Capsaicin, the pungent principle in hot peppers, acts to deter mammals from consuming pungent pepper pods. Capsaicinoid biosynthesis is restricted to the genus Capsicum and results from the acylation of the aromatic compound, vanillylamine, with a branched-chain fatty acid. The presence of capsaicinoids is controlled by the Pun1 locus, which encodes a putative acyltransferase. In its homozygous recessive state, pun1/pun1, capsaicinoids are not produced by the pepper plant. HPLC analysis confirmed that capsaicinoids are only found in the interlocular septa of pungent pepper fruits. Immunolocalization studies showed that capsaicinoid biosynthesis is uniformly distributed across the epidermal cells of the interlocular septum. Capsaicinoids are secreted from glandular epidermal cells into subcuticular cavities that swell to form blisters along the epidermis. Blister development is positively associated with capsaicinoid accumulation and blisters are not present in non-pungent fruit. A genetic study was used to determine if the absence of blisters in non-pungent fruit acts independently of Pun1 to control pungency. Screening of non-pungent germplasm and genetic complementation tests identified a previously unknown recessive allele of Pun1, named pun12. Sequence analysis of pun12 revealed that a four base pair deletion results in a frameshift mutation and the predicted production of a truncated protein. Genetic analysis revealed that pun12 co-segregated exactly with the absence of blisters, non-pungency, and a reduced transcript accumulation of several genes involved in capsaicinoid biosynthesis. Collectively, these results establish that blister formation requires the Pun1 allele and that pun12 is a recessive allele from C. chinense that results in non-pungency.
Capsaicin and its analogues, collectively called capsaicinoids, are the pungent principle of peppers (Thresh, 1876). Peppers (Capsicum spp.) are well known for their ability to cause an intense organoleptic sensation of heat when consumed. Capsaicinoids are produced only within the genus Capsicum and their presence has driven the domestication of several Capsicum species (Walsh and Hoot, 2001). Peppers probably originated in Bolivia, as this area contains many of the 20–27 recognized species of Capsicum (Andrews, 1984; Hunziker, 2001; Walsh and Hoot, 2001). Bell pepper seeds were first traded nearly 500 years ago, and archeological evidence indicates peppers have a history of human use dating back nearly 9000 years, making Capsicum one of the oldest plant genera to be domesticated (Boswell, 1937; Pickersgill, 1966; Heiser Jr 1969; Basu and Krishna De, 2003).
Ecological studies suggest that capsaicinoids evolved as a means of directed deterrence (Tewksbury and Nabhan, 2001). Birds, the favoured seed-dispersal agents of wild Capsicum, do not perceive capsaicinoids. After ingestion by avian frugivores, seeds are transported within the birds’ gut, and then excreted, often in favourable habitats under trees that provide appropriate light levels for plant growth. Furthermore, passage through the avian digestive system appears to promote the germination of pepper seeds (Tewksbury and Nabhan, 2001). In contrast to birds, mammals are extremely sensitive to capsaicinoids, perceiving pain when capsaicinoids contact epidermal tissues. Furthermore, mammals possess crushing molars and harsh, acidic, digestive systems that harm pepper seeds (Tewksbury and Nabhan, 2001). The pain response to capsaicinoids is due to the activation of the TRPV1 (VR1) receptor in mammals by capsaicinoids; in birds, this receptor does not respond to capsaicinoids (Caterina et al., 1997, 2000; Jordt and Julius, 2002). In fact, the TRPV1 receptor is involved in several pain-sensing pathways, which has prompted much basic and clinical research on the use of capsaicinoids to treat a variety of human ailments, including arthritis, bladder and digestive problems, and cancer (Deal et al., 1991; Chancellor and De Groat, 1999; Han et al., 2001; Surh, 2002; Cruz, 2004; Srinivasan, 2005). Natural capsaicin analogues, which may function as neurotransmitters, have been found in mammalian brains, and endorphins have been shown to be released in the brain when capsaicin is consumed (Appendino et al., 2002; Huang et al., 2002; Chu et al., 2003).
It is generally accepted that capsaicinoids are produced solely in pepper fruits, although the location of the biosynthesis and accumulation of capsaicinoids within the fruits has been debated. A recent report describes the detection of capsaicinoids in vegetative organs, and others have reported small amounts of capsaicinoids in seeds (Ohta, 1962; Balbaa et al., 1968; Estrada et al., 2002). Within the pepper fruit, capsaicinoids chiefly accumulate along the epidermal cells of the interlocular septum which defines the fruit locules and is derived from the tissue connecting the placenta to the pericarp (Judd et al., 1999). The epidermal cells of the interlocular septum have been implicated in capsaicinoid biosynthesis, based on morphological changes during fruit development and the existence of osmiophilic granules in these cells (Furuya and Hashimoto, 1954; Ohta, 1962; Suzuki et al., 1980). In pungent varieties, epidermal protrusions or blisters arise from the lifting of the cuticle layer from the cell wall during the filling of subcuticular cavities with capsaicinoids (Rowland et al., 1983; Zamski et al., 1987).
