Introduction
After being part of the daily diet of hunter-gatherers, peas (Pisum sativum L.) are being cultivated since antiquity participating in the diet of both modern humans and animals (Cousin, Reference Cousin1997; Smýkal et al., Reference Smýkal, Aubert, Burstin, Coyne, Ellis, Flavelle, Ford, Hýbl, Macas, Neumann, McPhee, Redden, Rubiales, Weller and Warkentin2012). The most widely cultivated pea varieties are P. sativum subsp. sativum var. sativum L. and P. sativum subsp. sativum var. arvense (L.) Poiret, utilized for human consumption and livestock feed, respectively (Smýkal et al., Reference Smýkal, Aubert, Burstin, Coyne, Ellis, Flavelle, Ford, Hýbl, Macas, Neumann, McPhee, Redden, Rubiales, Weller and Warkentin2012; Weeden, Reference Weeden2018). Wild subspecies, like P. sativum subsp. elatius (Bieb.) Aschers. et Graebn. s.l., are considered as the basis for the domestication of cultivars (Weeden, Reference Weeden2007).
In Greece, split legume grains, mainly Pisum and Lathyrus, are used to cook fava puree (or fava), a local dish dating back to antiquity (Halstead and Glynis, Reference Halstead and Glynis1989; Glynis and Halstead, Reference Glynis and Halstead1993). In insular Greece (e.g. islands Schinoussa, Amorgos, Skyros), small-seeded landraces of P. sativum subsp. sativum are used for fava; these landraces belong to P. sativum subsp. sativum, but are closely related to subsp. elatius, the wild subspecies (Stavridou et al., Reference Stavridou, Lagiotis, Karapetsi, Osathanunkul and Madesis2020).
Pisum sativum is a species with inherently high diversity in agronomic and seed traits (Annicchiarico et al., Reference Annicchiarico, Romani, Cabassi and Ferrari2017; Solberg et al., Reference Solberg, Yndgaard, Poulsen and von Bothmer2017; Gali et al., Reference Gali, Sackville, Tafesse, Lachagari, McPhee, Hybl, Mikić, Smýkal, McGee, Burstin, Domoney, Ellis, Tar'an and Warkentin2019), an indispensable prerequisite for the genetic improvement of the crop (Swarup et al., Reference Swarup, Cargill, Crosby, Flagel, Kniskern and Glenn2021). In this line, landraces, which are genetically diverse, locally adapted, lacking of formal crop improvement populations (Camacho Villa et al., Reference Camacho Villa, Maxted, Scholten and Ford-Lloyd2005), may host genetic diversity potentially exploitable for crop improvement in the challenging future conditions (Dwivedi et al., Reference Dwivedi, Ceccarelli, Blair, Upadhyaya, Are and Ortiz2016; Preston et al., Reference Preston, Ford-Lloyd, Smith, Sherman, Munro and Maxted2019).
Pea genotypes show high variability for grain nutritive components like protein or starch (Annicchiarico et al., Reference Annicchiarico, Romani, Cabassi and Ferrari2017; Solberg et al., Reference Solberg, Yndgaard, Poulsen and von Bothmer2017; Tao et al., Reference Tao, Afshar, Huang, Mohammed, Espe and Chen2017; Mohammed et al., Reference Mohammed, Chen, Walia, Torrion, McVay, Lamb, Miller, Eckhoff, Miller and Khan2018), but also for their phenolic content and thus antioxidant activity (Devi et al., Reference Devi, Sanwal, Koley, Mishra, Karmakar, Singh and Singh2019; Zhao et al., Reference Zhao, Su, Qin, Wang and Kang2020). It has been found that phenolic content is associated with seed coat (testa) colour with dark-seeded peas being enriched in phenolics (Xu et al., Reference Xu, Yuan and Chang2007; Jha et al., Reference Jha, Purves, Elessawy, Zhang, Vandenberg and Warkentin2019) and it was suggested that they could be used as a functional food (Zhao et al., Reference Zhao, Su, Qin, Wang and Kang2020). Apart from testa colour, hilum colour has been found to associate with seed traits like phenolic content (Zhao et al., Reference Zhao, Su, Qin, Wang and Kang2020). If associated with certain seed traits, both testa and hilum colour could be used as handy and friendly biomarkers for phenotypic breeding/selection efforts (Chukwumah et al., Reference Chukwumah, Walker and Verghese2009; Reynolds et al., Reference Reynolds, Chapman, Crespo-Herrera, Molero, Mondal, Pequeno, Pinto, Pinera-Chavez, Poland, Rivera-Amado, Saint Pierre and Sukumaran2020).
