Pepper (Capsicum annuum L.) is an economically important crop with added nutritional value. Production of capsaicin is an important quantitative trait with high environmental variance, so the development of markers regulating capsaicinoid accumulation is important for pepper breeding programs. In this study, we performed association mapping at the gene level to identify single nucleotide polymorphisms (SNPs) associated with capsaicin pathway metabolites in a diverse Capsicum annuum collection during two seasons. The genes Pun1, CCR, KAS and HCT were sequenced and matched with the whole-genome sequence draft of pepper to identify SNP locations and for further characterization. The identified SNPs for each gene underwent candidate gene association mapping. Association mapping results revealed Pun1 as a key regulator of major metabolites in the capsaicin pathway mainly affecting capsaicinoids and precursors for acyl moieties of capsaicinoids. Six different SNPs in the promoter sequence of Pun1 were found associated with capsaicin in plants from both seasons. Our results support that CCR is an important control point for the flux of p-coumaric acid to specific biosynthesis pathways. KAS was found to regulate the major precursors for acyl moieties of capsaicinoids and may play a key role in capsaicinoid production. Candidate gene association mapping of Pun1 suggested that the accumulation of capsaicinoids depends on the expression of Pun1, as revealed by the most important associated SNPs found in the promoter region of Pun1.
Citation: Reddy UK, Almeida A, Abburi VL, Alaparthi SB, Unselt D, Hankins G, et al. (2014) Identification of Gene-Specific Polymorphisms and Association with Capsaicin Pathway Metabolites in Capsicum annuum L. Collections. PLoS ONE 9(1): e86393. https://doi.org/10.1371/journal.pone.0086393
Editor: Mark Gijzen, Agriculture and Agri-Food Canada, Canada
Received: October 21, 2013; Accepted: December 6, 2013; Published: January 27, 2014
Copyright: © 2014 Reddy et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This project was supported by USDA-NIFA (Contract #2010-38821-21574). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Pepper (Capsicum annuum L.) is a crop of major agricultural and economic importance. It is known for its pungency, rich flavor, and nutritional value. World production of pepper in 2011 was estimated to be 29,939,029 metric tons; the United States alone recorded the production of 1,018,490 metric tons . Pepper contributes a range of beneficial metabolites, such as carotenoids, flavonoid glycosides and vitamins, to the human diet . The most unique metabolites are the alkaloids denominated by capsaicinoids, which make peppers pungent and are produced mainly in the placenta of the fruits . Capsaicinoids have been widely used in food and for pharmaceutical purposes –. The most important pharmaceutical role of capsaicin is in pain perception. The transient receptor potential of vanilloid type 1 receptor (TRPV1) is activated by capsaicin in mammalian nociceptor cells, triggering inflammation and pain responses , . Prolonged exposure to capsaicin numbs the TRPV1 over time, for long-term pain relief.
The use of molecular markers can save time and money in breeding programs by detecting particular traits before costly phenotyping is performed. Thus, genetic markers able to detect pungency and/or capsaicinoid profiles during the seedling stage are valuable tools in pepper breeding. Mazourek et al.  proposed a model integrating the capsaicin biosynthesis pathway and mapped genes. The acyl moieties of capsaicinoids are derived from catabolism of amino acids with subsequent fatty acid elongation , . In later studies, Aluru et al.  reported that the transcript level of the placental-specific β-ketoacyl carrier protein synthase I (KAS1) was positively associated with pungency. Abraham-Juarez et al.  silenced KAS by virus-induced gene silencing in Capsicum chinense and created plants with undetectable levels of mRNA and capsaicinoids, thus providing further evidence for the important role of this gene in altering pepper pungency. A crucial branching point in the capsaicin pathway is the metabolite p-coumaric acid, which is also important in synthesis of a wide variety of secondary metabolites such as lignins, flavonoids, hydroxycinnamic polyamides and pigments. Cinnamoyl CoA reductase (CCR) reduces coumaroyl, feruoyl and sinapoyl-CoA esters to their respective cinnamaldehydes; therefore, CCR is considered important in lignin biosynthesis and is a major control point of phenylpropanoid metabolic flux. It may have a role in determining capsaicinoid levels .
Capsaicinoids are alkaloids generated from the condensation of vanillylamine derived from the phenylpropanoid pathway and a variable branched chain fatty acid. A major dominant locus that alters capsaicin was mapped to chromosome 2 of pepper and named the C locus –. Kim et al.  identified SB2-66, a cDNA clone from a suppression subtractive hybridization library constructed from pungent C. chinense and further characterized to be homologous with acyl transferase. Interestingly, SB2-66 was found to express only in the placenta of pungent peppers. Stewart et al.  genotyped a mapping population with SB2-66 and noted that its relevant restriction fragment-length polymorphisms (RFLPs) co-segregated exactly with the pungency trait and mapped close to the C locus. Subsequently, Stewart et al.  sequenced a full-length transcript as well as genomic DNA, along with a 1.8-kb promoter, and named the locus Pun1. Pun1 encodes AT3, an acyl transferase from the BAHD acyl transferase superfamily. Allelic tests for Pun1 identified a 2.5-kb deletion unique to C. annuum. Later, the loss of pungency in C. chinense, Capsicum frutescens and Capsicum chacoense was found to be caused by species-specific independent events , . Hill et al.  genotyped 43 pepper accessions, 40 belonging to C. annuum, and discovered seven homologs of Pun1 and reported the presence of three acyl transferases. Nevertheless, Pun1 is the only known locus to have a qualitative effect on pungency in C. annuum complex. Han et al.  demonstrated that Pun1 functions in capsinoid synthesis. Yumnam et al.  reported 79 single nucleotide polymorphisms (SNPs) in Pun1 from sequences of 15 pepper accessions of landraces from India. To date, no association mapping has been performed to measure the effects of individual SNPs on the accumulation of capsaicinoids.