Very little is known about the regulation of capsaicinoid biosynthesis. Capsaicinoids are produced by condensation of a branched-chain fatty acid, derived from either valine or leucine, with vanillylamine, derived from phenylalanine (Fig. 1a) (Bennett and Kirby, 1968; Leete and Louden, 1968; Suzuki et al., 1981; Sukrasno and Yeoman, 1993). In previous research, the identity of Pun1, a putative acyltransferase named AT3, was reported (Stewart et al., 2005). The Pun1 locus, first investigated nearly a century ago, is responsible for non-pungency throughout C. annuum based on inferred breeding pedigrees and sequencing results (Webber, 1911; Stewart et al., 2005). Non-pungent C. annuum genotypes were observed to have a 2.5 kb deletion spanning the putative promoter and first exon of AT3. The pun1 allele defined by this large deletion is the only known mutation to date that has a qualitative effect on the presence/absence of capsaicinoids (Webber, 1911; Blum et al., 2002).
Studies of this locus have revealed that AT3 transcript is expressed specifically in the placenta and begins to accumulate along with several other capsaicinoid biosynthetic genes at 20 days post-anthesis (dpa). It was noted that several capsaicinoid biosynthetic genes, including AT3, were either not expressed or reduced in expression when non-pungent and pungent genotypes were compared (Stewart et al., 2005). Similarly, Curry et al. (1999) and Aluru et al. (2003) observed that transcript accumulation of several capsaicinoid biosynthetic genes was correlated with the level of pungency.
In the book, The Paprika, Somos described the presence of blisters that contained capsaicin and suggested it may be possible to regulate capsaicinoid production by regulating the presence of blisters (Somos, 1984). A recent report termed these blister structures ‘vesicles’ and stated that the absence of these capsaicinoid-accumulating vesicles is inherited as a single recessive gene at a new locus designated ‘loss-of-vesicles’ (lov) (Votava and Bosland, 2002). This locus, lov, was proposed as a second locus, in addition to Pun1, that has a qualitative effect on pungency. The current study was undertaken critically to test the genetic relationship between blisters, capsaicinoid biosynthesis, lov, and Pun1.
The localization of capsaicinoid biosynthesis and accumulation to the epidermal cells of the interlocular septa in pungent fruits is reported here. The presence of blisters in pungent fruits represents sites of capsaicinoid secretion and storage and is precisely related to capsaicinoid accumulation. A recessive allele of the Pun1 locus was identified and it is shown that in the allelic state at Pun1, transcript accumulation of capsaicinoid biosynthetic genes and capsaicinoid accumulation are highly correlated. Based on these results, it is concluded that mutations at a single locus are responsible for non-pungency within C. annuum and C. chinense and that this locus is also responsible for the presence/absence of the blistered structures that contain capsaicinoids.
Materials and methods
Seeds of C. chinense ‘Habanero’ and C. annuum ‘Maor’ were obtained from commercial sources (Tomato Growers Supply Co., Ft. Myers, FL). C. chinense NMCA 30036, previously described by Votava and Bosland (2002), was obtained from the NMSU Chile Pepper Institute.
For genetic studies in the field, plants were sown in the greenhouse and then transplanted to fields located in Varna, NY during the summer of 2004. For gene expression studies, all plant materials were grown and maintained in greenhouse conditions at 27 °C day/night with supplemental lighting unless otherwise noted.
Scanning-electron and light microscopy
Interlocular septum tissue from fresh, unfixed, immature green Habanero and NMCA 30036 fruits were removed and dissected cross-sectionally by hand into 5 mm2 pieces using a razor blade. The samples were mounted on aluminium stubs and coated twice under vacuum with a film of fine-grain gold using a Denton Vacuum Sputter Coater. The surface of the sections was examined under a Hitachi S-3200N Variable Pressure Scanning Electron Microscope (SEM).
For light microscopic analysis, the interlocular septa of ripe Habanero fruits were fixed and embedded into low viscosity resin (Electron Microscopy Sciences, Hatfield, PA) as described in Rowland et al. (1983) with the following modifications. After fixation, samples were washed three times with cold 50 mM PIPES buffer, pH 5.6, dehydrated in a graded ethanol series, 30 min at room temperature for each step, and incubated overnight in 100% ethanol at 4 °C. Ethanol was removed through a graded acetone/ethanol series, 30 min at 4 °C for each step, and soaked in acetone for 2–3 h. Samples were infiltrated with resin as described in the manufacturer's protocol. Sections 1 μm thick were cut using a glass knife and stained using a 1% (w/v) crystal violet solution. An Olympus BX60 compound microscope attached to a digital camera and computer for video-capture was used to view samples.
Blisters on the epidermis of the interlocular septum of Habanero and NMCA 30036 were photographed at 10 d intervals through fruit development using a dissecting microscope attached to a digital camera mounted with the ScopeTronix MaxView Plus System (Scopetronix, Cape Coral, FL). Images were taken of multiple fruit from multiple plants.
