Given the global climate changes, heat stress is becoming a major challenge to crop production and food safety. As per Intergovernmental Panel on Climate Change, the current rate of global warming is 0.2°C per decade and is predicted to reach 1.5°C between 2,030 and 2,052 (https://www.bbc.com/news/newsbeat-48947573). Such an increase in temperatures leads to heat stress and costs the global economy US$2.4 trillion a year (https://news.un.org/en/story/2019/07/1041652). More than 15% of the global land area becomes exposed to high levels of heat stress with an additional 0.5°C increase to the 2°C (Sun et al., 2019). Heat stress, besides affecting producers directly, also reduces labor productivity (Kjellstrom, 2016), further compounding the effects of increasing temperature on crop yields. In recent years, shifts toward more sustainable and healthy diets, which are typically characterized by high consumption of vegetables and legumes, have been evidenced (Scheelbeek et al., 2018).

Chickpea (Cicer arietinum L.) is an important cool season grain legume crop cultivated in the arid and semi-arid regions across the globe. It is an excellent source of proteins, essential amino acids, vitamins, and minerals (Jukanti et al., 2012). Major chickpea producing countries are India, Australia, Pakistan, Turkey, Russia, Myanmar, Iran, Mexico, Canada, and USA. In India, Madhya Pradesh, Maharashtra, Rajasthan, Karnataka, Uttar Pradesh, and Andhra Pradesh are the major chickpea growing states. Although India is the largest producer of chickpea, in order to attain self-sufficiency by 2050, the chickpea production in the country needs to reach 16–17.5 Mt from an area of about 10.5 Mha with an average productivity of 15–17 q/ha (Dixit et al., 2019). Drought and heat, the two most important environmental factors, can cause more than 70% yield loss in chickpea (Varshney et al., 2019). Traditionally, chickpea requires prolonged winter for better growth and cultivation in the northern states of India. However, in the northern states, the pulse area especially chickpea cultivation was reduced due to the green revolution. Southern and Central India, where significant chickpea area increased, are exposed to drought and heat stresses. The rise in ambient air temperature (≥35°C) that coincides with the reproductive processes leads to various anomalies in reproductive events, especially during fertilization, pod formation, and pod filling in chickpea (Devasirvatham et al., 2013; Kaushal et al., 2013; Gaur et al., 2014).

The genetic mechanism of heat stress in different crop plants has been reviewed extensively (see Janni et al., 2020). In general, the impact of heat stress depends on the intensity, duration of exposure, and degree of the elevated temperature. In the case of legumes like chickpea, heat stress has deleterious effects on the morphology, physiology, and reproductive growth (Sita et al., 2017). The effects of heat stress on the development of various male and female tissues in different legume species have been reviewed recently (Liu et al., 2019). In the case of legume crops, heat shock proteins (HSP), HSP gene families, and various metabolites were reported to control heat stress response (see Janni et al., 2020). Heat stress adversely affects pollen viability, fertilization, and seed development, which leads to a reduced harvest index consequently, and these events greatly impact chickpea yield. In the cool season, legumes such as chickpea, lentil, faba bean, and field peas, the temperature above 30°C lead to yield losses (Jiang et al., 2015; Bishop et al., 2016; Bhandari et al., 2017). As heat stress is a complex trait governed by many genes/QTLs, breeding for heat stress tolerance in chickpea remains challenging (Krishnamurthy et al., 2011; Devasirvatham et al., 2013). Therefore, the effects of heat stress on chickpea growth, development, and yield are important to understand by observing agronomic traits to develop high-temperature-tolerant cultivars.

Genomic revolution, during the last two decades, simplified understanding of the complex responses to biotic and abiotic stress in several crop plants (Roorkiwal et al., 2020; Thudi et al., 2020). Chickpea research community has access to genome sequence (Varshney et al., 2013), genome-wide variations among diverse germplasm lines at the sequence level (Thudi et al., 2016a,b; Varshney et al., 2019) for trait dissection, and the development of climate-resilient chickpea varieties (Mannur et al., 2019; Bharadwaj et al., 2020). The genotyping-by-sequencing (GBS) approach has been extensively used for single-nucleotide polymorphism (SNP) discovery and mapping traits in several crops for genetic research and breeding applications (Chung et al., 2017), including chickpea (Jaganathan et al., 2015; Thudi et al., 2020). Besides proteomic and metabolomic approaches to understanding the molecular mechanism of heat tolerance (Parankusam et al., 2017; Salvi et al., 2018), efforts were made to map QTLs and markers associated with heat tolerance in chickpea (Thudi et al., 2014; Paul et al., 2018a; Varshney et al., 2019; Roorkiwal et al., 2020).

In this study, we reported the construction of a high-density genetic map using SNPs derived from the GBS approach and major QTLs for phenological, physiological, yield, and yield-related traits based on phenotyping of recombinant inbred line (RIL) population (DCP 92-3 × ICCV 92944) under two environments (normal and late sown) for 2 years (2017–2018 and 2018–2019). In addition, we also reported the potential candidate genes implicated for heat tolerance in the QTL regions.

