lnheritance of steroidal glycoalkaloids in potato tuber flesh

PENG Zhen , WANG Pei , TANG Die, SHANG Yi LI Can-hui HUANG San-wen, , ZHANG Chunzhi

1 The CAAS-YNNU Joint Academy of Potato Sciences, Yunnan Normal University, Kunming 650500, P.R.China

2 Genome Analysis Laboratory of the Ministry of Agriculture and Rural Affairs, Agricultural Genomics Institute at Shenzhen,Chinese Academy of Agricultural Sciences, Shenzhen 518124, P.R.China

3 Key Laboratory of Biology and Genetic Improvement of Horticultural Crops of the Ministry of Agriculture and Rural Affairs,Sino-Dutch Joint Laboratory of Horticultural Genomics, Institute of Vegetables and Flowers, Chinese Academy of Agricultural Sciences, Beijing 100081, P.R.China

Abstract Potato (Solanum tuberosum L.) is the third most important food crop worldwide after wheat and rice in terms of human consumption. A critical domestication trait for potato was the decrease of toxic steroidal glycoalkaloids (SGAs) in tuber flesh.Here, we used a diploid F2 segregating population derived from a cross between S. tuberosum and the wild potato species Solanum chacoense to map the quantitative trait loci (QTLs) associated with the regulation of SGAs content in tuber flesh.In a three-year study, we identified two QTLs on chromosomes 2 and 8 affecting SGAs content in tuber flesh. The QTL on chromosome 8 harbors 38 genes that are co-expressed with the GLYCOALKALOID METABOLISM genes. These findings lay the foundation for exploiting the genes controlling SGAs content in tuber flesh and they provide a theoretical basis for the use of wild germplasm in potato breeding.

Keywords: steroidal glycoalkaloids, domestication, tuber flesh, QTL mapping

1. Introduction

Potato (Solanum tuberosum L.), the third most important food crop worldwide after wheat and rice in terms of human consumption (Stokstad 2019), was domesticated 8 000-10 000 years ago. A major domestication trait was the reduced levels of antinutritional compounds steroidal glycoalkaloids (SGAs) in potato tubers. SGAs are toxic specialized metabolites primarily found in members of the Solanaceae and Liliaceae plant families. SGAs have complex, diverse chemical structures and exhibit a variety of pharmacological activities, including anti-tumor, antibacterial and anti-inflammatory properties (Lee et al. 2004; Friedman et al. 2005; Friedman 2006). Toxic SGAs protect plants from pest and pathogen damage, however, they are dangerous to humans and other animals that consume them (Friedman 2006; Mweetwa et al. 2012). How the SGAs content in tubers decteased during domestication remains largely unknown.

In potato, α-solanine and α-chaconine are often referred to as total SGAs because they account for ＞90% of the glycoalkaloids in the tuber (Allen et al. 1986). A series of GLYCOALAKLOID METABOLISM (GAME) genes involved in SGA biosynthesis in potato and tomato was recently identified on chr06 (GAME8), chr07 (GAME6, GAME7 and GAME11) and chr12 (GAME4 and GAME12) (Itkin et al.2013). The metabolic pathway from acetyl-CoA to the formation of cholesterol, solanidine, and finally α-solanine and α-chaconine was recently elucidated (Itkin et al. 2013;Cardenas et al. 2015; Sonawane et al. 2016). Although SGA biosynthesis in Solanaceae species is relatively well understood, the mechanisms regulating the process in various tissues require further study. Several quantitative trait loci (QTLs) controlling foliar SGAs content have been identified, but these QTLs vary among different segregating populations (Yencho et al. 1998; Sagredo et al. 2006; Manrique-Carpintero et al. 2014). A major QTL associated with SGAs content in tubers was mapped to chr01 using a diploid BC1population derived from S. tuberosum×S. sparsipilum (Sorensen et al. 2008). Coexpression analysis revealed that a cluster of APETALA2/Ethylene Response Factor (ERF) genes on chr01 harboring GLYCOALKALOID METABOLISM 9 (GAME9)controls SGAs content by upregulating the transcription of multiple biosynthetic genes in potato leaves and tuber skin (Cardenas et al. 2016). However, no QTLs specifically regulating SGAs content in tuber flesh have thus far been identified.