Immunolocalization of capsaicinoids
Thick sections, approximately 8–10 mm wide, were dissected from the interlocular septa of immature green Habanero and NMCA 30036 fruits (approximately 30 dpa) using a double-edge razor blade. Samples were immediately placed in fixative consisting of 4% paraformaldehyde in 0.1 M sodium phosphate buffer, pH 7.2, and incubated overnight at 4 °C. Samples were washed five times with 0.1 M sodium phosphate buffer and dehydrated through a graded ethanol series, 30 min at room temperature for each step. Samples were incubated overnight at 4 °C in 100% ethanol, then moved through a graded Histoclear (National Diagnostics, Atlanta, GA) series, 1.5 h at each step, and held overnight in 100% Histoclear at ambient temperature. Samples were infiltrated with Paraplast (Kendall, Mansfield, MA) as per the manufacturer's instructions. Sections, approximately 8 μm thick, were cut, paraffin removed, and rehydrated in a graded ethanol series. A rabbit capsaicin antibody (Beacon Analytical Systems Inc., Portland, ME) was used for immunolocalization at a 1:250 dilution using the ImmPRESS Anti-Rabbit Ig Peroxidase Kit (Vector Laboratories, Burlingame, CA) following the manufacturer's instructions. According to the manufacturer, the capsaicin antibody has reactivity against capsaicin and dihydrocapsaicin, the dominant capsaicinoids in commercial peppers. Sections were viewed using an Olympus BX50 microscope attached to a digital camera.
HPLC and ELISA analysis of pungency
Extracts of Habanero, Maor, NMCA 30036, and F2 progeny from the cross of Habanero and NMCA 30036 were prepared and analysed for capsaicinoid accumulation via HPLC as previously described in Stewart et al. (2005).
Interlocular septa from ripe Habanero fruits were removed and either directly dried or gently brushed with a small spatula to remove blisters and then dried. Capsaicinoids were extracted and analysed via HPLC as previously described in Stewart et al. (2005).
The presence/absence of pungency in fruit produced for the genetic complementation tests was assessed with ELISA due to the greater sensitivity of this method for capsaicinoid detection. Multiple ripe fruits from Maor, NMCA 30036, and the F1 progeny of the complementation cross were analysed for capsaicinoid accumulation using the Capsaicin High Sensitivity Plate Kit (Beacon Analytical Sys. Inc, Portland, ME) that utilizes an antibody that specifically recognizes the capsaicinoid small molecule as the epitope. The manufacturer's protocol was followed with the followng modification. Methanol extracts of the dried pepper samples were not diluted prior to analysis.
Inheritance and co-segregation of blisters, pungency, and Pun1
Ripe fruit from F1 and F2 populations (n=495), derived from C. chinense Habanero and C. chinense NMCA 30036 and its reciprocal cross, were analysed for pungency and presence/absence of blisters, and then genotyped to determine which Pun1 allele(s) were present. Plants were scored twice independently as pungent or non-pungent by a panel of three trained individuals. Non-pungent phenotypes were confirmed using HPLC as described above. The presence/absence of blisters along the interlocular septum of the fruits was visually scored and the absence of blisters was independently verified by two observers. A PCR-based, co-dominant marker based on the following primers 5′-ATGTCAACCGGCCAGCAGCAT-3′, 5′-CTGATTCTTCTGCCACCTTCAATCCC-3′, 5′-CCCACAATCAAACTTACTTGAAC-3′ was used to determine the allelic state at Pun1 for all individuals. PCR was performed using Advantage 2 Polymerase (Clontech, Palo Alto, CA) on DNA isolated from the interlocular septa according to the method of Doyle and Doyle (1990) with one modification. Microfuge tubes were treated with polyvinylpolypyrrolidone (Sigma P6755), prior to DNA extraction to minimize interference of phenolic compounds during PCR. RNA and total protein from interlocular septa of mature green F2 plants, confirmed homozygous dominant or recessive at Pun1, were isolated as described below to analyse gene expression. Capsaicinoid accumulation in the same samples was analysed via HPLC as described above.
Genetic complementation tests to determine whether the absence of pungency in NMCA 30036 and Maor is due to defects at the same or different loci were conducted by crossing NMCA 30036×Maor. The interspecific F1 progeny were confirmed to be hybrids visually and by genotyping. Reciprocal crosses in the opposite direction did not pollinate. Ripe F1 fruits from each cross were assayed for capsaicinoid accumulation via ELISA as described above.
