Introduction
The genus Capsicum includes 30 species, five of which (C. annuum L., C. chinense Jacq., C. frutescens L., C. baccatum L. and C. pubescens Ruiz & Pavon) have been domesticated and are among the most popular foods consumed and cultivated worldwide (Antonio et al., Reference Antonio, Wiedemann and Veiga Junior2018; Florentino-Ramos et al., Reference Florentino-Ramos, Villa-Ruano, Hidalgo-Martínez, Ramírez-Meraz, Méndez-Aguilar, Velásquez-Valle, Zepeda-Vallejo, Pérez-Hernández and Becerra-Martínez2019). More than 200 cultivars and landraces of these five species contribute to the genetic diversity of the genus with a global production of 4,255,050 tons in 2019 (FAO, 2021). In Mexico, C. annum is the most important economically, and the Yucatan Peninsula is an important region of its diversification (Morales-Soriano et al., Reference Morales-Soriano, Kebede, Ugás, Grauwet, Van Loey and Hendrickx2018; del López Castilla et al., Reference del López Castilla, Hernández, Castillo Aguilar, Martínez-Hernández, Ortiz-García and Andueza-Noh2019). Morphological (Alam et al., Reference Alam, Syazwanie, Mahmod, Badaluddin, Mustafa, Alias and Prodhan2018), chemical and genetic diversity of Capsicum species have been studied worldwide, and habanero pepper (C. chinense Jacq.) has been the most studied in the Yucatan Peninsula (Cisneros-Pineda et al., Reference Cisneros-Pineda, Torres-Tapia, Gutiérrez-Pacheco, Contreras-Martín, González-Estrada and Peraza-Sánchez2007; Pino et al., Reference Pino, González, Ceballos, Centurión-Yah, Trujillo-Aguirre, Latournerie-Moreno and Sauri-Duch2007; Moreira et al., Reference Moreira, Ruas, de Ruas, Baba, Giordani, Arruda, Rodrigues, Simões and Gonçalves2018; Fabela-Morón et al., Reference Fabela-Morón, Cuevas-Bernardino, Ayora-Talavera and Pacheco2019; Florentino-Ramos et al., Reference Florentino-Ramos, Villa-Ruano, Hidalgo-Martínez, Ramírez-Meraz, Méndez-Aguilar, Velásquez-Valle, Zepeda-Vallejo, Pérez-Hernández and Becerra-Martínez2019; Gu et al., Reference Gu, Cao, Zhang, Zhang, Zhao, Zhang, Wang, Li and Wang2019; Kethom et al., Reference Kethom, Tongyoo and Mongkolporn2019). Nevertheless, the phenotypic diversity of the numerous chilli pepper landraces preserved by Maya farmers in the western region of the Yucatan Peninsula has been scarcely studied. In the traditional Maya crop system known as milpa, three of the five domesticated species of Capsicum (C. annum, C. chinense and C. frutescens), which can be distinguished by flavour, pungency and morphological characteristics, are grown. These species represent an important source of valuable traits for crop improvement programmes and are cultivated by traditional Maya farmers, who whether by choice or by need, conserve and develop local plant genetic resources by preserving landraces and associated local knowledge. However, until now, the characterization of genetic diversity and conservation status of these species in the western Yucatán Península has not been studied. Therefore, the chilli pepper landraces preserved by Maya farmers requires an understanding of their conservation status and genetic variation. Morphological characterization of landraces is a first step to guide efforts to conserve biodiversity and provide farmers with quality seeds of improved cultivars (Ghalmi et al., Reference Ghalmi, Malice, Jacquemin, Ounane, Mekliche and Baudoin2010; Carvalho et al., Reference Carvalho, Ragassi, Bianchetti, Reifschneider, Buso and Faleiro2014).