Capsaicin and dihydrocapsaicin are the major capsaicinoids, and they differ only in the saturation of their fatty acid chain. Capsaicin and dihydrocapsaicin make up approximately 90% (66% and 22%, respectively) of total capsaicinoids . Various Capsicum species and accessions within the species accumulate capsaicinoids in different proportions , . Iwai et al.  indicated that capsaicin does not interconvert to dihydrocapsaicin, and some capsaicinoids do not undergo changes during different growth stages, which suggests unique regulatory effects on the expression of various enzymes in the capsaicin metabolic pathway.
In this study, we aimed to sequence Pun1, CCR, KAS, and hydroxycinnamoyl transferase (HCT) genes and use the underlying polymorphisms for association mapping to identify markers responsible for variation in capsaicinoids and the other metabolites in the capsaicin pathway among a diverse C. annuum population.
Materials and Methods
We investigated 94 accessions of C. annuum from various countries representing a wide geographical area of the world (Table S1). These selfed accessions were grown in three replications during the summers of 2011 and 2012 (seasons 1 and 2). Genomic DNA isolation involved use of the DNeasy plant mini kit (QIAGEN cat# 69104).
Detailed metabolite profiling involved gas chromatography coupled with mass spectrometry (GC/MS) performed at the University of Illinois. For metabolic profiling, dried polar extracts were derivatized with 80 µl methoxyamine hydrochloride (20 mg ml-1) for 60 min at 50°C, 80 µl MSTFA for 120 min at 70°C, then 2-hr incubation at room temperature. An amount of 10 µL of the internal standard (hentriacontanoic acid, 10 mg/mL) was added to each sample before derivatization. Samples were analyzed on a GC/MS system (Agilent Inc, Palo Alto, CA, USA) consisting of an Agilent 7890 gas chromatograph, an Agilent 5975 mass selective detector, and a HP 7683B autosampler. Gas chromatography involved an HP-5MS capillary column (60 m×0.25 mm I.D. and 0.25-µm film thickness) (Agilent Inc, Palo Alto, CA, USA). The inlet and MS interface temperatures were 2500°C, and the ion source temperature was adjusted to 2300°C. An aliquot of 1 µL was injected with the split ratio of 10∶1. The helium carrier gas was kept at a constant flow rate of 1.5 ml min-1. The temperature program was 5-min isothermal heating at 700°C, followed by an oven temperature increase of 50°C min−1 to 3100°C and a final 10 min at 3100°C. The mass spectrometer was operated in positive electron impact mode (EI) at 69.9 eV ionization energy in m/z 30–800 scan range. The spectra of all chromatogram peaks were compared with those in electron impact mass-spectrum libraries NIST08 (NIST, MD, USA), W8N08 (Palisade Corp., NY, USA), and a custom-built library. To allow comparison between samples, all data were normalized to the internal standards in each chromatogram. The spectra for all chromatogram peaks were evaluated by use of the programs HP Chemstation (Agilent, Palo Alto, CA, USA) and AMDIS (NIST, Gaithersburg, MD, USA). Metabolome concentrations are reported as “(analyte concentration relative to hentriacontanoic acid) per gram Wet Weight” (relative concentration) (i.e., as target-compound peak area divided by the internal standard [IS] peak area [IS concentration is the same in all samples]): Ni = Xi × X-1hentriacontanoic acid × g wet weight−1. Hentriacontanoic acid (C31H62O2) is a fatty acid that is usually absent in any real sample we had dealt with. Calibration curves could not be built for all identified metabolites because some are not commercially available as pure standards. Relative concentration (RC) is an accepted way to compare the same metabolite between different samples but does not allow for comparisons between different metabolites within a sample because of different MSD responses to various compounds.
Capsaicinoids were extracted by diluting 100 mg dried powder with 2 mL pure acetonitrile after thorough mixing on a vortex. The mixture was incubated at 50°C for 1 hr followed by 1-hr sonication before centrifugation at 10,000 rpm for 15 min. The supernatant was filtered through a Phenomenex 0.2-µm PTFE membrane filter (Torrance, CA, USA) before analysis. Capsaicin and dihydrocapsaicin were quantified by use of a Waters high-performance liquid chromatography (HPLC) system equipped with 1525 binary HPLC pump, 2707 autosampler and 2998 Photodiode array detector (Waters Corp., Milford, MA, USA). Acetonitrile with 2% acetic acid was used as mobile phase at a flow rate of 0.6 ml/min. Separation of capsaicinoids involved an X-Bridge C18 column (4.6×150 mm; 5 µm) coupled with a guard column (Waters Corp.). Capsaicin and dihydrocapsaicin were detected at 280 nm. Injection volume was set to 10 µL. Retention times for capsaicin and dihydrocapsaicin were 9.3 and 9.7 min, respectively. Stock solutions of capsaicin and dihydrocapsaicin (Sigma–Aldrich) were prepared in acetonitrile for a linear standard curve from 12.5 to 500 ppm. Metabolite concentrations were normalized by log2 transformation before further analysis.