RNA gel blot analysis
RNA was extracted from pepper fruits at time points through fruit development for gel blot analysis as described in Stewart et al. (2005). Fruits from multiple plants were pooled together to minimize variation due to environmental effects. Due to small fruit size, whole fruits with seeds removed were used for all 10 dpa time points. Ten μg of RNA was transferred to Hybond N+ membrane (Amersham Bioscience, Piscataway, NJ). Full-length AT3 cDNA was radiolabelled using the Prime-It RmT Random Labeling Kit (Stratagene, La Jolla, CA) and hybridized to the membrane according to standard protocols (Sambrook and Russell, 2001). Pal and Kas, known capsaicinoid biosynthetic genes, were radiolabelled and used as probes as described previously (Curry et al., 1999; Aluru et al., 2003; Blum et al., 2003). Filters were washed twice at 65 °C in 2× SSC/0.1% SDS for 10 min and twice in 1× SSC/0.1% SDS for 10 min. Filters were exposed on a PhosphorImager storage screen (Molecular Dynamics). RNA gel bots were replicated at least twice from independent extractions.
Immunoblotting of AT3
Habanero, Maor, and NMCA 30036 fruits were harvested at 10 d intervals. Total protein was extracted and blotted as described in Stewart et al. (2005). For detection of the AT3 protein, a rabbit polyclonal antibody against AT3 and an anti-rabbit secondary antibody conjugated to phosphatase (Amersham Bioscience, Piscataway, NJ) were used at 1:2000 and 1:5000 dilutions, respectively. The production, use, and specificity of the AT3 antibody were described previously in Stewart et al. (2005). Coomassie staining of replicate gels from the same extractions or of the probed membrane was used to verify equal protein loading. Blots were developed using the ECL kit (Amersham Biosciences, Piscataway, NJ). Chemiluminescence emitted from the filter was detected using X-ray film. Immunoblots were replicated at least twice.
Cloning and sequencing
Primer sets based on Habanero genomic DNA were used to isolate homologous sequence from NMCA 30036 using PCR. The pGEM-T Easy Vector System I was used for all subcloning and sequencing (Promega, Madison, WI). Sequencing was performed by the Bioresource Center, Cornell University (www.brc.cornell.edu).
Localization of capsaicinoid accumulation to the epidermis of the Capsicum interlocular septum
A cross-section of a ripe C. chinense Habanero fruit with seeds removed, with the pericarp, placenta, and interlocular septum labelled, is shown in Fig. 1b. Capsaicinoids were extracted from Habanero leaf, root, pericarp, interlocular septum, and seed tissues. Capsaicinoid accumulation, monitored by HPLC, was present in the interlocular septum, 52 000±11 000 parts per million (ppm), and absent from all other tissues sampled (Fig. 1c). Residual capsaicinoids were detected in the pericarp, probably due to imperfect separation of the interlocular septum from the pericarp tissues. Close inspection revealed that when the fruit is cut open or handled roughly, capsaicinoids from the blisters (see below) can contaminate nearby seeds. These results are consistent with early radioisotope analysis localizing capsaicinoids to the interlocular septum of pepper fruits (Iwai et al., 1979).
Immunolocalization studies using a capsaicinoid antibody were conducted to determine the cellular location of capsaicinoid accumulation within the interlocular septum. Capsaicinoids were not detected in C. chinense NMCA 30036, a non-pungent variety, or negative controls (Fig. 2a–c). Figure 2d shows the immunolocalization of 3-keto-acyl ACP synthase (KAS), a capsaicinoid biosynthetic enzyme, to the epidermal cells consistent with previous reports (Aluru et al., 2003). Similarly, capsaicinoids were localized to the epidermal cells of the interlocular septum (Fig. 2e). Signal was also detected in the cell layer immediately subtending the epidermal cells, but it was not possible to determine whether this was actually due to the presence of capsaicinoids or an artefact of sample preparation. Capsaicinoids and KAS were detected in all the epidermal cells; their expression was not spatially restricted and did not occur in a noticeable pattern in the epidermal cells. Staining of the specimens with crystal violet, a general cytological stain, revealed cellular differentiation between cells under the blisters compared with cells not under the blisters (Fig. 2g). This suggests their involvement in secretion of capsaicinoids into the subcuticular cavity. The results of the organ-specificity of capsaicinoid biosynthesis and immunolocalization studies taken together establish that capsaicinoid accumulation and biosynthesis are localized to the epidermal cells of the interlocular septum in the pepper fruit.
Association of blisters and presence of capsaicinoids
Examination of the blistered structures observed on the interlocular septa of pepper fruit showed that they were only present in the pungent genotype, Habanero, and were entirely absent from the non-pungent genotype, NMCA 30036 (Fig. 3a). To examine the relationship between blisters and pungency more closely, the capsaicinoid accumulation along the interlocular septum was measured before and after brushing the tissue to remove blisters along the surface (Fig. 3b). Similar to studies where exudates are produced in glandular trichomes, it was not possible to remove all blisters from the interlocular septum by brushing (Fridman et al., 2005). Capsaicinoid accumulation in brushed interlocular septa was 20 600±3700 ppm compared with 52 000±6800 ppm in unbrushed samples, a reduction of 60% (Fig. 3).