Pungency, perhaps the best known trait of peppers, is determined by the content of capsaicinoids, formed by the union of a vanillylamine group with a fatty acid. These alkaloids are found exclusively within the Capsicum genus and are synthesized in the placenta. The total capsaicinoid content is affected most by genotype, cultivar, ripening stage and growing conditions (Ananthan et al., Reference Ananthan, Subhash and Longvah2018; Antonio et al., Reference Antonio, Wiedemann and Veiga Junior2018; Naves et al., Reference Naves, de Ávila Silva, Sulpice, Araújo, Nunes-Nesi, Peres and Zsögön2019). Capsaicin and dihydrocapsaicin are the most abundant capsaicinoids, responsible for up to 90% of the pungency in peppers (Canto-Flick et al., Reference Canto-Flick, Balam-Uc, Bello-Bello, Lecona-Guzmán, Solís-Marroquín, Avilés-Viñas, Gómez-Uc, López-Puc, Santana-Buzzy and Iglesias-Andreu2008; Fayos et al., Reference Fayos, Savirón, Orduna, Barbero, Mallor and Garcés-Claver2019; Naves et al., Reference Naves, de Ávila Silva, Sulpice, Araújo, Nunes-Nesi, Peres and Zsögön2019). Cisneros-Pineda et al. (Reference Cisneros-Pineda, Torres-Tapia, Gutiérrez-Pacheco, Contreras-Martín, González-Estrada and Peraza-Sánchez2007) reported capsaicinoid content from 1010 to 71,789 μg/g of dry mass for seven cultivars of chilli peppers from Yucatan State, indicating that variation in the content is due to genetic factors or environmental conditions. Therefore, the capsaicinoid content is a key component to characterize the chemical diversity of peppers.
The aim of this work was to evaluate the phenotypic diversity and capsaicinoid content of nine chilli pepper landraces from the western region of the Yucatan Peninsula, to generate useful information that can be used in participatory breeding programmes to generate new varieties of peppers with better agronomics traits and to develop in situ and ex situ conservation programmes.
Material and methods
Plant material collection and cultivation
Nine chilli pepper landraces collected from the milpa of four Maya communities in Campeche state (Calkiní, Campeche, Champotón and Palizada) were obtained from a germplasm collection held at the Colegio de Posgraduados Campus Campeche, Mexico and representing three species of Capsicum (C. annuum with seven landraces known locally as: Bobo, Bolita, Dulce, Verde, landrace-Xcat'ik1, landrace-Xcat'ik2 and Maax; C. frutescens with one landrace named Pico Paloma and C. chinense with one landrace known locally as: Rosita). These nine chilli pepper landraces are the most important pepper varieties for the Maya people in the Yucatan Peninsula (del López Castilla et al., Reference del López Castilla, Hernández, Castillo Aguilar, Martínez-Hernández, Ortiz-García and Andueza-Noh2019). Twenty plants from each chilli pepper landrace were grown in a completely randomized blocks design with three replicates in a greenhouse at the experimental field of Colegio de Postgraduados Campus Campeche. Rows were 1.5 m apart, and plants were spaced at 0.25 m. The experimental unit consisted of 13 plants, of which 10 were evaluated. During the experiment, the plants were subjected to typical agronomic management (Soria-Fregoso et al., Reference Soria-Fregoso, Trejo-Rivero, Tun-Suárez and Terán-Saldivar2002).
Phenotypic characterization
For each of the 10 plants for each landrace, 54 traits including 15 quantitative and 39 qualitative extracted from the chilli pepper descriptors of IPGRI (1995) were measured. These included characteristics of the seedling, plant, flower, fruit and seed (online Supplementary Fig. S1 and Table S1).
Total capsaicinoid extraction and quantification
Capsaicinoids were extracted using the method of Collins et al. (Reference Collins, Wasmund and Bosland1995). A representative sample of total fruit production obtained randomly from each fresh pepper landrace was crushed and mixed with HPLC-grade methanol in a 1:10 (w/v) ratio, then heated in a water bath (Buchi B-491, Switzerland) at 80°C for 4 h. Experiments were done in triplicate. Samples were removed from the bath, cooled to room temperature, then the supernatant of each sample was filtered through 0.45 μm filter paper into an amber HPLC sample vial using a 5 ml disposable syringe (Millipore, Bedford, MA, USA). The vial was stored at 4°C until further analysis. Chromatographic analyses were done in a UPLC Acquity H Class chromatograph (Waters, Milford, MA, USA) equipped with a sample manager (UPSMFTN), a quaternary pump (UPQSM) and a photodiode array detector (UPPDALTC). Data were acquired using the Empower 3 software program (Waters). Capsaicinoids were separated in a Gemini C18 column (5 μm, 110A, 150 × 4.6 mm i.d.) (Phenomenex, Torrance, CA, USA) using the ASTA 21.3 method with slight modifications: 0.5 ml/min isocratic system flow and 10 μl injection volume. Capsaicin and dihydrocapsaicin were quantified using calibration curves for the respective analytical standards (5–300 μg/ml). All measurements were done in triplicate. Results were expressed in mg/kg dry mass for independent compounds and in Scoville units for total content using a factor of 16 as reported for the ASTA methodology. Moisture content (%) to express capsaicinoid content was determined according to the AOAC (2005) official method 934.06.