Primer design and amplification
Gene-specific primers were designed with sequences available in Genbank for HCT (Genbank: EU616565), CCR (Genbank: EU616555), KAS (Genbank: HQ229922) and Pun1 (Genbank: AY819029). Primer pairs were designed to amplify overlapping fragments of ∼500 to 1000 bp that covered full template sequences by use of Primer 3 software . Sequences and annealing temperatures of primers are in Table S2. PCR amplification was performed in a total reaction volume of 50 µl containing 40 ng of genomic DNA with 25 µl GoTaq colorless master mix (Promega, Madison, WI, USA), 10 pmol each of forward and reverse primers, and completed with nuclease free water. Thermocycling conditions were an initial denaturing step of 95°C for 5 min, followed by 45 cycles of 95°C for 30 sec, corresponding annealing temperature for 30 sec and 72°C for 1 min, with a final extension step of 72°C for 2 min. Amplification of fragments was confirmed by visualization in a 1% agarose gel pre-stained with ethidium bromide under UV light. The amplified products were purified by polyethylene glycol precipitation.
Sequencing involved the BigDye terminator cycle sequencing kit v.3.1 (cat# 4337455, Life Technologies) and an ABI 3130x/Genetic analyzer sequencer. Sequence fragments were aligned by use of the software Sequencher 4.9 (Gene Codes Corp., MI, USA). Exons and introns for each gene were determined by aligning available cDNA sequences of Pun1 (Genbank: GU300812), KAS (Genbank: AF085148) and CCR (Genbank: EU616555) to the obtained genomic sequence with the software Spidey . Chromosomal assignment and position on the physical map of candidate genes were deduced from the Whole Genome Sequence draft for hot pepper (CM334) (kindly provided by Drs. Park and Choi of Seoul National University). Phylogenetic trees were constructed for the four candidate genes. First, sequences for each gene were aligned in Sequencher 4.9 and the alignment was exported to MEGA 5.2  to construct neighbor-joining trees. The nucleotide diversity (π) and Tajima’s test for selection were calculated on the alignments by use of DNASP 5.0 . Consensus sequences for the promoter sequence of Pun1 and intron sequences of CCR and KAS1 were searched in the PLACE database for identification of known cis-regulatory elements .
Candidate gene association mapping
Linkage disequilibrium (LD) was estimated as the correlation between all pairs of SNPs in individual candidate genes by use of the SNP & Variation Suite (SVS) v7.7.6 (www.goldenhelix.com). Haplotype blocks were computed with the default settings for the Gabriel et al.  algorithm imbedded in SVS v7.7.6. Haplotype frequencies for each defined haplotype block in all three genes were calculated by the estimation maximization (EM) method , with a frequency threshold of 0.01. Only SNPs with a minimum minor allele frequency > 0.1 were considered for LD studies and candidate gene association mapping. To visualize LD throughout the gene, heat maps were produced on the basis of pair-wise correlation estimates of all marker pairs. The Q and K matrices were adapted from previously performed simple sequence repeat (SSR) analysis . Q matrix was adapted from K-5 cluster of SSR data obtained by use of Structure v2.2. The Mixed Linear Model (MLM) of TASSEL v3.0 was used for association mapping for Pun1, KAS1 and CCR. HCT did not undergo association mapping because of minimum minor allele frequency < 0.1 for SNPs discovered in this gene. The SNP P-values obtained were not subjected to sequential Bonferroni correction  or FDR . Considering the sample size and number of polymorphisms used in our study, corrections for population structure and kinship were sufficient for association tests.
Principal component analysis
Numeric principal component analysis (PCA) of the metabolic profiles involved use of SVS v7.7.6. Before PCA, concentrations of metabolites directly or indirectly involved in the capsaicinoid pathway for 93 pepper accessions were normalized by log2 transformation. Accessions were categorized by their recorded pungency level from HPLC analysis. Analysis of accessions grouped by pungency involved plotting values of the first two eigen vectors of PCA with use of SVS v7.7.6.
PCA with normalized concentration values for various metabolites (Table S3) obtained by GC/MS and HPLC revealed non-pungent peppers with trace amounts of capsaicin and those with low pungency and a few moderately pungent accessions remaining on the negative side of the Y-axis, with only moderate-, high- and very high-pungent accessions located on the positive side of the Y-axis (Fig. 1). Tepin produced the highest amount of capsaicin, followed by Prikkinu and Bird’s eye baby during season 1 (Table 1). In season 2, all peppers showed a significant decrease in capsaicin, which indicated a high degree of environmental variance. In season 2, Hot Ornamental Prairie Fire produced the most capsaicin, followed by Tepin and Bolivian rainbow.
Accessions are labeled by their pungency level (EV = eigenvalue).
Association and diversity studies of Pun1
All primer pairs belonging to the Pun1 locus were successfully amplified in high-, moderate- and low-pungent accessions but not non-pungent peppers. This finding was expected because of a large deletion in the Pun1 locus reported for non-pungent accessions. Because the fragments were purified for direct sequencing, the presence of homologous bands with similar size could not be resolved in 1% agarose gel nor sequenced, especially the amplicons of primer pairs Pun1_1 and Pun1_3 (Fig. 2). We obtained a fragment of 3197 bp for 43 genotypes, with the exception of a fragment that contained a 201-bp gap pertaining to the Pun1_3 fragment. Thus, only a 2996-bp portion of the gene was successfully sequenced from the available 3753-bp genomic sequence (Genbank: AY819029).
