Characterization of Capsicum germplasm collected in Northwestern Argentina based on morphological and quality traits

Paprika production is concentrated in Northwestern Argentina, where around 1500 ha are annually grown. Farmers mainly from Andean valleys, maintain their local landraces or “criollas” under different environmental conditions, using their farming traditions and selection criteria. These traditional varieties represent valuable genetic resources for conservation and selection of agronomical and quality traits. Between 2005 and 2007 eight collecting trips were made to Northwestern Argentina in order to recover local germplasm. As a result, five hundred new vegetable seed samples have been incorporated into the Germplasm Bank. Among these accessions 32 belong to Capsicum, the majority for paprika production, that were evaluated in field conditions using a randomized block design with three repetitions of eight plants per accession; 54 morphological and agronomical traits were measured during two production cycles. Multivariate analyses determine three groups, where fruit characters were the most efficient for accession differentiation. Accessions suitable for paprika production were further examined, and variability for fruit pungency and color was found. Promising accessions were selected for ex situ conservation and breeding purposes.


InTrODuCTIOn
Andean civilizations in America domesticated several species for different purposes; peppers are found among these species, (Popenoe et al., 1990).An area comprising Bolivia, Northern Argentina, Paraguay and middle and Southern Brazil has been considered the center of origin of Capsicum species (Pickersgill 1969(Pickersgill , 1997)).In Northwest Argentina, farmers have grown pepper crops following their traditional practices maintaining local populations under different environments over many years, resulting today in genetically different populations (Asprelli et al., 2011).These types of populations are named "criollas", "local" or landraces (Zeven, 1998).Nevertheless, there have not been systematic efforts in Argentina to recover these valuable resources, which are important for breeding programs (Thul et al., 2006;Rêgo et al., 2011b).
Genetic diversity among and within species of the genus Capsicum has been extensively investigated (Souza & Maluf, 2003); the level of variation within the domesticated species is lower than in their wild relatives, and variation in peppers with elongated and sweet fruit is limited compared to the diversity found among pungent peppers (Lefebvre et al., 1993;Portis et al., 2006, Marame et al., 2009).Morphological and agronomical characters have been useful to evaluate local pepper collections in many countries like United States (Smith et al., 1987), Yugoslavia (Zewdie & Zeven, 1997), Colombia (Medina et al., 2006;Pardey et al., 2006), Italy (Portis et al., 2006), Turkey (Bozokalfa et al., 2009), Peru (Ortiz et al., 2010), Brazil (Rêgo et al., 2009;2011a,b;Rodrigues Monteiro et al., 2010).In Argentina, genetic resources of vegetable crops are preserved ex situ at the Germplasm Bank of the Agricultural Experimental Station La Consulta INTA (Mendoza).Pepper genetic diversity maintained there was considered limited and local landraces were poorly represented in the germplasm collection until the year 2004.This concern brought the need for the recovery and evaluation of local populations.
Our work was specially directed towards the characterization of Capsicum germplasm suitable for paprika production.In Argentina around 13,000 ha of pepper are cultivated annually among them 1,500 ha are for paprika production, which is concentrated in Northwestern Argentina, in the provinces of Salta, Tucumán and Catamarca (Galmarini, 2000).Two paprika populations of Capsicum annuum L. are mainly cultivated, one with elongated fruits, locally called "Trompa de elefante" (elephant trunk), and another with round fruits known as "Bolita Salteño" or "Ñora" (Galmarini, 1993).Traditionally, Argentine market demands sweet paprika with intense color for dehydrated products.However, a common practice done by farmers in Northwest Argentina is to cultivate spicy peppers in family orchards in the vicinity of sweet paprika crops (Occhiuto et al., 2006;Occhiuto 2009).As a consequence, contamination with pollen coming from plants with pungent fruits is quite common, since natural crosses may occur depending on varieties, environment conditions and the type and frequency of pollinators (Portis et al., 2004;Thul et al., 2006).
The main objective of this work was to evaluate local germplasm using morphological and agronomical traits.This characterization contributes to assess the variability of the accessions, to classify them for ex situ preservation in the Germplasm Bank, and also to determine promising materials suitable for breeding programs.