Statistical analyses
Qualitative and quantitative data were analysed with a multivariate analysis to identify traits that differentiate the landraces of chilli pepper. For morphological relationships, a principal component analysis (PCA) was conducted with the software SPSS version 20 (IBM, Armonk, NY, USA) and a cluster analysis with the UPGMA method using Euclidean distance in the PAST program (Hammer et al., Reference Hammer, Harper and Ryan2001). Mean capsaicinoid levels among the landraces were analysed for differences with a one-way analysis of variance (ANOVA), followed by Tukey's HSD test using Statgraphics Centurion 18 (Statgraphics Technologies, The Plains, VA, USA).
Results
Quantitative data analyses
The first three components in the PCA explained 86.77% of the total variation. The first principal component (PC1) explained 58.90% of the total variation and was influenced by cotyledon length, plant height, plant width, stem length, anther length, filament length, fruit length, fruit mass, fruit wall thickness and seed diameter. PC2 explained 17.15% of the total variation and was influenced mainly by mature leaf length, mature leaf width and fruit pedicel length. PC3 explained 10.71% of the variation and was influenced by cotyledon width and fruit width (Table 1).
Principal component analysis (PCA) of quantitative morphological data of nine landraces of Capsicum spp., cultivated in the Yucatan Peninsula
Values in bold indicate the variables with the major contribution for each principal component.
The PCA scatter diagram on the first two principal components allowed the identification of two well-defined groups (named Group I and Group II) with 76.05% of the total accumulative variance (Fig. 1), indicating that the graphical representation of the first two main components is appropriate to visualize the relationships between the landraces of chilli pepper studied. Group I was the most diverse (Fig. 1) and included Rosita, Bobo, Dulce, Xcat'ik1, Xcat'ik2 and Verde landraces. Plants in Group I have an intermediate size with medium to large fruits. Group II was formed by the Maax, Bolita and Pico Paloma landraces. Tall plants with very small fruits characterized landraces in Group II.
Principal component analysis of quantitative morphological data for nine landraces of Capsicum spp. from the Yucatan Peninsula.
Cluster analysis constructed with the UPGMA method (Fig. 2) supported the observations in the PCA scatter diagram (Fig. 1) by identifying two well-defined groups (Group I and Group II; Fig. 2), with bootstrap values of 86 and 90%, respectively. Each group included the same landraces as the groups in the PCA.
UPGMA dendrogram based on quantitative morphological data for nine landraces of Capsicum spp. cultivated in the Yucatan Peninsula.
Qualitative data analyses
The PCA results for qualitative data explained 77.35% of the total variation in the first four components. PC1 explained 33.51% of total variation, influenced most by hypocotyl pubescence, leaf blade margin, corolla spot colour, filament colour, pedicel attachment to fruit, seed colour, seed size and number of seeds per fruit. The variability in PC2 (18.34%) was influenced by stem shape, plant growth habit, number of flowers per axil and fruit cross-sectional corrugation. PC3 contributed 13.73% and PC4 11.76% of the total variability, with mature fruit colour and anther colour, respectively, contributing the most to the variation (online Supplementary Table S2).
