Safflower (Carthamus tinctorius L.) belongs to the Asteraceae family and it is believed to have domesticated in the Fertile Crescent region (Knowles and Ashri, 1995). Kupsow (1932) first reported that the safflower had two centers of variability: Ethiopia and Afghanistan. Later, Hanelt (1961) reported the Afghanistan and India area a primary center of variability and the Ethiopia-Egypt area a secondary center. Up to now, ten geographic centers of similarity throughout the world have been proposed based on plant morphology of safflower (Chapman et al., 2010). Safflower is phenotypically differentiated from its progenitor species, grown over a much larger land area, and had a long history of cultivation, and it has been described as a ‘strongly domesticated’species (Demperwolf et al., 2008). India, Pakistan, Iran and Afghanistan had identified as centers of safflower (Knowles, 1969). However, the differences in safflower gene pool regarding the phenotypic characters are reported between the centers (Chapman et al., 2010). Safflower is considered as neglected crop though it provides opportunity through great genetic diversity and diverse agro-ecological adaptation (Padulosi et al., 1999; Thies, 2000). Many germplasm of safflower have been collected in Korea from different countries since few years and this crop is expected to develop in future for the diversified nutritional needs of people.
Safflower is a highly branched, herbaceous, thistle-like annual plant usually with many long sharp spines on the leaves. Plants are 30-200 cm tall and have strong taproot which support them to thrive in dry climates. It is considered as multipurpose crop species, grown for oil, medicinal and industrial uses from Mediterranean region to the Pacific Ocean (Khan et al., 2009). The safflower oil is very low in saturated fatty acids (Knowles, 1958). The safflower germplasm had been identified much variation in high oleic and linoleic acid and the variation ranged from 3.1 to 90.6% and from 3.9 to 88.8%, respectively (Fernanadez- Martinez et al., 1993). Johnson et al. (1999) used the oil and fatty acid content variation to differentiate accessions. Oleic acid and linoleic acid are the predominant fatty acids in safflower which showed a strong negative correlation (Khan et al., 2009). Studies on germplasm assessment for morphological and fatty acid composition of crop plants had been done by many researchers (Velasco et al., 1998; Johnson et al., 1999, Shim et al. 2004, Uzun et al. 2008; Khan et al., 2009). These studies revealed the wide variation in the proportions of saturated and unsaturated fatty acids, which offered the possibilities of developing superior quality edible oils and specialized industrial oils (Bajpai et al. 1999).
The National Agrobiodiversity Center (NAC) of Rural Development Administration (RDA) has been collected the safflower germplasm from different countries since few years and these germplasm should have considerable variation for morphological characters and oil content along with fatty acid compositions. But studies on morphological and biochemical characters are limited in safflower germplasm of RDA. Therefore, the objective of this research was to assess the variation in agro-morphological and biochemical characters of safflower accessions collected from different regions, and to examine the relationship among the variables.
MATERIALS AND METHODS
Seed materials of 172 accessions from National Agrobiodiversity Center, RDA were obtained for this study (Table 1). The accessions from India and Pakistan were categorized into South-Central Asia (S-C Asia, n=102) whereas the accessions from Iran and Afghanistan were categorized into South-West Asia (S-W Asia, n=70). The seeds were sown in April 2014 at the experimental field, Gyeonggi-do (Latitude, 37°11′59″ North and Longitude, 126°49′53″ East). The experimental field was located at 38 m above the sea level and the field can be characterized by having sandy loam soil with pH 6.0. The seedlings were transplanted in April 25, 2014 and each accession had 21 seedlings in a plot with row spacing of 30 cm, and plant to plant spacing of 20 cm at non-replicated design. All accessions were provided with the same agronomic practices and plants were drip irrigated, and all the plants were harvested manually in August, 2014.
Agro-morphological traits including plant height, leaf length and width, spines, days to flowering, flower color, seed length and seed weight were recorded. Plant height (cm) was measured at maturity when plants were in full flower from group level to the tip of the main stem. Leaf length and width (cm) were measured from the base to the tip of the leaf using the scale. The spines of bract were recorded in 0 to 2 score, 0 = absence of spine or spineless, 1 = short spine (< 2 mm), and 2 = long spine (> 2 mm). Days to flowering was counted as number of days from the date of sowing to the date when at least three plants showed open flowers (Khan et al., 2009). Flower color was recorded in 1 to 5 score (1 = yellow, 3= orange and ivory, 5 = red). Seed was harvested manually at maturity. Seed length (cm) was measured using scale whereas seed weight (g) was measured using digital balance. The measurement on morphological characters was taken on three samples and was averaged them. Gas Chromatography (GC, Agilent 7890A, USA) was used to analyze the biochemical characters (palmitic acid, stearic acid, oleic, linoleic and linolenic acid) using the procedure described by Velasco et al. (1998), and total oil content was determined by SoxtecTM 2043.