The distribution of blisters was examined to understand the relationship between blister development and capsaicinoid accumulation. Blisters were not observed on interlocular septa of ripe non-pungent NMCA 30036 or 10 dpa Habanero fruits (Fig. 4a, b). Blisters were first observed at 20 dpa in Habanero fruit and persisted throughout fruit development (Fig. 4c–f). Generally, the size of the individual blisters increased markedly during fruit development although definitive measurements were hard to make. Consistent with this trend, capsaicinoids did not begin to accumulate in Habanero until 20 dpa, gradually increasing throughout fruit development (Fig. 4g). These results and the previous finding that capsaicinoid biosynthesis is not restricted to cells under the blisters suggest that the blisters act primarily as a site of capsaicinoid storage for materials synthesized in the epidermal cells.
Genetic analysis of pungency, blisters, and Pun1
An F2 population (n=495) derived from reciprocal crosses of Habanero×NMCA 30036 was generated to confirm an earlier report that the presence/absence of blisters along the interlocular septum is controlled by the single recessive gene, lov, at a second locus controlling pungency (Votava and Bosland, 2002). Habanero is blistered and pungent while NMCA 30036 is non-blistered and non-pungent. The resulting F1 plants were blistered and pungent, confirming the dominance of blisters and pungency. Results for the F2 population, shown in Table 1, indicate that absence of blisters precisely co-segregated with non-pungency. Blisters were never observed in non-pungent plants. In a testcross population, blisters co-segregated exactly with pungency consistent with a 1:1 ratio, confirming monogenic inheritance and perfect association of these traits in C. chinense (Table 1).
|P2 NMCA 30036||22||0||0||0||22|
|F2 [(P1×P2)×(P1×P2)]||214||169||0||0||45||3:1||1.80 (0.180)|
|F2 [(P2×P1)×(P2×P1)]||281||202||0||0||79||3:1||1.45 (0.228)|
|Pooled F2||495||371||0||0||124||3:1||0.0007 (0.979)|
|P2 NMCA 30036||22||0||0||0||22|
|F2 [(P1×P2)×(P1×P2)]||214||169||0||0||45||3:1||1.80 (0.180)|
|F2 [(P2×P1)×(P2×P1)]||281||202||0||0||79||3:1||1.45 (0.228)|
|Pooled F2||495||371||0||0||124||3:1||0.0007 (0.979)|
A F2 population (n=495) derived from reciprocal crosses Habanero (Pun1/Pun1, pungent, blistered)×NMCA 30036 (pun12/pun12, non-pungent, not blistered) was analysed for the inheritance and co-segregation of blisters, pungency, and Pun1. Bl=blistered; NB=not blistered.
The genomic sequence at the Pun1 locus was determined in NMCA 30036 to develop PCR-based markers for testing whether Pun1 segregated independently from the blister phenotype. During the development of these primers, the Pun1 locus from C. chinense was sequenced. A deletion of four base pairs (bp) located centrally in the first exon was identified in C. chinense NMCA 30036 at the Pun1 locus (Fig. 5). This deletion was confirmed by sequencing cDNA from 30 dpa NMCA 30036 fruit. Sequence analysis predicted that this 4 bp deletion causes a frameshift mutation which results in a truncated AT3 protein. The truncated AT3 is not predicted to include the putative active site, HXXXDG (St Pierre and De Luca, 2000).
Based on this deletion, a co-dominant marker system (Fig. 5) for Pun1 in C. chinense was developed. Three primers were used for genotyping, with primer 3 spanning the deletion found in NMCA 30036. When all the primers are mixed together during PCR, the reaction favours primer 3 as the reverse primer in the NMCA 30036 genotypes, generating a 0.5 kb amplicon (the shorter of two possible amplicons). Alternatively, in Habanero genotypes, primer 2 is the preferred reverse primer and generates a 1.1 kb amplicon. Primer 3 does not generate a significant amplification product in Habanero genotypes due to base pair mismatching between the primer and template. Genotyping of F2 individuals from the segregating population described above revealed that this 4 bp deletion, tentatively named pun12, co-segregated precisely with the absence of blisters and non-pungency in the F2 population shown in Table 1. These results suggest that non-pungency in NMCA 30036 previously attributed to a second locus, lov, is actually perfectly associated with a 4 bp deletion at the Pun1 locus. To confirm this result, genetic complementation analysis was performed to assess whether lov in C. chinense and Pun1 in C. annuum are allelic.
Complementation tests to determine allelic diversity of the Pun1 locus
To test the hypothesis that the 4 bp deletion identified in the Pun1 locus of NMCA 30036 is the gene previously designated lov, a genetic complementation test was performed by crossing C. chinense NMCA 30036×C. annuum Maor (non-pungent, pun1/pun1). Capsaicinoid accumulation was not detectable in either parent, NMCA 30036 or Maor (Table 2). Capsaicinoid accumulation in all F1 plants was below the lower detection limit of the ELISA assay system (0.50 ppm) indicating that the loss-of-pungency mutations in Maor and NMCA 30036 fail to complement. These data conclusively establish that the locus previously designated lov is actually a second recessive allele of the Pun1 locus, defined by a 4 bp deletion. According to conventions in Capsicum gene nomenclature, it is proposed that this allele be designated pun12, and listed as synonymous with lov. The F1 plants that were recovered from this cross did produce fruit, but did not produce viable seeds and were deformed and exhibited virus-like symptoms, which is often observed in the progeny of such crosses. The reciprocal cross could not be obtained.
|Genotype||n||Capsaicinoid accumulation (ppm)|
|Genotype||n||Capsaicinoid accumulation (ppm)|
Parents and F1 hybrids were assayed for capsaicinoid accumulation using an ELISA system.
nd, Not detectable.