Germplasm
Eight collecting trips were performed to Northwestern Argentina between 2005 and 2007.The itinerary included small villages, some of them isolated in high mountain valleys, in the provinces of Catamarca, Tucumán, Salta and Jujuy.In these rural communities, 103 growers were interviewed through informal surveys requiring information about traditional cultivation practices, consumption habits, criteria for selection of the fruits and seed maintenance methods (Peralta et al., 2008).A total of 511 vegetable crop samples, representing 33 of different genus were recovered, among them 32 Capsicum accessions.These pepper samples were mainly collected in Andean valleys where paprika production is concentrated (Table 1, Figure 1).Seed samples were incorporated into the Germplasm Bank of the Agricultural Experimental Station La Consulta INTA.

Experimental field and design
The trial was carried out on the experimental field of the Institute of Horticulture (33º 00,268'S, 68° 52,254' W, 912 m.a.s.l.) at the Faculty of Agricultural Sciences, Chacras de Coria, Luján de Cuyo, Mendoza, Argentina.Thirty two pepper accessions collected in Andean valleys of Northwestern Argentina, and a commercial pepper UCODULCE INTA selected for paprika production (Galmarini & Fuligna, 2003) were evaluated in two consecutive seasons (September 2005-April 2006and September 2006-April 2007).A randomized block design with three repetitions of eight plants per accession was used.The plants were placed at 0,45 m within row and 0,8 m between rows, cultivation was carried out under anti-hail mesh and drip irrigation.

Characterization
Characterization of the accessions was based on international descriptors for Capsicum (IPGRI, AVRDC and CATIE, 1995).Thirty five qualitative and 19 quantitative traits belonging to morphological, phenological, agronomical and fruit quality characters were evaluated (Table 2).Morphological, phenological and agronomical variables were calculated as averages of eight plants per repetition.External color, pungency, pH and soluble solids were measured on 10 ripe fruits from each replicate.
Fruit pungency was assessed using a sensorial method (Barbosa et al., 2002), the percentage of pungent fruits per accession was established,  1) and the angular transformation of the average value was calculated.The external fruit color was measured with a Konica Minolta CR-400 colorimeter.The Hunter scale parameters (L, a, b), where "L" measured the brightness of the sample; "a" measured the tone of red to green colors; and "b" the tone of yellow to blue colors.With these parameters the hue angle was calculated (hue= arc tan b/a) and purity of color (chroma= √(a2 + b2); the value of "L" brightness was obtained as a direct measure of the colorimeter (Moreno-Pérez et al., 2006).A multiple correspondence analysis was applied to the qualitative variables using the programme Statistica 6.0 (Statsoft, 2004).The significance of the quantitative variables was tested through the analysis of variance, for a p< 0,05.
A grouping analysis of the 33 accessions was performed using significant quantitative variables.Only significant continuous variables were included in the analysis, and the algorithm of Ward was used.A detailed analysis for pepper accessions suitable for paprika production was performed.
The explanatory variables were as well identified (Matus et al., 1996).

Quantitative Characters
There were no significant effects for blocks and years (data not shown), so the average was calculated using the data of two seasons.Among the 19 quantitative traits, four (plant width, leaf length and width, and days from transplant to fruiting), did not show significant differences for a level of p< 0,05, all the others were polymorphic (Table 4).

Cluster Analysis
The groups obtained were similar to those found using multiple correspondence analyses.Fruit characters were the most influential variables in clustering results (Figure 3, Table 5).
Accessions corresponding to groups I and II belonged to the species C. annuum.Group I included 5 entries, with triangular fruits of greater pulp thickness (0,36 cm), wider (6,21) and smaller length.
Group II, included 25 entries with elongated fruits.This group subdivided two subgroups: "a" including fruits of greater length and lower flesh acidity, and "b" grouping less pungent and reduced pulp thickness fruits.
Group III included 3 accessions of spicy peppers, corresponding to C. baccatum, characterized by the largest number of fruits per plant (113,53), less fruit weight (6,94 g) and size, and pungent fruits; the subgroup "c" was highlighted by its elongated fruits.A subgroup "d" was separated by its square fruit shape.