Alignment of exons from the cDNA sequence of Pun1 to the Capsicum genome draft showed that Pun1 is on the negative strand of chromosome 2 (Table 2). A total of 36 polymorphisms were identified in Pun1: 19 were localized in the promoter, seven in the first exon, seven in the intron and three in the second exon: 20 were transversions, 15 were transitions and one was an indel of two nucleotides (Table S4). SNP positions are numbered by orientation (down or upstream) and position from the transcription start site. Annotations of Cis regulatory elements for various SNP positions are presented in Table S4. SNPs –483 and –482 have the binding sequence of SEBFCONSSTPR10A, which is a silencing factor of resistance gene PR-10 in potato. In addition, SNP –116 has the sequence for MYB1LEPR which regulates defense related gene expression in tomato. The pattern of LD distribution in Pun1 is presented in Fig. 3. On association mapping by the MLM approach, polymorphisms in Pun1 were found associated with variation in six main metabolites in the capsaicin pathway (capsaicin, dihydrocapsaicin, isoleucine, leucine, pyruvate and valine). In addition, three more metabolites that are produced as result of deviation of the capsaicin pathway (naringenin, spermidine and vanillic acid) were found associated with the polymorphisms located in Pun1. The associated SNPs with respective p-values and correlation values are in Table S5. The six reported SNPs showed an association in both seasons. These six SNPs had the highest correlation values and allelic effects and are presented in Table 3. The SNPs causing synonymous and non-synonymous mutations are in Table 4: SNPs 653 and 654 were associated with capsaicinoids in season 1 and acyl moiety precursors; they cause non-synonymous mutations at 33 and 34 amino acids away from the active site, respectively.
Four haplotype blocks were defined in Pun1 (Fig. 3). Markers contained in each block and haplotype frequencies calculated with the EM method are in Table S6. Block 3 was the largest and contained 13 markers with six distinguished haplotypes. The estimated probabilities of only two haplotypes totaled 0.85 of the total probabilities for this block, whereas the remaining ∼0.15 is represented by four less-probable haplotypes. Block 4 was the smallest haplotype block, comprising the last three markers of the second exon in Pun1. For the remainder of the haplotype blocks, the same trend was observed, with most of the haplotype estimated probabilities (>0.80) represented by two haplotypes. The SNPs associated with capsaicin in both seasons were located in blocks 1 and 2. Block 1 contained the associated SNPs –1392, –1390, –1386, –1120 and –1077. Seven different haplotypes were estimated for this block.
The overall nucleotide diversity (π) for the Pun1 locus including the promoter sequence was 0.0041 and that of the transcribed sequence 0.00387. When using a sliding window of 100 bp under a step size of 25 bp, the region near the active site (541–555 bp) had a nucleotide diversity of π = 0, then after 575 bp π increased rapidly and peaked (π = 0.01699) in the region located near the end of the first exon at base 650 (Fig. 4). Subsequently, nucleotide diversity dropped near the splicing region of exon1 and increased again in the intron sequence (π = 0.0139), with a gradual drop to zero up to base 1432 located in exon 2. Tajima’s D for Pun1 was calculated by genomic sequence alignment and only transcribed sequence alignment. Coding sequence alignment returned a D value of –0.665 considering 34 segregating sites. Meanwhile, genomic sequence alignment resulted in a lower D value of –1.027 calculated by data from 59 segregating sites, thus showing more evidence of purifying selection in Pun1. The discrepancy between the values can be explained by the number of segregating sites used, nevertheless, both analyses show that Pun1 is under purifying selection as is common for domesticated traits.
Phylogenetic analysis was performed for the genomic and transcribed sequence alignments. The neighbor-joining algorithm separated all the accessions into two main clusters based on the polymorphisms located on the coding sequences of Pun1 (Fig. S1). One clade was composed of only 9 accessions that included highly pungent Tepin. The second clade had two sister clades: one contained 28 accessions and the other the remaining eight accessions.
Association and diversity studies of CCR
CCR homologs are common in several plant families including Capsicum. The first primer pair of CCR (CCR_1) amplified two homologs, of ∼1.1 and 1.2 kb (data not shown). Primer pair CCR_2 yielded a single 1292-bp band and could be amplified across 53 pepper accessions. Alignment to the Capsicum genome draft positioned CCR on the (+) strand of chromosome 3 (Table 2). The full length of CCR in the genome is of 2764 bp; the alignment shows that CCR has 5 exons and 4 introns. The 1292-bp sequence extends from the beginning of the fourth exon at position 1419 to 2711 bp, toward the end of the gene. Sequence analysis showed that the NWYCY active site of CCR is well conserved in all accessions and is located in exon 4. Additionally, we report for the first time the presence of an intron between exons 4 and 5. A total of 32 polymorphisms were found in the CCR genomic fragment (Table S4). In all, 26 polymorphisms were located in the fourth intron, three in exon 4 and the remaining three in exon 5. Fifteen SNPs were transitions, and 13 were transversions. Additionally, we found two single-nucleotide insertions, another insertion with three nucleotides and a deletion of five nucleotides. Association mapping with MLM revealed CCR associated with caffeic acid and p-coumaric acid during season 1 (Table S5). A total of 14 polymorphisms were found associated with caffeic acid and also showed significant association with pyruvate, vanillate and p-coumaric acid. Haplotype analysis reported one block in CCR (Fig. 5). The block contained 28 markers, from the first polymorphism at 1460 bp to the SNP at 2426 bp. The first haplotype is represented by the major alleles and was estimated to have a probability of 0.52, while the rare alleles represented the second most frequent haplotype with an estimated probability of 0.30 (Table S6). Construction of a neighbor-joining tree allowed us to distinguish two main clades resolved by the polymorphisms located in CCR (Fig. S1). The largest clade contained 32 genotypes and the second contained the remaining 21 accessions.