Asterisk denotes capsaicinoid accumulation was below limits of assay detection.
Regulation of AT3, Pal, and Kas gene expression in peppers containing the alleles, Pun1, pun1, and pun12
Transcript accumulation of capsaicinoid biosynthetic genes is hypothesized to correlate with the level and presence of pungency (Curry et al., 1999; Aluru et al., 2003). The expression patterns of AT3, a putative acyltransferase encoded by the Pun1 locus, and phenylalanine ammonia lyase (Pal) and Kas, two structural genes implicated in capsaicinoid biosynthesis, were examined in Habanero (Pun1/Pun1), Maor (pun1/pun1), and NMCA 30036 (pun12/pun12). In Habanero, Pal and Kas expression was similar; both were strongly expressed by 20 dpa before decreasing gradually throughout later fruit development (Fig. 6a). Pal and Kas were specific to the interlocular septum, although a small amount of signal was detected in flowers and seeds. This background probably resulted from the fact that both Pal and Kas are members of gene families and are essential genes in phenylpropanoid and fatty acid metabolism, respectively (Blum et al., 2003).
AT3 expression was also specific to the interlocular septum and was first detected at 20 dpa. AT3 expression persisted throughout much of fruit development but was not detectable by 50 dpa. The expression patterns of Pal, Kas, and AT3 closely followed that of capsaicinoid accumulation throughout fruit development (Fig. 4g). Capsaicinoids were first detected at 20 dpa and increased through fruit development. At 50 dpa capsaicinoids still accumulated, even though expression of Pal, Kas, and AT3 was not detected. It is plausible that metabolic turnover, for example, a reduction in the degradation of capsaicin, could account for the continued capsaicinoid accumulation at 50 dpa (Bernal et al., 1993; Estrada et al., 2000).
In Maor, both Pal and Kas had significantly reduced transcript accumulation throughout much of fruit development relative to Habanero (Fig. 6a). Similar to Habanero, however, there was background expression of both Pal and Kas in leaf, flower, seed, and pericarp tissues. In addition, there was strong expression of Pal and Kas in the seeds, probably due to lignins and fatty acids known to accumulate in seeds (Buchanan et al., 2000). AT3 expression in Maor was not detectable during fruit development. This observation was consistent with the lack of capsaicinoid accumulation and was also consistent with previous results showing that the deletion in AT3 disrupts transcription and translation (Fig. 4g) (Stewart et al., 2005).
Pal and Kas had similar expression profiles in NMCA 30036 (Fig. 6a). Both Pal and Kas were strongly expressed early in fruit development, 10 dpa versus 20 dpa in Habanero. At 20 dpa, however, both Pal and Kas were less abundant in NMCA 30036 than in Habanero. Expression levels of Pal and Kas were higher in NMCA 30036 than in Maor, perhaps due to genotype- or species-specific differences. In NMCA 30036, AT3 transcripts were expressed at significantly lower levels when compared with Habanero from 20 dpa through 50 dpa (Fig. 6a).
AT3 protein accumulation in Habanero followed the same pattern as transcript accumulation, first detected at 20 dpa, and reduced by 50 dpa (Fig. 6b). AT3 protein was not detectable in Maor or NMCA 30036 genotypes, except in seeds. The absence of AT3 protein accumulation in these genotypes is consistent with their mutations, a 2.5 kb deletion in pun1 and a 4 bp deletion in pun12, which disrupt translation. A cross-hybridizing band was detected in the seed tissues of all genotypes tested. This band persisted despite variations in protein extraction procedures and was consistently found in pungent and non-pungent genotypes. There are probably several acyltransferases involved in the production and storage of fatty acids and oils operating in seed tissues (Buchanan et al., 2000). A similar cross-hybridizing band was reported in the isolation and characterization of DAT, an acyltransferase involved in vindoline biosynthesis (St Pierre et al., 1998).
