell responses to extracellular stimuli [95]. In a range of smut pathogenesis-related biological processes, Huang et al. [59] reported that STKs may act as receptors or signaling factors, such as in Ca 2+ signaling, that then activate defense responses. Umezawa et al. [96] reported that SRK2C is annotated as a putative and osmotic-stress-activated STK in Arabidopsis; it may mediate signal transduction and regulate a series of drought stress-response genes that enhance expression of 35S:SRK2C-GFP to improve drought tolerance in plants.

Conclusion
To understand genetic basis and select reliable markers are important in the variety breeding process for bottle gourd. The accurate QTL-seq results have clearly shown that the recessive heat tolerance inheritance and three polymorphic SNPs involved in pollen sterility, intracellular transport, and signal recognition were exhibited significant marker-trait association. Intensity of global warming will increase in next two to three decades and will exacerbate the negative effects on Cucurbitaceae plants. It will help in identification of heat tolerance genotypes in breeding program of bottle gourd for further heat tolerance strengthening of other Cucurbitaceae plants by grafting, such as those for watermelon. The results also revealed the novel region (from 11.03 to 19.25 Mb on chromosome 2) harbored qHT2.1, which first provide the promising direction for further exploration on heat tolerance in bottle gourd for fine mapping to discover the new gene and understanding the complex mechanisms.
Supporting information S1