The overall nucleotide diversity (π) for CCR was 0.0011. With use of a sliding window of 100 bp under a step size of 25 bp, the highest nucleotide substitution was at about bp 1888 (π = 0.0443) located in intron 4, and the value was ∼2 times higher than the highest value observed for Pun1. The nucleotide diversity decreased to 0.0248 from bases 1938 to 2039, where the conserved motif for splicing factor is located (Fig. 6). Subsequently, nucleotide diversity dropped close to 0 near the splicing region of exon 5. Testing for selection revealed that CCR was under positive selection, with Tajima D = 0.91, calculated with 47 segregating sites from 53 genotypes.
Association and diversity studies of KAS
The sequence used for the study of KAS gene was the genomic isolate (Genbank: HQ229922), which was derived from the cDNA sequence of C. chinense (Genbank: AF085148). A BLASTX search of this sequence revealed KASI and KASII domains, and a nucleotide BLAST search aligned the sequence to the KAS1 gene of tomato; hereafter, we refer to the gene studied as KAS1. Of eight primer pairs designed for KAS1, three were sequenced. We obtained a sequence of 1313 bases starting at position 149 of the KAS1 gene and ended at base 1,461 from 62 genotypes by using the overlapping primer pairs for KAS1_1, KAS1_2 and KAS1_3. Alignment of the available cDNA sequence to the genomic sequence in Spidey revealed eight exons for this gene. The sequence obtained was extended from the last seven bases for the first exon to 232 bases for the second exon, while passing through an intron. No polymorphisms were detected in the coding regions, but six SNPs were identified in the intron. Association mapping with MLM revealed linkage of SNP 447 with isoleucine, leucine, pyruvate and valine, the major precursors of the fatty acid moieties in capsaicin (Table S5). The constructed neighbor-joining tree showed that Nepalese (H137) pepper has a distinct KAS1 haplotype that separates it from the rest (Fig. S1). Nucleotide diversity for KAS1 was calculated to be 0.0026 considering 29 segregating sites. Testing for neutrality indicated that KAS1 is under negative selection, with Tajima D = –1.84.
Association and diversity studies of HCT
For HCT, we amplified 778 bp in exon 2 using the primer pairs HCT_2 and HCT_3. The alignment did not reveal any SNPs with frequency > 0.1, so we did not perform association mapping. In fact, nucleotide diversity for HCT was 0.0003 and was calculated from seven segregating sites. The Tajima’s D was –2.044, indicating negative selection for the HCT locus.
Our association-mapping results revealed Pun1 associated with six main metabolites in the capsaicin pathway (capsaicin, dihydrocapsaicin, isoleucine, leucine, pyruvate and valine) as well as three other metabolites produced from deviations of the capsaicin pathway (naringenin, spermidine and vanillic acid). Three SNPs, –483, –482 and 1559, controlled variation in major precursors for the acyl moieties pyruvate, valine, leucine and isoleucine, which are used in the synthesis of all known capsaicinoids. These metabolites are precursors of the fatty acid moieties that are used in the synthesis of capsaicinoids . SNPs causing non-synonymous substitution of amino acids in the coding region affected only the levels of capsaicinoids and valine, leucine and pyruvate in season 1. Pun1 greatly influenced the concentration of acyl moiety precursors, possibly because of the demand for production of the fatty acid moieties for capsaicinoid synthesis. Pyruvate is needed both for the synthesis of acetyl-CoA used in the fatty acid elongation pathway and as a precursor for the synthesis of valine, which is converted to iso-butyril and is elongated to the acyl moieties in two major capsaicinoids: capsaicin and dihydrocapsaicin . Wahyuni et al.  studied metabolic profiles of Capsicum spp and found that variation in volatile compounds corresponded well to differences in pungency. In addition, our association mapping showed that Pun1 is important in determining concentrations of narignenin, spermidine and vanillic acid, compounds resulting from deviation of the capsaicin pathway. Naringenin is a metabolite in the flavonoid pathway; 4-coumaroyl-CoA is derived from vanillin production and is converted to chalcone with subsequent isomerization to naringenin . In fact, three SNPs (75, 653 and 714), associated with naringenin, were found in exon 1 of AT3 and represented non-synonymous amino acid substitutions. Markers 653 and 714 were among the top three SNPs with the highest correlation values (r2 = 0.12) for naringenin. Three metabolites in the phenylpropanoid branch of the capsaicin pathway (coumaroyl-CoA, caffeoyl-CoA and feruoyl-CoA) are transferred to spermidine for the synthesis of hydroxycinnamic acid amides . However, vanillic acid results from vanillin oxidation . Our results showed that Pun1 is a key regulator of the major metabolites in the capsaicin pathway.