Figure 3 shows a scatter diagram with three groups based on the first two PCs, which accounted for 51.85% of the total variation: Group I included five landraces belonging to C. annuum species (Dulce, Bobo, Xcat'ik1, Xcat'ik2 and Verde), which are characterized by white flowers, white filaments and larger seeds; Group II included the Rosita landrace (C. chinense) with a cylindrical stem, prostrate growth habit, three flowers per axil, yellow-green corolla, light purple filaments, bell-shaped fruits and orange ripe fruit; Group III comprised two landraces of C. annuum (Maax and Bolita) and one landrace of C. frutescens (Pico Paloma). These landraces share deciduous attachment of the fruit to the pedicel, less than 15 seeds per fruit and small seeds. The UPGMA conglomerate analysis supported the results obtained in the PCA with three main groups (Fig. 4).
Principal component analysis of qualitative morphological data for 39 traits for nine landraces of Capsicum spp., from the Yucatan Peninsula.
UPGMA dendrogram of qualitative morphological data for nine landraces of Capsicum spp. cultivated in the Yucatan Peninsula.
Moisture and capsaicinoid content
Moisture contents ranged from 60.42 to 90.41%, with the lowest value in Verde landrace and the highest in Xcat'ik1, statistically similar to Rosita (89.52%), Dulce (89.4%), Xcat'ik2 (88.52%) and Bobo (87.72%) (Table 2). Capsaicinoid content differed significantly (P < 0.05) among the landraces, with Rosita (Capsicum chinense) having the most capsaicin (9948.0 μg/g D.W), dihydrocapsaicin (4114.3 μg/g D.W) and total capsaicinoid content (14,062.3 μg/g D.W) (Table 2). The landraces Bolita, Maax and Pico Paloma had intermediate levels of capsaicin (4335.0, 2644.1 and 2201.9 μg/g D.W, respectively), dihydrocapsaicin (1593.1, 794.3 and 937.0 μg/g D.W, respectively) and total capsaicinoids (5928.1, 3438.4 and 3138.9 μg/g D.W, respectively; Table 2). The landraces Verde, Xcat'ik1, Bobo and Xcat'ik2 had the lowest levels of capsaicin (829.1, 524.5, 336.6 and 198.8 μg/g D.W, respectively), dihydrocapsaicin (548.7, 242.9, 370.4 and 238.1 μg/g D.W, respectively) and total capsaicinoids (1377.8, 767.4, 707.0 and 436.9 μg/g D.W, respectively). The Dulce landrace did not have any capsaicinoids and therefore lacked pungency.
Mean (±SD) of moisture, capsaicin, dihydrocapsaicin and total capsaicinoid content of nine landraces cultivated in the Yucatan Peninsula
Different letters in the same column indicate significant difference among landraces (Tukey's test, α = 0.05); DW, dry mass; SHU, Scoville heat unit.
Discussion
Our study is the first comprehensive evaluation of the phenotypic diversity and capsaicinoid content of the nine most economically important chilli pepper landraces from the western Yucatan Peninsula. The diversity in morphological characteristics and pungency among the chilli pepper landraces analysed contribute greatly to the important genetic diversity in the chilli germplasm cultivated in the western Yucatan Peninsula.
Phenotypic diversity using quantitative data
The UPGMA and PCA analyses (Figs. 1 and 2) show the separation of two main groups. In Group I, all landraces belong to the two most commonly cultivated species in the Yucatan Peninsula, and Group II includes landraces with wild characteristics (i.e. those with differences in characteristics related to the domestication, but able to survive in nature without human intervention). In this study, the different shapes of each genotype were key to observing the diversity among the nine chilli pepper landraces. The results show that high diversity is related to leaf, flower, fruit and plant characteristics, similar to the results of De la Cruz Lázaro et al. (Reference De la Cruz Lázaro, Márquez Quiroz, Osorio Osorio, Preciado Rangel and Márquez Hernández2017), who reported that plant and leaf traits (leaf length, leaf width, plant height, plant width and stem diameter) provide the greatest variation in C. frutescens. Similarly, Narez-Jiménez et al. (Reference Narez-Jiménez, de-la-Cruz-Lázaro, Gómez-Vázquez, Cruz-Hernández, Márquez-Quiroz and Castañón-Nájera2014) reported the greatest variability in fruit and leaf traits among different species of Capsicum. The separation of the chilli pepper landraces into two different groups in our study was similar to the results of Carvalho et al. (Reference Carvalho, Ragassi, Bianchetti, Reifschneider, Buso and Faleiro2014), who evaluated morphological and genetic relationships among wild and domesticated forms of C. frutescens and C. chinense and found