Controlling the SGAs content in tuber flesh to acceptable safety levels (＜200 mg kg-1flesh weight) is crucial for potato breeding, especially when wild potato species are used in breeding. Wild potato species have many desirable properties, such as insect resistance, disease resistance and high dry matter content (Jansky et al. 2014; Spooner et al. 2014). The introgression of wild species can expand the genetic variation of cultivated potatoes, but can also result in some genetic drags, such as high levels of SGAs in tubers. The potato cultivar Lenape, which was derived from crosses with the wild species S. chacoense Bitter,a species that is particularly resistant to Colorado potato beetle (Leptinotarsa decemlineata Say) but produces bitter tubers, had to be withdrawn from the market due to high levels of SGA (Sinden et al. 1984). This example illustrates the need to understand the genetic basis of SGAs content in tuber flesh.

In this study, we constructed an F2segregating population from a cross of diploid potatoes species, S. tuberosum×S. chacoense. In a three-year study, we identified two QTLs affecting the SGAs content in tuber flesh. Our findings provide insights into the regulation of SGAs content in tuber flesh, as well as the application of wild germplasm in potato breeding.

2. Materials and methods

2.1. Plant materials

This study used the diploid potato accessions S. tuberosum group Tuberosum RH89-039-16 (RH), S. tuberosum group Tuberosum E, S. chacoense 34-28, an F2population from the cross between E and 34-28, and the germplasm resource accessions 4 to 25. The E172 clone, an F1individual from a cross between E and 34-28, is characterized by bitter tubers, self-compatibility and strong growth vigor. F2seeds from the E172 clone were sown, and 40-day-old seedlings were transplanted into 36 cm×36 cm black nursery pots for tuberization. Approximately four months after transplanting,the tubers were synchronously harvested. Since many F2tubers were very small, they did not spout the following year.To ensure the accuracy of QTL-seq, three independent experiments using different F2seeds were performed from 2016 to 2018. The number of F2individuals examined in 2016, 2017 and 2018 was 431, 260 and 350, respectively.All the plant resources were provided by the Germplasm Resources Library of CAAS-YNNU Joint Academy of Potato Sciences, Yunnan Normal University, Kunming, China.

To obtain tuber samples, the peel (outer 1.5 mm tissue)and flesh were separately collected and graded according to tuber weight. Leaf samples from RH were harvested from the first to ninth expanded leaf arrangement, labeled L1 to L9. The stems, stolons and roots of RH were sampled from morphologically upper, middle and lower sections of these structures. RH flowers (carpel, stamen and perianth tissues)were also sampled. All samples were quickly collected and immediately frozen in liquid nitrogen.

2.2. Extraction and analysis of α-solanine and α-chaconine

Plant tissue extracts were prepared as described by Itkin et al.(2011). To analyze SGAs content, liquid chromatographyquadrupole time-of-flight mass spectrometry (LC-QTof-MS)was performed using a Waters Xevo G2-XS QTof system operated in selected ion monitoring. The LC-QTof-MS was equipped with a SunFire C18 Column (100?, 5 μm, 2.1 mm×150 mm, Waters, USA), which was operated at a temperature of 40°C. The mobile phases consisted of eluent A (acetonitrile containing 0.1% (v/v) formic acid) and eluent B (ultra-pure water containing 0.1% (v/v) formic acid) at a flow rate of 0.200 mL min-1, with 1.000 μL samples loaded per injection. The gradient profile started at 10% A for 1 min and increased linearly to 60% A for 10 min, and the column was then washed and equilibrated with 10% A for 5 min before the next injection. The levels of α-solanine and α-chaconine were calculated based on the ratio of peak area at m/z 868 and 852, respectively, from positive ion scans using a calibration curve of authentic samples (ChromaDex,USA), with both coefficients of determination: r2＞0.998.

2.3. Generation and analysis of QTL-seq data

Genomic DNA was isolated from the fresh leaves of plants from the F2population using the CTAB method. Two extreme pools were constructed for QTL-seq based on the distribution of SGAs content in the F2population: High-pool(top 5% of SGAs content) and Low-pool (bottom 5% of SGAs content). Each pool was sequenced on the Illumina HiSeq X Ten platform, generating 40 Gb of clean data. The sequences were aligned to the potato reference genome DM(v4.03) (PGSC 2011) with BWA Software, and SNP-calling was performed with SAMtools and BCFtools Software. In a previous study, two haplotypes of the F1clone E172 were phased using ～50 F2individuals (Zhang et al. 2019). Taking the haplotype B of E172 as the reference, the SNP-index was calculated using the sliding window method (window size=1 Mb, step=10 kb). The Δ(SNP index), representing the absolute value of differences in the SNP index between the two pools, was calculated to identify candidate QTLs(Abe et al. 2012; Takagi et al. 2013). QTL mapping was performed three times in three successive years.