For CCR in Capsicum, we could sequence only a fragment of the 1292 bp because of multiple bands. Other studies have described multiple CCR homologs for Arabidopsis and Populus, and in Oryza up to 26 CCR and CCR-like genes have been reported . Similar to these studies, the first 400 bp of the CCR cDNA sequence aligned with another region of the pepper chromosome away from where the functional copy of CCR is located. This finding indicates the presence of the whole CCR gene family in Capsicum. Previous work on CCR has involved cDNA. In contrast, we used genomic DNA because intronic and genomic areas reveal more detailed information than the exons . In this study, we reported data for the sequence of the fourth intron of the CCR gene in C. annuum. Additionally, our sequence analysis of CCR revealed that the conserved catalytic motif NWYCY of CCR  is located after two bases from the beginning of the fourth exon in C. annuum. As expected, CCR showed a major association with p-coumaric acid and caffeic acid. CCR is known to act on coumaroyl, caffeoyl and feruoyl-CoA, converting them to their respective aldehydes . CCR activity is considered the first committed step in lignin biosynthesis , and our data support that the flux of coumarate and caffeate is highly controlled by CCR. Surprisingly, pyruvate and malonate were highly associated with CCR as well. Malonyl-CoA is used for fatty acid elongation and is synthesized from acetyl-CoA, which can be produced from pyruvate . CCR appears to have an indirect influence in the fatty acid branch of the capsaicin pathway by determining the flow of p-coumaric acid used for capsaicin synthesis. However, three malonyl-CoA molecules are needed to synthesize chalcone from p-coumaric acid for flavonoid biosynthesis . The association of CCR with malonate and pyruvate could be explained by CCR being the principal regulator of coumaroyl-CoA flux. In brief, CCR is a major determinant of p-coumaric acid, caffeic acid, pyruvate and vanillic acid concentrations in Capsicum fruits but also controls, in a minor manner, other metabolites in the capsaicin pathway.
Sequencing of KAS1 in the current study was hampered by the presence of similar band-sized homologs. In accordance, Mazourek et al.  mapped the KAS1 gene to seven different chromosomal locations in an integrated AFLP and RFLP map. Nevertheless, achieving direct sequencing of the fragments in the first intron indicates that this intron is highly conserved in sequence as well as size for all KAS1 homologs of C. annuum. Our 2012 study revealed association of capsaicin and KAS. In a study by Aluru et al. , KAS expression was positively correlated with pungency, and silencing of the KAS gene led to lower levels of capsaicinoids as well . In our study, all major precursors of capsaicinoid acyl moieties were found to be associated to KAS1. KAS genes are known to greatly affect the fatty acid composition of plants. For example, overexpression of KASIII in tobacco, Arabidopsis and rapeseed increased levels of 16:0 fatty acids . Leonard et al.  report that the introduction of a Cuphea wrightii KAS gene homologous to KASII transformed in Arabidopsis shifted fatty acid profiles towards short 8:0 and 10:0 chains. In addition, glutamine and γ-amino butyrate were among the metabolites associated with KAS1. Catabolism of amino acids for producing branched acyl moieties in capsaicinoids requires several transfers of amino groups by branched-chain-amino-acid aminotransferase (BCAT) . Although glutamate is considered the amino donor/acceptor in these steps, glutamine or γ-amino butyrate could also participate in the BCAT amino transfer reactions. Furthermore, γ-amino butyrate is a product of glutamate degradation. The low nucleotide diversity reported for HCT and the negative selection reflected by a –2.044 Tajima D value indicated that this gene is a locus of major importance for the phenylpropanoid pathway and plant development in general.
Our results show Pun1 as a regulator of major compounds in the capsaicin pathway, mainly capsaicinoids and also precursors for acyl moieties of capsaicinoids in C. annuum. Six different SNPs lying in the promoter sequence of Pun1 were found associated with capsaicin in plants from two different growing seasons by the candidate gene association-mapping approach. The results of candidate gene association mapping of Pun1 indicated that even though Pun1 is the only known qualitative trait for pungency, accumulation of capsaicinoids depends more on different genomic regions regulating the expression of the enzymes in the pathway. Indeed, the most important SNPs were found in the promoter region of Pun1. We report the presence of an intron sequence for CCR in C. annuum, and an SNP in a conserved intron motif involved in pre-mRNA splicing affects concentrations of caffeic acid and p-coumaric acid. Our results also support CCR as an important control point for the flux of p-coumaric acid to specific biosynthesis pathways. Consistent with previous reports, we found that KAS regulates the major precursors of acyl moieties of capsaicinoids and may play a key role in capsaicinoid production. Functional characterization of these SNPs will provide further details into their effects on capsaicinoid metabolism, thus elucidating the mechanism of capsaicinoid level control.
Neighbor-joining trees constructed with (A) transcribed sequence alignment of Pun1; (B) sequence alignment of Pun1 including promoter; (C) sequence alignment of CCR; and (D) sequence alignment of KAS.
Primer pairs for candidate genes used in the study.
Log2 values for metabolite concentrations.
Details of polymorphisms in Pun1 and annotations.
Association mapping of Pun1, CCR and KAS.
We thank Dr. Lucas (Zhong) Li, Director Metabolomics Center, and University of Illinois at Urbana-Champaign, for performing metabolome analysis. We thank Dr. Doil Choi, Seoul National University, for kindly providing genomic locations of various genes before publication. Authors thank Laura Smales for critical review.