that the wild forms grouped separately from the domesticated. In the UPGMA to determine the closest relationships among the landraces, the Rosita landrace was the most morphologically distant in Group I because this landrace belongs to C. chinense and the other landraces in Group I belong to C. annuum. This result agrees with del López Castilla et al. (Reference del López Castilla, Hernández, Castillo Aguilar, Martínez-Hernández, Ortiz-García and Andueza-Noh2019) who evaluated the structure, diversity and genetic relationships of nine accessions of Capsicum spp., and observed that the Rosita landrace was separated forming an individual group. Another interesting aspect to highlight is that the Bobo landrace was related closely to landraces Xcat'ik2 and Dulce, because Bobo resulted from a spontaneous cross between Xcat'ik and Dulce pepper (González et al., Reference González, Casanova, Gutiérrez, Torres, Contreras, Peraza, Durán and Méndez2010). Such crosses are very common when two landraces of C. annuum are sown in the same plot. This result is similar to the results of del López Castilla et al. (Reference del López Castilla, Hernández, Castillo Aguilar, Martínez-Hernández, Ortiz-García and Andueza-Noh2019) and González et al. (Reference González, Casanova, Gutiérrez, Torres, Contreras, Peraza, Durán and Méndez2010). Within Group II, Maax was the most distantly related to the other landraces, and Bolita and Pico Paloma landraces were the most closely related (90%). This result can be attributed to the genetic origin of these landraces: the Maax landrace belongs to the C. annuum var. Aviculare, while Bolita and Pico Paloma belong to C. annuum and C. frutescens, respectively, their morphological characteristics are very similar, since both are landraces with wild characteristics.
Phenotypic diversity using qualitative data
The analysis of qualitative traits indicated a high morphological diversity among the landraces that was supported mainly by 14 variables. IPGRI (1995) reported that 14 of the 39 qualitative variables evaluated in this study are highly discriminatory for Capsicum species. However, in the present study, only four variables (seed colour, number of flowers per axil, number of seeds per fruit and plant growth habit) coincided with the discriminatory variables in the descriptor guide. Additionally, variation in the margin of the leaf blade in this study allowed the discrimination of variation among the nine landraces. Similar results were observed by Orobiyi et al. (Reference Orobiyi, Loko, Sanoussi, Agré, Korie, Gbaguidi, Adjatin, Agbangla and Dansi2018). In contrast, Narez-Jiménez et al. (Reference Narez-Jiménez, de-la-Cruz-Lázaro, Gómez-Vázquez, Cruz-Hernández, Márquez-Quiroz and Castañón-Nájera2014) reported that the leaf blade margin and seed colour did not vary significantly. The 10 other traits that contributed to the greatest variation in the present study (hypocotyl pubescence, leaf blade margin, corolla spot colour, filament colour, pedicel attachment to fruit, seed size, stem shape, fruit cross-sectional corrugation, mature fruit colour and anther colour) were not reported in other studies as providing significant variation, but here they were important keys to the separation of the groups.
The pattern of clustering using the quantitative data was very similar to what was observed with qualitative data. One exception was the Rosita landrace (C. chinense), which grouped with landraces belonging to C. annuum when it was analysed with quantitative data. When analysed with qualitative data, however, Rosita formed an independent group, because its qualitative characteristics were very different from the other landraces studied. Piña-Razo (Reference Piña-Razo1982) indicated that C. chinense and C. annuum are mainly differentiated based on qualitative traits such as the presence of three to four flowers per axil in the plant. Smith and Heiser (Reference Smith and Heiser1957) indicated that C. chinense is the only species of the annuum complex with a yellow-green corolla, which could influence the separation of the Rosita landrace (belonging to C. chenense) as an independent group when it was analysed with qualitative data. With either quantitative or qualitative data, the landraces of chilli pepper belonging to C. annuum species formed a large group; similarly, Maax, Bolita and Pico Paloma landraces (C. annuum var. aviculare, C. annuum and C. frutescens, respectively) formed a solid group, perhaps because they are the least domesticated of the genotypes analysed here.