In the context of climate change, every degree increase in aerial temperature has a severe impact on crop production, especially on chickpea that is predominantly cultivated on residual soil moisture in marginal environments (Gaur et al., 2014). Therefore, understanding the nature, impact, and molecular mechanisms of heat stress tolerance will help in designing strategies to overcome production losses. In chickpea, previously, very few studies were focused on understanding the nature, impact, and existing diversity in germplasm lines as well as identifying the genomic regions responsible to some extent. In this study, we reported major QTLs and novel genes in these genomic regions, which are associated with responsible for heat stress tolerance.

Late sowing, a simple and effective field screening technique for reproductive-stage heat tolerance in chickpea developed at the ICRISAT, Patancheru, India, was adopted for phenotyping the RILs for heat stress tolerance. The late sowing approach was adopted earlier in understanding the genetic variability for heat stress among genotypes as well as in identifying the genomic regions responsible for heat stress tolerance; for instance, in cool season crops, namely, chickpea (Paul et al., 2018b), wheat (Sareen et al., 2020), brassica (Branham et al., 2017), and rice (Prasanth et al., 2016). As the selection was based on yield per se results in a slower response because of genotype × environmental interactions, we also phenotyped the mapping population for physiological traits like CMS, NDVI, NBI, and CHL, which could be used as an indirect selection criterion to improve heat tolerance in chickpea as this was used in other crop plants. A sufficient amount of genetic variability for various phenological, physiological, and yield-related traits was noted in both the parents, and the RIL population was studied under NS and HTS environments for both years. The similar results were also reported earlier in chickpea based on evaluating the germplasm lines as well as one RIL population under HTS environment (Krishnamurthy et al., 2011; Devasirvatham et al., 2013; Paul et al., 2018b). High heritability for physiological traits like CMS, NDVI, NBI, and CHL contents under HTS environments indicates that the selection for heat tolerance relying on these traits could be effective. Earlier, heat stress was reported to reduce the total CHL and showed moderate to high heritability for NDVI, CMS, and CHL content under stress condition in wheat (Bhusal et al., 2018; Condorelli et al., 2018; Pradhan et al., 2020), maize (Naveed et al., 2016), carrot (Nijabat et al., 2020), and pea (Tafesse et al., 2020). Low heritability for HI trait under HTS environments observed in this study indicates that the election for this trait will not enhance yield under stress. As yield traits remain the primary objective for improving the heat tolerance in all crop plants including chickpea, a positive and significant correlation among the yield and yield-related traits especially, FP, SYPP, and BYPP could serve as an important parameter for developing heat-tolerant chickpea genotype.

In this study, using a RIL population with a dense genetic map and phenotyping under NS and HTS allowed us to precisely identify major QTLs for heat stress in chickpea. Our genetic map has approximately 3-fold more markers compared to the previous study reporting the QTLs for heat stress tolerance (Paul et al., 2018a). In addition to reporting QTLs for phenological, yield, and yield-related traits under heat stress environments, to the best of our knowledge, this study is the first comprehensive study that reports QTLs for physiological traits like CHL, NBI, NDVI, CMS, and NPP in chickpea. A better understanding of phenology in response to heat stress will enable designing the breeding strategies. Minor QTLs for DFI were identified on CaLG06 and CaLG08, while a major QTL was identified for DM on CaLG01. Physiological traits like CMS, NDVI, and NBI, to date, have been used as proxy for grain yield under stress mostly in cereals (ElBasyoni et al., 2017; Bhusal et al., 2018; Condorelli et al., 2018; Getahun et al., 2020; Khanna-Chopra and Semwal, 2020). In this study, we reported major QTLs for these traits in chickpea, which may be used, after validation, for marker-assisted breeding for heat stress tolerance in chickpea.

In the case of chickpea, four flowering time genes (efl-1 from ICCV 96029, efl-3 from BGD 132, and efl-4 from ICC 16641) and their allelic relationships were reported (Gaur et al., 2015), and major QTLs corresponding to these genes were mapped on CaLG04, CaLG08, and CaLG06, respectively (Mallikarjuna et al., 2017). Furthermore, marker trait associations for flowering time were reported earlier (Thudi et al., 2014; Upadhyaya et al., 2015; Varshney et al., 2019). However, in this study, we reported QTLs for DFI on CaLG06 and CaLG08 under heat stress environments as well as based on the pooled data of 2 years. Furthermore, interestingly, QTLs for DPI and DFI co-localized or mapped in the same genomic region under HTS environments of both years as well as based on pooled data. These observations indicate that introgression of one of the traits simultaneously improves both the traits, which are key for achieving resilience to heat stress.