Conceived and designed the experiments: UKR AA GH PN. Performed the experiments: AA DU VLA SBA. Analyzed the data: UKR AA DC MP PN. Contributed reagents/materials/analysis tools: UKR PN DC MP. Wrote the paper: UKR AA.
- 1. FAOStat (2011) Available: http://faostat.fao.org/site/567/DestopDefault.aspx?PageID=567#ancor.FAO.
- 2. Wahyuni Y, Ballester A-R, Sudarmonowati E, Bino RJ, Bovy AG (2011) Metabolite biodiversity in pepper (Capsicum) fruits of thirty-two diverse accessions: Variation in health-related compounds and implications for breeding. Phytochemistry 72: 1358–1370.
- 3. Fujiwake H, Suzuki T, Iwai K (1982) Capsaicinoid Formation in the Protoplast from the Placenta of Capsicum Fruits. Agricultural and Biological Chemistry 46: 2591–2592.
- 4. Minguez-Mosquera MI, Hornero-Mendez D (1993) Separation and quantification of the carotenoid pigments in red peppers (Capsicum annuum L.), paprika, and oleoresin by reversed-phase HPLC. Journal of Agricultural and Food Chemistry 41: 1616–1620.
- 5. Lee RJ, Yolton RL, Yolton DP, Schnider C, Janin ML (1996) Personal defense sprays : effects and management of exposure. Journal of the American Optometric Association 67: 548–560.
- 6. McCormack P (2010) Capsaicin Dermal Patch. Drugs 70: 1831–1842.
- 7. Caterina MJ, Schumacher MA, Tominaga M, Rosen TA, Levine JD, et al. (1997) The capsaicin receptor: a heat-activated ion channel in the pain pathway. Nature 389: 816–824.
- 8. Marzo VD (2004) Signalling at vanilloid TRPV1 channels. European Journal of Biochemistry 271: 1813–1813.
- 9. Mazourek M, Pujar A, Borovsky Y, Paran I, Mueller L, et al. (2009) A dynamic interface for capsaicinoid symstems biology. Plant Physiology 150: 1806–1821.
- 10. Leete E, Louden MCL (1968) Biosynthesis of capsaicin and dihydrocapsaicin in Capsicum frutescens. Journal of the American Chemical Society 90: 6837–6841.
- 11. Kopp B, Jurenitsch J (1981) Biosynthese der Capsaicinoide in Capsicum annuum L. var. annuum. Planta Med 43: 272–279.
- 12. Aluru MR, Mazourek M, Landry LG, Curry J, Jahn M, et al. (2003) Differential expression of fatty acid synthase genes, Acl, Fat and Kas, in Capsicum fruit. Journal of Experimental Botany 54: 1655–1664.
- 13. Abraham-Juárez MR, Rocha-Granados MC, López MG, Rivera-Bustamante RF, Ochoa-Alejo N (2008) Virus-induced silencing of Comt, pAmt and Kas genes results in a reduction of capsaicinoid accumulation in chili pepper fruits. Planta 227: 681–695.
- 14. Dixon RA, Paiva NL (1995) Stress-induced phenylpropanoid metabolism. The Plant Cell 7: 1085–1097.
- 15. Lacombe E, Hawkins S, Van Doorsselaere J, Piquemal J, Goffner D, et al. (1997) Cinnamoyl CoA reductase, the first committed enzyme of the lignin branch biosynthetic pathway: cloning, expression and phylogenetic relationships. The Plant Journal 11: 429–441.
- 16. Lefebvre V, Palloix A, Caranta C, Pochard E (1995) Construction of an intraspecific integrated linkage map of pepper using molecular markers and doubled-haploid progenies. Genome 38: 112–121.
- 17. Blum E, Liu K, Mazourek M, Yoo EY, Jahn M, et al. (2002) Molecular mapping of the C locus for presence of pungency in Capsicum. Genome 45: 702–705.
- 18. Blum E, Mazourek M, O'Connell M, Curry J, Thorup T, et al. (2003) Molecular mapping of capsaicinoid biosynthesis genes and quantitative trait loci analysis for capsaicinoid content in Capsicum. Theoretical and Applied Genetics 108: 79–86.
- 19. Kim M, Kim S, Kim S, Kim B-D (2001) Isolation of cDNA clones differentially accumulated in the placenta of pungent pepper by suppression subtractive hybridization. Molecules and Cells 11: 213–219.
- 20. Stewart C, Kang B-C, Liu K, Mazourek M, Moore SL, et al. (2005) The Pun1 gene for pungency in pepper encodes a putative acyltransferase. The Plant Journal 42: 675–688.
- 21. Stewart C, Mazourek M, Stellari GM, O’Connell MA, Jahn M (2007) Genetic control of pungency in C. chinense via the Pun1 locus. Journal of Experimental Botany 58: 979–991.
- 22. Stellari GM, Mazourek M, Jahn MM (2010) Contrasting modes for loss of pungency between cultivated and wild species of Capsicum. Heredity 104: 460–471.
- 23. Hill TA, Ashrafi H, Reyes-Chin-Wo S, Yao J, Stoffel K, et al. (2013) Characterization of Capsicum annuum Genetic Diversity and Population Structure Based on Parallel Polymorphism Discovery with a 30K Unigene Pepper GeneChip. PLoS ONE 8: e56200.