Means, standard deviation, maximum, and minimum values for morphological and biochemical characters of the groups were analyzed. Pearson correlation coefficient was calculated for all the agro-morphological and biochemical variables. SPSS Statistics 17.0 was used to analyze the data for descriptive statistics and correlation coefficient. Principal component analysis (PCA) was used to analyze on the morphological and biochemical variables using Microsoft Excel 10, Multibase for Windows.
RESULTS
The evaluated safflower germplasm of two geographical regions showed the variation in all the morphological characters studied (Table 2). The mean value of plant height was higher in S-W Asia accessions than S-C Asia. The range of variability for plant height and leaf length were higher in S-W Asia accessions than S-C Asia. The short and long spine accessions were recorded in the both S-C and S-W Asia accessions. S-C accessions had flowered earlier than S-W accessions but the variation for days to flowering was higher in S-W Asia accessions. The flower color was ranged from yellow to red in the accessions of both regions. The variation for seed length was not differed between two groups of accessions. The mean values for seed weight was higher (69.5 g) in S-C Asia accessions and the maximum value was higher as 9.7% than S-W Asia accessions. The variation for seed weight was higher in S-C Asia than S-W Asia accessions (Table 2).
Variation in fatty acid composition and total oil content of safflower germplasm is given in Table 3. The germplasm exhibited large variation in all biochemical variables. Mean values of palmitic acid and linolenic acid did not differ among two groups. However, S-C Asia accessions exhibited higher mean oleic acid and total oil content than S-W Asia accessions. Mean value for oleic acid in SC Asia accessions was 6% higher than S-W Asia accessions (Table 3). Similarly, the variation for oleic acid was higher S-C Asia accessions. The mean values for linoleic acid was higher in S-W Asia accessions whilst maximum values for linoleic acid was about 1.2% higher than S-W Asia accessions. Similarly, the variation for linoleic acid was higher in S-C Asia accessions. Linolenic acid had not much differed between two groups of accessions. With respect to total oil content, the mean values were higher in the accession of S-C Asia and maximum values for total oil content was 6.5 % (data not shown) was higher than SW Asia accessions.
The results of correlation analyses among plant morphological and oil composition of the accessions are shown in Table 4. Plant height had significantly positively correlated with leaf length, days to flowering, palmitic acid, and stearic acid but negatively correlated with seed length and seed weight. Likewise, leaf length had a significant positive correlation with leaf width, days to flowering, stearic acid and linolenic acid. Leaf width was significantly positively correlated with seed length. Days to flowering had a positive correlation with palmitic acid, stearic acid, linolenic acid and total oil content but it had a negative correlation with seed length and width. The seed length had significantly positively correlated with seed weight and linoleic acid but negatively correlated with stearic, oleic and linolenic acid. The seed weight had a significant negative correlation with stearic acid, oleic acid, and linolenic acid. Likewise, palmitic acid had a significant positive correlation with stearic acid and linoleic acid but it had a significant negative correlation with oleic and total oil content. Stearic acid was positively correlated with linolenic and total oil content whereas oleic acid had a strong negative correlation with linoleic acid. The oleic acid had significantly positively correlated with linolenic acid. The linoleic acid had a negative correlation with linolenic acid.
The PCA analysis on all morphological and biochemical variables revealed that the first four principal components (PC4) accounted 66.0% of the total variation (Table 5). Total oil content, stearic acid and leaf traits contributed to this variation. The first principal component (PC1) contributed 26.0 % of the variation, showing the highest contributions from the proportion of spines, seed weight, seed length and days to flowering. Likewise, the third principal component (PC3) contributed 57.0% of variation exhibiting highest contributions from leaf length, width, seed length and weight. The second principal component (PC2) contributed 45% variation showing the highest contribution from oleic acid and linoleic acid.