Earlier two major genomic regions harboring QTLs for heat tolerance-related traits were mapped on CaLG05 and CaLG06; however, none on the QTLs explained >20% PVE (Paul et al., 2018a). Nevertheless, HI in this study explained >30% PVE. Similarly, CaLG04 also harbored QTLs for five traits (CHL, CMS, NDVI, DPF, and 100SDW), among these QTLs for traits CHL and 100SDW explained >30% PVE. Except HI (PVE 39.13%, NS 2018), none of the QTLs mapped on CaLG06 had PVE >15%. Similarly, QTLs for traits like CHL, NDVI, and HI were mapped on CaLG05 which explained 6.78–18.69% PVE. Earlier, a genomic region refereed as “QTL-hotspot” was reported to harbor several QTLs for different drought tolerance-related traits including 100SDW on CaLG04 (Varshney et al., 2014). A genomic region on CaLG07 harbors QTLs explaining >30% PVE for DPI and 100SDW as well as QTL for NBI explaining >10% PVE. For 100SWD, a total of four major QTLs were identified on CaLG01, CaLG04, and CaLG07 under HTS, and no QTLs were detected under NS. For SYPP trait, two major QTLs were identified on CaLG02 and CaLG06 under HTS. Only one QTL was identified on CaLG06 under NS. However, yield-related QTLs were not consistently recorded under all the conditions suggesting their environmental specific expression. Likewise, QTLs contributing to pods per plant, seed yield per plant, and seed weight were reported on CaLG01 and CaLG06 (Bajaj et al., 2015; Srivastava et al., 2016). In the case of cowpea, four QTLs were identified for pod set number per peduncle under HS and markers, which were utilized in breeding applications (Lucas et al., 2013; Pottorff et al., 2014). Similarly, in lentils, a major QTL controlling the seedling survival and pod setting traits under heat stress was noticed (Singh et al., 2017). In addition, QTLs for SYPP and BYPP (full names) were mapped in the same genomic region under NS environments of both years as well as based on pooled data. For yield and yield-related traits like DPI, DPF, and SYPP under HTS major alleles were contributed by ICCV 92944. For the 100SDW trait under HTS major alleles were contributed by DCP 92-3. On the other hand, almost all of the traits DPI, BYPP, and SYPP were contributed by DCP 92-3 under NS. For the trait HI under NS major alleles was contributed by both parents. In addition to these major QTLs, several QTLs were identified that were environmentally specific under NS and HTS, which only appeared in this study in the first year (NS or HTS). A total of nine major QTLs were located on CaLG06, which highlight the importance of this region in the heat tolerance mechanism in chickpea. Some QTLs were largely affected by environmental factors and that could be detected in only one season, and for these QTLs, further verification should be required.

HSP genes play a pivotal role in heat stress tolerance. In this study, 32 genes were identified in the QTL regions of CaLG01, CaLG02, CaLG04, and CaLG07. Similarly, in the case of soybean, 38 Hsfs were identified that were located on 15 soybean chromosomes (Li et al., 2014). HSP gene families were reported to be involved in drought and heat stress responses in soybean seedlings (Zhang et al., 2015; Das et al., 2016; Liu et al., 2019). HSP90 gene families in five legumes and expression profiles in chickpea were reported earlier (Agarwal et al., 2016). Furthermore, based on genome-wide association studies, especially eight flowering time-regulating genes (efl1, FLD, GI, Myb, SFH3, bZIP, bHLH, and SBP) were reported (Upadhyaya et al., 2015). The genes reported in this study can be further explored for haplotypes based on the germplasm sequence information available in the public domain that has the potential for genetic improvement of the trait (Varshney et al., 2020). In addition, pollen-specific leucine-rich repeat extensin-like protein 1 genes identified in the QTL regions were reported to synergistically maintain pollen tube cell wall integrity; thus, they play critical roles in pollen germination and pollen tube growth (Wang et al., 2018). Recently, cloning of SHY in tomato, a pollen-specific gene that encodes a leucine-rich repeat (LRR) protein, demonstrated its role in a signal transduction pathway mediating pollen tube growth (Guyon et al., 2004).

In this study, we identified a total of 37 major QTLs across the genome for 12 traits. DFI, DPI, and DM are the key traits for escaping the heat stress in chickpea especially reproductive heat stress that hampers chickpea production. In this study, we reported major QTLs explaining >30% PVE for these key traits that contribute to yield under heat stress. In addition, we also reported for the first time major QTLs for proxy traits (physiological traits like CHL, NBI, NDVI, and CMS). Furthermore, 32 candidate genes in the QTL regions that encode the HSP, heat shock transcription factors, genes are involved in flowering time regulation as well as pollen-specific genes. The major QTLs reported in this study may be useful in molecular breeding for developing heat-tolerant superior lines or varieties.

The original contributions presented in the study are publicly available. This data can be found here: NCBI repository, BioProject ID PRJNA695065 (http://www.ncbi.nlm.nih.gov/bioproject/695065).

UJ and MT conceived the idea. UJ, HN, and RJ conducted the experiments. UJ performed the statistical analysis. MT, UJ, RP, and RV prepared the manuscript. VV and PB performed the GBS analysis. NS contributed to the development and phenotyping of mapping population. AC and RV contributed to consumables and the generation of genotyping data. All authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.