2.4. KASP genotyping assay

The genotyping assay was carried out using the Kompetitive Allele Specific PCR (KASP) SNP genotyping platform(Semagn et al. 2013). SNPs exhibiting polymorphism between the parents in the candidate QTL regions were selected, converted to KASP markers and used to screen the parents and nine F2individuals to confirm the polymorphisms before genotyping the entire F2population. The extent of scoring errors was evaluated for two no-template controls(NTC) that had been included across several plates of samples submitted for genotyping. For genotype calling, the SNP haplotype detected by KASP markers was converted to the ‘a(chǎn), b and h’ genotypes. The SNP genotype identical to that of haplotype A of E172 was defined as a, and the other genotype was defined as b. The SNP markers used in this study are listed in Appendix A.

2.5. RNA-seq analysis

Fresh flesh was sampled from large tubers of E, 34-28 and E172 prior to natural plant death. Total RNA was extracted from the samples and subjected to RNA-seq on the Illumina HiSeq X Ten platform. The insert size of the libraries was 300 bp, and the read length was 150 bp. For each sample,4 Gb of data were generated and aligned to the potato reference genome version 4.03 (PGSC 2011). RNA-seq analysis was conducted using HISAT2 (v2.1) for alignment and StringTie (v1.3.3b) to calculate expression levels.Transcripts were annotated with InterProScan (v5.21).Genes showing an adjusted P-value＜0.05 identified by DESeq were considered to be differentially expressed. All candidate genes were analyzed using Heatmap (http://www.omicshare.com/tools) with a 5% false positive detection threshold.

3. Results

3.1. The distribution of SGA in various organs/tissues of potato

We measured the SGAs content (including α-solanine and α-chaconine) in various organs of the diploid potato clone RH. The SGAs content varied among organs and in the same organ of different developmental stages (Fig. 1-A).Flowers and young leaves contained high levels of SGAs (up to 15 312 and 6 952 mg kg-1fresh weight (FW), respectively);the major SGA in both organs was α-solanine. In other organs/tissues, the ratio of α-solanine to α-chaconine ranged from 0.89 to 2.14. The SGAs levels of tuber peel was similar to that of old leaves and stolons, whereas tuber flesh contained very few SGAs (Fig. 1-A). These results demonstrate that the SGA biosynthesis pathway in cultivated potato is intact, but its regulatory mechanisms vary among organs. We then compared the SGAs content in the tubers of 10 diploid wild species, ten diploid landraces and two tetraploid modern cultivars (Table 1). The SGAs content in tuber peel did not significantly differ between wild and cultivated accessions (t-test, P=0.26) (Fig. 1-B).However, the SGAs contents were high in the tuber flesh of wild potatoes, at levels far beyond the safety threshold of 200 mg kg-1FW (Fig. 1-B). By contrast, the tuber flesh of cultivated potatoes was safe for consumption. This difference indicates that the mechanism regulating SGA biosynthesis in the tuber flesh of cultivated potatoes likely changed during domestication.

3.2. Measurement of SGAs content in an F2 population

Fig. 1 Distribution of steroidal glycoalkaloids (SGAs; sum of α-solanine and α-chaconine) in potato. A, SGAs content in different organs or tissues of diploid clone RH. L1 to L9 represent samples from the first to ninth expanded leaves. B, SGAs content in different diploid wild species (accessions 4 to 13), diploid landraces (accessions 14 to 23) and tetraploid modern cultivars of potato(accessions 24 to 25). Bars are SD.

To identify the genetic loci controlling SGAs content in tuber flesh, we constructed an F2segregating population derived from a cross between S. tuberosum clone E and S. chacoense clone 34-28 (Fig. 2-A). Since the male parent, 34-28, harbors the S-locus inhibitor (Sli) gene, F1clone E172 was self-compatible and was used to generate the F2population. The SGAs contents in tuber flesh of E,34-28 and E172 were 16.07, 2 003.69 and 273.77 mg kg-1FW, respectively (Fig. 2-B). The SGAs contents in the F2population ranged from 22.09 to 1 487.29 mg kg-1FW(Fig. 2-C), suggesting that SGAs content in tuber flesh is a quantitative trait. Similar results were obtained in three independent experiments.