Segregation of gene expression and Pun1 genotype
RNA gel blots were made from representative samples of mature green (approximately 40 dpa) fruit from the F2 homozygous progeny classes resulting from the cross of Habanero×NMCA 30036. Pal gene expression did not segregate between the different F2 classes (Fig. 7b). This is consistent with the gene expression studies shown in Fig. 6 which show that Pal is similarly expressed in both Habanero and NMCA 30036 at 40 dpa. Kas appeared to accumulate more in Pun1/Pun1 F2 progeny even though Kas expression at 40 dpa between the two parental genotypes is not significantly different. It is plausible that this difference in Kas gene expression is a result of environmental variation between the parental genotypes shown in (Fig. 6) and F2 progeny (Fig. 7b). AT3 was clearly more abundant in the Pun1/Pun1 progeny than in the pun12/pun12 progeny. Within the Pun1/Pun1 class there were not major differences in AT3 expression even though total capsaicinoid accumulation varied significantly (Fig. 7d). AT3 protein was not detectable in pun12/pun12 progeny, consistent with the genetic lesion that produces the truncated protein (Fig. 7c). Taken together, the results clearly establish a strong relationship between the allelic state of Pun1, pungency, and AT3 and Kas expression.
Due to centuries of human manipulation, taxonomists disagree on the criteria necessary to identify some of the species and varieties of Capsicum. In particular, C. annuum, C. frutescens, and C. chinense are referred to as the C. annuum complex due to the difficulty of demarcating the individual species. Members of this complex include pungent and non-pungent, wild and domesticated varieties which interbreed and are morphologically similar (Walsh and Hoot, 2001). In previous research, it was established that the pun1 allele is responsible for non-pungency within C. annuum as a result of a large deletion at Pun1 that has been conserved and propagated for several centuries (Stewart et al., 2005). In this report, another recessive allele of Pun1, namely pun12, is identified. DNA sequencing of pun12 revealed a 4 bp deletion in the centre of the first exon of AT3. Inheritance studies revealed that pun12 co-segregated with the absence of blisters, non-pungency, and decreased expression of two capsaicinoid biosynthetic genes, Kas and AT3. Genetic complementation tests confirmed that pun12 is allelic to pun1. We have tentatively identified another allele of pun1 in C. frutescens (GM Stellari and M Mazourek, unpublished data). This study establishes that, to date, only mutations in Pun1 define known genetic sources of non-pungency in domesticated Capsicum spp.
Results from this study support the long-held assertion that capsaicinoids are synthesized primarily or exclusively in the interlocular septa of pungent fruits. This tissue has independently been referred to as the septum (Ohta, 1962; Rowland et al., 1983), dissepiment (Furuya and Hashimoto, 1954; Balbaa et al., 1968), cross wall (Huffman et al., 1978), or placenta (Iwai et al., 1977a, b, 1979; Fujiwake et al., 1980; Suzuki et al., 1980), generating confusion in the literature. In small-fruited varieties, the distinction between the placenta and interlocular septum is difficult to discern and in other varieties this tissue breaks down during ripening (Rao and Paran, 2003). Huffman et al. (1978) attributed the small amounts of capsaicinoids detected in the seeds to surface contamination during dissection. There has been persistent debate in the literature regarding the existence of glands that accumulate capsaicinoids. At least two reports concluded that the existence of specialized glands containing capsaicinoids could not be verified (Erwin, 1932; Huffman et al., 1978). Conversely, other reports have documented the presence of capsaicin-secreting glands, sometimes called receptacles, vesicles, or blisters, along the interlocular septum of pungent genotypes (Furuya and Hashimoto, 1954; Ohta, 1962; Suzuki et al., 1980; Rowland et al., 1983; Zamski et al., 1987; Votava and Bosland, 2002). In botany, the term ‘gland’ is used broadly, often having the connotation of specialized secretory structures (e.g. glandular trichomes of tomato or mint). However, any cells secreting specialized metabolites can be referred to as glandular even if they lack glandular trichomes (Fahn, 1979). For example, the papillae cells of the stigma surface are called glandular based on their secretion of carbohydrates and lipids to aid pollen tube growth, yet they lack formal secretory structures (Buchanan et al., 2000).
The term blister has been used to describe the epidermal swellings produced by the filling of subcuticular cavities along the interlocular septum of pungent peppers (Rowland et al., 1983; Zamski et al., 1987). In an analogous system, the resin exudate of epidermal cells along young leaves and stipules in Populus spp. (Salicaceae) is secreted into the space between the outer wall of the cells and the cuticle, forming a bulge or blister (Langehheim, 2003). In this article, the term blister was used in reference to such epidermal swellings in pungent pepper fruits and the term glandular cells in broad reference to any cells that secrete capsaicin.
Blisters in Habanero begin forming at 20 dpa concomitant with capsaicinoid accumulation. This link between specialized morphological features, blisters, and their specialized metabolites, capsaicinoids, is consistent with studies in sage and thyme (Lamiaceae), showing that the development of the glandular trichomes is correlated with essential oil production (Croteau et al., 1981; Yamaura et al., 1989). Immunolocalization of capsaicinoids and Kas, a subunit of the fatty acid synthase complex involved in capsaicinoid biosynthesis, showed that the epidermal cells are the site of capsaicinoid biosynthesis. Capsaicinoid biosynthesis is not restricted to the cells immediately subtending the blisters but occurs along the entire epidermis. Based on differential staining observed during microscopic analysis, the cells immediately under the blister are glandular cells, or idioblasts, that secrete capsaicinoids. Capsaicinoids are secreted into spaces between the cell wall and the cuticle and, as the pressure builds, the cuticle is separated and bulges to form a blister with a subcuticular cavity (Furuya and Hashimoto, 1954; Ohta, 1962; Rowland et al., 1983; Zamski et al., 1987). Similar to the blisters on resin-secreting tissues in Populus spp., the pepper blisters are formed from the exudate secreted from multiple cells; however, it is not known whether capsaicinoids are transported intercellularly from nearby cells not immediately under a blister (Langehheim, 2003). These results are consistent with the conclusion of Rowland et al. (1983), who reported that capsaicinoids were only found in the glandular areas along the interlocular septum of Jalapeño peppers.