DISCUSSION
Our study showed the variation on morphological characters between the germplasm of S-C and S-W Asia. Plant height, leaf length and days to flowering exhibited large variation between two groups of germplasm. We observed large variation in plant height, leaf length and days to flowering within the accessions of S-W Asia. The accessions of S-C Asia exhibited greater variation in seed length and weight. In the study of Khan et al. (2009), they reported a large amount of diversity for agro-morphological traits among the different groups of safflower accessions. Other studies (Ashri 1975; Jaradat and Shahid, 2006) had reported a wide range of variation in safflower collections for agro-morphological traits which might be evolved from human and natural selection. Variation in traits suggests good potential for selection in a breeding program (Pascual-Villalobos and Albuquerque, 1996).
Our results exhibited large differences in palmitic acid, oleic acid, and linoleic acid within the germplasm of S-C Asia whereas total oil content revealed large variation within the accession of S-W Asia. Uzun et al. (2008) reported the percentage of palmitic acid, stearic, oleic, linoleic acid in the seed ranged between 8.0-10.3, 2.1-4.8, 40.7-49.3, and 29.3-41.4 % in sesame seed, respectively. But our study exhibited the percentage of palmitic, stearic, oleic, and linoleic acid in the safflower seed of S-C accessions ranged between 4.8-7.6, 1.3-3.3, 10.7-75.7 and 15.5- 79.9 %, respectively. We found higher oleic acid in S-C Asia accessions. Knowles and Mutwakil (1963) reported the germplasm of India and Bangladesh had high oleic acid which confirmed our results. In the study of Khan et al. (2009), they reported 63% to 83% linoleic acid in Eastern European safflower accessions but in our study, it ranged from 15.5 to 79.9% in S-C Asia accessions and from 69.0% to 78.7% in S-W Asia accessions. Gambacorta et al. (1997) reported that safflower oil contained 7.4% palmitic, 2.9% stearic, 14.3% oleic and 74.3% linoleic acids in Italian accessions. We observed different values for fatty acid composition in various accessions but environmental condition influence more in fatty acid composition than does the genotype of the variety (Robertson et al. 1978). The palmitic acid, oleic acid, and linoleic acid appeared at much diversity in S-C Asia accessions than in S-W Asia. However, total oil content showed more diversity in S-W Asia accessions. Therefore, the scope for substantial improvement in oleic acid and linoleic acid using the germplasm of S-C Asia germplasm has greater than S-W Asia germplasm. The linoleic acid was the major fatty acid of safflower and the averaged values of linoleic acid from SC Asia and S-W Asia accessions were 70.9 and 75.9 %, respectively. Matthaus et al. (2015) reported the average value of linoleic acid was 70.6 % which is close to our findings.
The significant correlation (r = 0.625**) between days to flowering and plant height was found in this study which had also reported by Khan et al. (2009). Palmitic acid was positively correlated with both stearic acid (r = 0.282**) and linoleic acid (r = 0.444**) but negatively with oleic acid (r = –0.493**) which is inconsistent with the findings of Khan et al. (2009). Our study showed a strong negative relationship (r = –0.977**) between oleic acid and linoleic acid. This relationship is well documented and reported in safflower (Johnson et al., 1999; Knowles and Hill, 1964), sesame (Uzun et al., 2008; Yermanos et al., 1972; Were et al., 2006; Brar, 1982) and crucifers (Mandal et al., 2002). Knowles (1969) explained the genes involved in the inverse relationship between linoleic and oleic acid in safflower. Fernanez-Martinez et al. (1993) also reported that direction of correlation between certain fatty acids depend on the environment where the genotypes were grown. PCA identified oleic, linoleic and total oil content as traits responsible for variation in germplasm which indicates the possibility of improving oil quality through breeding (Fernandez- Martinez et al. 1993). In conclusion, variation in morphological and biochemical characters existed at both groups. The variance in plant height, leaf length and days to flowering was higher in S-W Asia accessions whereas S-C Asia accessions possessed higher variance for palmitic, oleic, and linoleic acid. S-C Asia accessions exhibited higher total oil content as compared to S-W accessions. The significant negative relationship observed between oleic and linoleic acid. Thus, the variation observed in morphological and biochemical characters in S-C and S-W accessions provides useful information to breeder for developing new safflower cultivars with high oil content with better fatty acid compositions. The safflower accessions rich in oil component could be used as parental lines in breeding program to enhance oil quantity and quality.