3.3. QTL mapping of SGAs content in tuber flesh

QTLs affecting SGAs content in tuber flesh were identified by bulked segregant analysis. Due to their quantitative inheritance and environmental effects, the genomic regions with high delta indices (the absolute value of difference in the SNP index between two pools) varied among years,but two loci on chr02 and chr08 were repeatedly detected in three experiments (Fig. 2-D and E and Appendix B).The QTL on chr08 showed the highest delta index in two of the three years (2016 and 2018), i.e., delta index≥0.45 in 2016 and ≥0.58 in 2018, both in regions between 3.5 and 6.3 Mb on chr08, as highlighted in Fig. 2-E and Appendix B.The highest delta index in 2017 was ≥0.50 in the genomic segments in the 27.0 to 32.3 Mb region of chr02, as highlighted in Appendix B. When the threshold of delta index was ≥0.45 in 2018, a peak on chr02 was detected at 28.0-30.7 Mb. Thus, although the highest peak slightly fluctuated during the three-year study, these two regions,i.e., 27.0-32.3 Mb on chr02 and 3.5-6.3 Mb on chr08, were selected as candidate QTLs responsible for SGAs content in tuber flesh; these QTLs were designated as qSTF2 and qSTF8, respectively.

We analyzed the correlation between the genotypes of these two QTLs and SGAs content. When the genomic region harboring qSTF8 was homozygous, there was no significant difference in the SGAs content between genotype‘a(chǎn)’ and ‘b’ of qSTF2 (Fig. 2-F). However, when the genomic region harboring qSTF2 was homozygous, genotype ‘b’ of qSTF8 had higher SGA levels than genotype ‘a(chǎn)’ (Fig. 2-G).These results indicate that, compared to qSTF2, qSTF8 might have a larger effect on SGAs content in tuber flesh.However, the influence of environmental and other factors on the effects of these QTLs is currently unclear.

Table 1 Potato (Solanum sp.) accessions used in this study

3.4. RNA-seq analysis of the genes in qSTF2 and qSTF8

To explore the candidate genes regulating SGAs content in qSTF2 (27.0-32.3 Mb) and qSTF8 (3.5-6.3 Mb), we sequenced the transcriptomes of tuber flesh from different clones, including E, 34-28 and E172. We investigated the expression patterns of known SGA biosynthesis genes(HMGS, HMGR, MVK, SQE, CAS, SQS, GAME7, GAME8,GAME11, GAME6, GAME4, GAME12, SGT1, SGT2 and SGT3) (Itkin et al. 2013; Sonawane et al. 2016) and the regulatory gene GAME9 (Cardenas et al. 2016). The expression levels of all GAME genes and sterol alkaloid glycosyltransferase genes, as well as GAME9, were positively correlated with SGAs content in tuber flesh(Fig. 3-A). However, none of the expression patterns of the cholesterol biosynthesis genes (HMGS, HMGR, CAS and SQS) were similar to those of the GAME genes (Fig. 3-A).We detected 354 gene transcripts at the qSTR2 locus,whereas no transcription factor genes were found among the 63 co-expressed genes (Appendix C). In the candidate 2.8 Mb region of qSTF8, 195 genes were predicted,142 of which were expressed. Heatmap analysis of the 142 expressed genes identified 38 co-expressed genes associated with SGAs content in tuber flesh (Fig. 3-B and Table 2). The co-expressed genes included seven transcription factor genes, such as the MYC transcription factor gene PGSC0003DMT400002963, which might regulate SGA biosynthesis in tuber flesh.

4. Discussion

In this study, we used LC-QTof-MS analysis to quantify the toxic SGAs content in various organs of the diploid potato clone RH, as well as the tuber flesh and peel of various wild species, landraces and modern tetraploid cultivars. We determined that the decrease in SGAs content in tuber flesh is a critical domestication trait. No QTL for SGAs content in tuber flesh has previously been mapped. One single ISSR marker on chr03 was previously found to be associated with SGAs content in tuber flesh in a segregating population of tetraploid S. tuberosum, but no QTLs were mapped in that study (Dam et al. 2003). In the current study, using a diploid F2segregating population, we identified two QTLs on chr02 and chr08 that accounted for SGAs content in tuber flesh.