Capsaicinoid content
Moisture and capsaicinoid content varied widely among the landraces of chilli pepper, highlighting their importance as a germplasm reservoir for genetic improvement. Similar results in capsaicin content of autochthonous pepper fruits of Capsicum annuum in Tunisia were reported by Lahbib et al. (Reference Lahbib, Dabbou, Bok, Pandino, Lombardo and Gazzah2017). Although capsaicin levels in Rosita and Xcat'ik landraces reported here were lower than in Habanero Orange (C. chinense) and Xcat'ik (C. annuum) chilli peppers reported by Cisneros-Pineda et al. (Reference Cisneros-Pineda, Torres-Tapia, Gutiérrez-Pacheco, Contreras-Martín, González-Estrada and Peraza-Sánchez2007), the Maax landrace (C. annuum var. aviculare) here had more capsaicin than they reported for landraces of Maax chilli pepper. The variations in capsaicinoid content among the nine chilli pepper landraces can be attributed to differences in the species, genotype, environmental conditions and the extraction method (González-Zamora et al., Reference González-Zamora, Sierra-Campos, Luna-Ortega, Pérez-Morales, Rodríguez Ortiz and García-Hernández2013). Capsicum chinense is one of the most cultivated species and has been selected for pungency. In addition, del López Castilla et al. (Reference del López Castilla, Hernández, Castillo Aguilar, Martínez-Hernández, Ortiz-García and Andueza-Noh2019) showed that the Rosita landrace had the lowest level of genetic diversity and Pico Paloma the highest, indicating that this genetic diversity may have influenced the general characteristics of the peppers. Islam et al. (Reference Islam, Sharma, Sinha, Negi, Neog and Tripathi2015) reported variation in the total capsaicinoid content from 20 to 69,240 μg/g in 139 accessions of Capsicum from India. Kethom et al. (Reference Kethom, Tongyoo and Mongkolporn2019) also reported large differences in pungency values for C. frutescens (TRF239) at 214,678 SHU and C. annuum (TRF239) 3220 SHU. Cisneros-Pineda et al. (Reference Cisneros-Pineda, Torres-Tapia, Gutiérrez-Pacheco, Contreras-Martín, González-Estrada and Peraza-Sánchez2007) reported variation in capsaicinoid content from 1010 to 71,789 μg/g for seven cultivars of chilli peppers from Yucatan State. Orellana-Escobedo et al. (Reference Orellana-Escobedo, Garcia-Amezquita, Olivas, Ornelas and Sepulveda2013) also reported a range of variation in capsaicinoid content between 22.8 and 9097 μg/g of Mexican chilli peppers, widely used in the state of Chihuahua, Mexico. These authors indicated that the variation in the capsaicinoid content is due to genetic factors or environmental conditions.
Conclusions
Morphological variability was high among the nine landraces of chilli pepper studied from Maya communities of the Yucatan Peninsula, highlighting the good conservation status of the landraces studied. The greatest variability was observed in the qualitative variables, for which three groups were differentiated morphologically. The qualitative variables demonstrated that although C. annuum, C. chinense and C. frutescens belong to the same genetic complex (annuum complex), these three species can be differentiated. Because morphological traits are influenced by the environment, biochemical or molecular markers are recommended to better estimate the diversity of these landraces. Capsaicinoid content also varied considerably among the nine landraces. Capsicum chinense (Rosita) had the highest level of capsaicinoids, while the Dulce landrace had none. The chilli pepper landraces with a high level of capsaicinoids characterized in this study will be useful for further use in participatory breeding programmes to generate new varieties with better agronomics traits. The results overall highlight the importance of the traditional milpa crop system and Maya farmers in maintaining and generating the genetic diversity of these species. We emphasize the importance of initiating programmes for the management, use and conservation of chilli pepper landraces involving the Maya farmers who are the principal heirs, custodians and breeders of native germplasm.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1479262121000204.
Acknowledgements
The authors thank all Maya farmers from Campeche state that provided the seeds of the chilli pepper landraces used in this study, and especially to the local farmer from Palizada, Campeche for the important role in maintaining the diversity of chilli pepper landraces. Also, we would like to thank the three anonymous reviewers and the Editor in Chief for their suggestions and comments.
Conflict of interest
None.