Several reports in the literature describe blisters in non-pungent Capsicum varieties (Rowland et al., 1983; Zamski et al., 1987). Such a phenomenon was not observed in this analysis. A possible source of these observations may be that intense yellow-pigmented areas occur in both pungent and non-pungent varieties; however, careful examination reveals no association with pungency. This is consistent with previous work that concluded that the presence of the yellow pigmentation does not necessarily indicate the presence of capsaicinoids (Huffman et al., 1978).
A previous report stated that the presence/absence of blisters constituted a second locus controlling pungency, lov, in the non-pungent C. chinense genotype NMCA 30036 (Votava and Bosland, 2002). Analysis of F2 and testcross populations indicated that the absence of blisters and non-pungency perfectly co-segregate in every population examined, consistent with the possibility that a single defect has pleiotropic effects on capsaicinoid biosynthesis and blisters. This implies that capsaicinoid biosynthesis is required for blister formation. Genetic complementation tests reported here definitively prove that non-pungency in NMCA 30036 is due to a recessive allele at the Pun1 locus herein designated pun12.
Our results also support the hypothesis that the transcriptional state of some genes in the capsaicinoid biosynthetic pathway is positively correlated with pungency (Curry et al., 1999; Aluru et al., 2003). The complete absence of AT3 protein accumulation in pun1 and pun12 genotypes is consistent with a lack of capsaicinoid accumulation. The pun1 allele is characterized by the absence of AT3 expression, owing to a 2.5 kb deletion of AT3, and residual levels of Pal and Kas expression (Stewart et al., 2005). In pun12 genotypes, a small 4 bp frameshift mutation results in the absence of AT3 protein accumulation. Also within pun12 both Pal and Kas are expressed at levels greater than pun1 but less than Pun1. A strong up-regulation of several capsaicinoid biosynthetic genes (pAMT, Pal, Kas, BCAT, FatA) occurs 20 dpa after flowering coinciding with capsaicinoid accumulation in pungent varieties of both C. annuum and C. chinense (Stewart et al., 2005). In addition, the transcript accumulation of AT3 and Kas at 20 dpa was higher in pungent genotypes than in non-pungent genotypes from both C. annuum and C. chinense. This pattern of expression was confirmed in an F2 population. This suggests that the association of elevated transcript levels of genes involved in capsaicinoid biosynthesis with capsaicinoid biosynthesis, itself, is consistent across multiple Capsicum species. While there were small differences in Pal expression between pungent and non-pungent varieties, its expression did not segregate in an F2 population. This indicates that coordinated transcriptional regulation of capsaicinoid biosynthesis occurs downstream of Pal with genes that are specific to this metabolic pathway. Current research efforts aim to characterize relevant gene expression in non-pungent Capsicum accessions drawn from diverse species and to analyse the genetic basis for variation in the expression of genes implicated in this biosynthetic pathway.
The identification of a new recessive allele of Pun1 and the development of co-dominant markers will be useful for marker-assisted selection in breeding programmes and studies of allelic diversity at loci relevant for agriculture. Growing evidence suggests that dramatic changes in phenotypes during plant domestication can result from perturbations in single dominant genes, such as Pun1. The gain and loss of strawberry flavour, the architecture of maize inflorescence, and the liberating of maize kernels from the hardened fruitcases of the maize progenitor teosinte are all traits controlled by single dominant genes (Aharoni et al., 2004; Gallavotti et al., 2004; Wang et al., 2005). Manipulation of such genes will be useful in the metabolic engineering of natural products for the benefit of humanity.
We would like to acknowledge Mary Kreitinger, George Moriarity, Matthew Falise, Arnon Ben-Chaim, Steve Czaplewski (Syngenta Seeds), Lynn Bohs, Soumitra Goshroy, and Dominick Paolillo for technical assistance and useful discussions. We would also like to thank Robert Jarret, USDA Capsicum Germplasm Collection, for assistance. We would like to acknowledge Beacon Analytical Systems, Inc. for donation of the capsaicinoid antibody. This work was supported in part by USDA IFAFS Plant Genome Award No. 2001-52100-11347, NSF Metabolic Biochemistry Award No. 0412056 and supplemental awards, NSF Plant Genome 0218166 and NSF MCB 0417056, to CS and MMJ and NIH GMS06-08136 to MOC.