Transcriptome analysis showed that the expression levels of 38 genes at the qSTF8 locus were positively correlated with SGA content and that seven transcription factor genes might be involved in the regulation of SGA biosynthesis and the domestication of specialized metabolites in tuber flesh.We were especially interested in the basic helix-loop-helix transcription factor MYC gene PGSC0003DMT400002963,which is highly homologous to NbbHLH2, CrMYC2 and SlMYC2. In Nicotiana benthamiana, NbbHLH2, together with NbbHLH1, positively regulate jasmonate-induced activation of the biosynthesis of the alkaloid nicotine (Todd et al. 2010). In Catharanthus roseus, the basic helix-loophelix transcription factor CrMYC2 controls the jasmonateresponsive expression of ORCA genes, encoding AP2/ERFdomain transcription factors that regulate the biosynthesis of the terpenoid indole alkaloid (Zhang et al. 2011).SlMYC2 regulates SGA biosynthesis in tobacco protoplasts(Cardenas et al. 2016). Moreover, a synergistic effect was observed when GAME9 was combined with SlMYC2, which activated the promoters of C5-SD, GAME4, GAME7 and HMGR1, pointing to the cooperative action of these two transcription factors in the regulation of SGA biosynthesis(Cardenas et al. 2016). Thus, whether this potato MYC transcription factor plays a role in regulating steroidal glycoalkaloid biosynthesis in tuber flesh is worth exploring.

Fig. 2 Genetic analysis of steroidal glycoalkaloids (SGAs) content in the tuber flesh of potato in an F2 population. A, F2 population derived from a cross between E and 34-28. The clone E172 is an F1 individual characterized by bitter tubers, self-compatibility and strong growth vigor. B, SGAs content (sum of α-solanine and α-chaconine) in the tuber flesh of parents (E and 34-28) and F1 plant (E172). C, distribution of SGAs content in the flesh of 431 F2 individuals. D, SNP-indices of the high-SGA and low-SGA pools on 12 chromosomes. E, The Δ(SNP index) of two bulks. The arrows indicate two QTLs identified on ch02 and ch08. F,SGAs content of different genotypes of qSTF2. G, SGAs content of different genotypes of qSTF8.

Fig. 3 Expression patterns of steroidal glycoalkaloid (SGA)-related genes in the tuber flesh of E, E172 and 34-28. A, expression levels of SGA biosynthesis genes. B, expression levels of genes at the qSTF8 locus. The co-expressed genes associated with SGAs content in tuber flesh were highlighted with a blue wireframe.

Table 2 Co-expressed transcription factor genes at the qSTF8 locus may be associated with steroidal glycoalkaloids (SGAs)content in tuber flesh

Our identification of the QTLs affecting SGAs content in tuber flesh should contribute to the utilization of wild resources for potato breeding. Scientists from the International Potato Center (CIP) have been actively researching how wild species can be used to help improve crop varieties (Gruber 2016). As our understanding of important domestication traits increases, the de novo domestication of wild species using genome editing should occur in crops such as potato (Li et al. 2018; Zsogon et al.2018). Our results will be helpful for exploiting the genes controlling SGAs content in tuber flesh, and they provide a theoretical basis for the utilization of wild germplasm in potato breeding.

5. Conclusion

In the present study, we quantified the SGAs content in wild species, landraces and modern cultivars of potato to explore whether changes SGAs content in tuber flesh occurred during domestication in potato. In a three-year study, we identified two candidate QTLs on chr02 and chr08 associated with domestication and/or the regulation of SGAs content in tuber flesh. The QTL on chr08 harbors 38 candidate genes that are co-expressed with GAME genes and associated with SGAs content in tuber flesh.

Acknowledgements

This work was supported by the Agricultural Science and Technology Innovation Program from the Chinese Academy of Agricultural Sciences (ASTIP-CAAS), the Agricultural Science and Technology Innovation Program Cooperation and Innovation Mission, CAAS (CAAS-XTCX2016), the Advanced Technology Talents in Yunnan Province, China(2013HA025), the PhD Scholar Newcomer Award in Yunnan Province, China (01701205020516025) and the Graduate Research and Innovation Fund Project, China (yjs201679).This work was also supported by the Ministry of Agriculture and Rural Affairs of China, the Shenzhen Municipal and Dapeng District Governments, China.

Appendicesassociated with this paper can be available on http://www.ChinaAgriSci.com/V2/En/appendix.htm