- 24. Han K, Jeong H-J, Sung J, Keum Y, Cho M-C, et al. (2013) Biosynthesis of capsinoid is controlled by the Pun1 locus in pepper. Molecular Breeding 31: 537–548.
- 25. Yumnam J, Tyagi W, Pandey A, Meetei NT, Rai M (2012) Evaluation of Genetic Diversity of Chilli Landraces from North Eastern India Based on Morphology, SSR Markers and the Pun1 Locus. Plant Molecular Biology Reporter 30: 1470–1479.
- 26. Bennett DJ, Kirby GW (1968) Constitution and biosynthesis of capsaicin. Journal of the Chemical Society C: Organic: 442–446.
- 27. Collins MD, Wasmund LM, Bosland PW (1995) Improved method for quantifying capsaicinoids in Capsicum using high-performance liquid chromatography. HortScience 30: 137–139.
- 28. Zewdie Y, Bosland PW (2001) Capsaicinoid profiles are not good chemotaxonomic indicators for Capsicum species. Biochemical Systematics and Ecology 29: 161–169.
- 29. Iwai K, Suzuki T, Fujiwake H (1979) Formation and accumulation of pungent principle of hot pepper fruits, capsaicin and its analogues, in Capsicum annuun var. annuun cv. Karayatsubusa at different growth stages after flowering. Agricultural and Biological Chemistry 43: 2493–2498.
- 30. Rozen S, Skaletsky HJ (1998) Primer3. Available: http://biotoolsumassmededu/bioapps/primer3_wwwcgi.
- 31. Wheelan SJ, Church DM, Ostell JM (2001) Spidey: a tool for mRNA-to-genomic alignments. Genome Research 11: 1952–1957.
- 32. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, et al. (2011) MEGA5: Molecular Evolutionary Genetics Analysis using Maximum Likelihood, Evolutionary Distance, and Maximum Parsimony Methods. Molecular Biology and Evolution 28: 2731–2739.
- 33. Librado P, Rozas J (2009) DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25: 1451–1452.
- 34. Higo K, Ugawa Y, Iwamoto M, Korenaga T (1999) Plant cis-acting regulatory DNA elements (PLACE) database: 1999. Nucleic Acids Research 27: 297–300.
- 35. Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, et al. (2002) The Structure of Haplotype Blocks in the Human Genome. Science 296: 2225–2229.
- 36. Fallin D, Schork NJ (2000) Accuracy of Haplotype Frequency Estimation for Biallelic Loci, via the Expectation-Maximization Algorithm for Unphased Diploid Genotype Data. The American Journal of Human Genetics 67: 947–959.
- 37. Abburi L (2013) Linkage disequilibrium and population structure analysis among Capsicum annuum L. cultivars for use in association mapping. WV, USA: West Virginia State University.
- 38. Holm S (1979) A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics 6: 65–70.
- 39. Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of royal statistical society 57: 289–300.
- 40. Winkel-Shirley B (2001) Flavonoid Biosynthesis. A Colorful Model for Genetics, Biochemistry, Cell Biology, and Biotechnology. Plant Physiology 126: 485–493.
- 41. Handrick V, Vogt T, Frolov A (2010) Profiling of hydroxycinnamic acid amides in Arabidopsis thaliana pollen by tandem mass spectrometry. Analytical and Bioanalytical Chemistry 398: 2789–2801.
- 42. Rao SR, Ravishankar GA (2000) Biotransformation of protocatechuic aldehyde and caffeic acid to vanillin and capsaicin in freely suspended and immobilized cell cultures of Capsicum frutescens. Journal of Biotechnology 76: 137–146.
- 43. Kawasaki T, Koita H, Nakatsubo T, Hasegawa K, Wakabayashi K, et al. (2006) Cinnamoyl-CoA reductase, a key enzyme in lignin biosynthesis, is an effector of small GTPase Rac in defense signaling in rice. Proceedings of the National Academy of Sciences of the United States of America 103: 230–235.
- 44. Wu FN, Mueller LA, Crouzillas D, Petiard V, Tanksley SD (2006) Combining bioinformatics and phylogenetics to identify large sets of single-copy orthologous genes (COSII) for comparative, evolutionary and systematic studies: a test case in the euasterid plant clade. Genetics 174: 1407–1420.
- 45. Pichon M, Courbou I, Beckert M, Boudet A-M, Grima-Pettenati J (1998) Cloning and characterization of two maize cDNAs encoding Cinnamoyl-CoA Reductase (CCR) and differential expression of the corresponding genes. Plant Molecular Biology 38: 671–676.
- 46. Funk C, Brodelius PE (1990) Phenylpropanoid Metabolism in Suspension Cultures of Vanilla planifolia Andr. II. Effects of Precursor Feeding and Metabolic Inhibitors. Plant Physiology 94: 95–101.
- 47. Dehesh K, Tai H, Edwards P, Byrne J, Jaworski JG (2001) Overexpression of 3-Ketoacyl-Acyl-Carrier Protein Synthase IIIs in Plants Reduces the Rate of Lipid Synthesis. Plant Physiology 125: 1103–1114.
- 48. Leonard JM, Knapp SJ, Slabaugh MB (1998) A Cupheaβ-ketoacyl-ACP synthase shifts the synthesis of fatty acids towards shorter chains in Arabidopsis seeds expressing Cuphea FatB thioesterases. The Plant Journal 13: 621–628.