Evaluation of pepper accessions suitable for paprika production
Principal Component Analysis (PCA): PCA was performed among 25 accessions of group II that were suitable for paprika production.Three components explained 60,4% of the variability (Figure 4).The first principal component (33,4%) included the variables: leaf length and width, fruit weight and width, and components of the external fruit color (brightness, tone and purity), length of peduncle, pulp thickness and fruit pungency.The second principal component explained 14% of the variation, and the variables were number of days to flowering and fruiting, and number of fruits per plant.Finally, on the third component (13%) the variables that contributed to explain variation were placenta and fruit length.
Accessions 29, 30, 31 and 32 have the best   Table 1) Table 5 Means, standard deviations, and ranges of significant quantitative characters of the accession groups
Regarding pungency the control (33) and the accession 28 were significantly different due to the absence of pungent fruits, while accession 16 presented the greatest pungency.The 22 remaining accessions were not statistically different among themselves (Figure 5).Ninety six percent (96%) of the evaluated samples presented different degrees of pungency.However, a great variation within each sample was found (the coefficient of variation was 55,3%).

DIsCussIOn
This is the first study that deals with Capsicum landraces variability in Argentina.Information about crop management practices, seed conservation and food uses by small farmers was recovered.A significant number of recovered accessions are now maintained in the Germplasm Bank at E.E.A La Consulta INTA.
Morphological and agronomical characters allowed the characterization of 33 accessions into 3 groups.The phenotypic variability in "criollo" pepper germplasm was demonstrated through various multivariate methods.The variables that discriminate the groups in this study coincide with other evaluations made to characterize pepper introductions in other countries (Medina et al., 2006;Pardey et al., 2006;Bozokalfa et al., 2009;Thul et al., 2009).The number of fruits per plant was inversely related to their weight, this relation was also described by Zewdie & Zeven (1997), Portis et al. (2006), Somashekhar and Salimath (2008) and Schuelter et al. (2010).
The higher the soluble solids content, the lower the water content in the fruit, this is a very important trait in Capsicum breeding for the dehydration industry (Lannes et al., 2007).In our study accessions of Group III showed the highest soluble solids content (13,88).
On the other hand, fruit wall thickness and fruit weight were positive correlated, which was also previously reported by Lannes et al. (2007).These traits are important for the development of fresh market varieties, as pepper fruits with thicker pericarp are more resistant to postharvest and transportation damage.
As a result of pepper germplasm evaluation, 14 promising accessions were selected to be incorporated in a paprika breeding program, which is important for the local industry.A great variability for pungency was found in the evaluated germplasm; this fact was also found in other studies carried out in other countries (Nuez et al., 1996;Zewdie & Bosland, 2000;Barbosa et al., 2002).Argentine market demands sweet paprika, therefore pungency of the fruits, after the intensity of color, is considered one of the most important features for cultivar selection.This variability would allow selecting plants with less pungency within the entries that have other important features for breeding (Rêgo et al., 2011a).This result confirms that farmers indicated during the collection trips, regarding the increase of pungency in recent years.It should be noted that pepper prices paid to growers decreases with the proportion of spicy fruits (Occhiuto, 2009).Our results suggest that isolated areas, where spicy peppers are not produced along with sweet paprika, should be used in order to avoid seed contamination problems.

aCKnOWLeDGemenTs
The authors thank María Makuch, Inés Lorello and Leonardo Togno for their collaboration in collection trips and introduction of seed accession in the Germplasm bank, Rafael Bottero for drawing the map, Adriana López for her advice in data analysis, and Cristian Delú Lozano for his collaboration in field evaluation.

Figure 1
Figure 1 Collection sites in Northwestern Argentina.Number indicates pepper samples corresponding to the accessions evaluated (Table1)

Figure 2 .
Figure 2. Multiple correspondence analyses, showing three groups based on 11 qualitative characters of 33 pepper accessions

Figure 3
Figure 3 Dendrogram of 33 pepper accessions (numbers are indicated in Table 1)

Figure 4 .Figure 5
Figure 4. Distribution of 25 pepper accessions in the first and second primary components.(Variables codes are indicated in Table4)

Table 2 .
Morphological and agronomical traits measured in

Table 3 .
Summary of differential qualitative characters present in different accessions groups green hypocotyls, erect plant growth habit, ovate leaf shape, dark green leaf color, purple anther, triangular fruit, fruit with 3 locules, cordate peduncle attachment II