Seed phenolics could act as a chemical defence against pea weevil (bruchid, Bruchus pisorum L.), the major insect pest of dry peas, which claims a very high (>60%) yield toll yearly (Teshome et al., Reference Teshome, Mendesil, Geleta, Andargie, Anderson, Rämert, Hillbur, Dagne and Bryngelsson2015). Selecting and growing tolerant genotypes is a low-cost, effective and sustainable way to cope with bruchids (Mishra et al., Reference Mishra, Macedo, Panda and Panigrahi2018). Traits like seed size, testa colour and seed toughness (ST) have been reported to associate with tolerance to bruchids. Small seeds, offering less food to developing larvae, are also tougher and thus more tolerant to bruchids (Sadakiyo and Ishihara, Reference Sadakiyo and Ishihara2012; Fricke and Wright, Reference Fricke and Wright2016; Tsialtas et al., Reference Tsialtas, Irakli and Lazaridou2018). Also, dark-seeded peas were found to be less susceptible, possibly due to the higher phenolic contents (Teshome et al., Reference Teshome, Mendesil, Geleta, Andargie, Anderson, Rämert, Hillbur, Dagne and Bryngelsson2015; Zhao et al., Reference Zhao, Su, Qin, Wang and Kang2020).
The present study aimed to identify putative associations between testa and hilum colour with seed traits (seeds/pod, protein concentration (PC), 1000-seed weight (TSW), phenolic contents, toughness, tolerance to bruchids) in four phenotypes of a field pea landrace in order colour to be used as phenotypic biomarker in selection schemes.
Materials and methods
Genetic material and experimental set up
Four phenotypic accessions were selected out of the seeds of ‘Katsouni’ pea landrace (P. sativum L. subsp. sativum; Tsialtas and Tan, Reference Tsialtas and Tan2017) from Schinoussa island (36°52´ N, 25°31´ E). The selection was based on testa (green or green with black spots) and hilum colour (white or black) and the four phenotypes were encoded as green testa and white hilum (TGHW), green testa and black hilum (TGHB), spotted testa and white hilum (TSHW) and spotted testa and black hilum (TSHB).
The experiment took place in the farm of Aristotle University of Thessaloniki (40°32´1 N, 22°59´3 E, 0 m a.s.l.) on a typic xerorthent loam, under rain-fed conditions, during 2016–2017 (hereafter 2017) and 2017–2018 (hereafter 2018) growing seasons. The growing season of 2018 (December to May) was wetter and warmer than the 2017 growing season (255 versus 179 mm and 13.5 versus 10.5°C).
The arrangement of the four phenotypes followed the randomized complete block (RCB) design with three replications. Each plot (1 m2 area) consisted of four, 1 m long rows, at 0.25 m separation in accordance to our previous works (Tsialtas et al., Reference Tsialtas, Irakli and Lazaridou2018, Reference Tsialtas, Theologidou, Bilias, Irakli and Lazaridou2020; Boulata et al., Reference Boulata, Irakli and Tsialtas2022). The seeding, at a rate of 40 seeds/m2 (10 seeds per row), was conducted by hand on 3 December 2016 and 7 December 2017. The blocks were separated by a 1.5 m buffer zone. No fertilization or irrigation was supplied and the weeds were removed by hand, when necessary.
Assessments on pods and seeds
At full seed maturity (BBCH 89; Lancashire et al., Reference Lancashire, Bleiholder, Van den Boom, Langelüddecke, Stauss, Weber and Witzenberger1991), all the pods were collected from the two internal rows of each plot and kept in paper bags. Fifteen randomly selected pods per plot were used to measure the number of seeds/pod (SpP) and the mean value per plot was calculated. Then, all the pods per plot were threshed by hand and seeds kept in paper bags at room temperature and darkness.
Hundred seeds per plot were randomly selected and placed in plastic boxes (150 ml) and kept at room temperature (25/20°C day/night temperature) for over three months until exits of the adult bruchids consummated. The damage caused by bruchids and its endoparasitoid wasp (Triaspis thoracica Curtis, 1860, Hymenoptera: Braconidae) on the seeds was estimated macroscopically by examining all the seeds per box as described by Tsialtas et al. (Reference Tsialtas, Irakli and Lazaridou2018). The bruchid infestation level (BI) was calculated as the sum of bruchid-damaged (BD) and parasitoid-damaged (PD) seeds and expressed in percentage (%).
Seed coat colour was assessed using a CR-400/410 chroma meter (Konica Minolta, Kyoto, Japan). Uninfected seeds of each plot were placed in a cylindrical petri dish, 39 mm internal diameter, firmly fixed to instrument's glass light projection tube (CR-A33a). Chroma meter measurements were based on the International Commission of Illumination (CIE) colour solid scale (L*, a*, b*). Specifically, L* represents lightness (from black = 0 to white = 100), a* is for greenness and redness (red = positive value and green = negative value) and b* corresponds to blueness and yellowness (yellow = positive value and blue = negative value). Details on the testa colour assessment are given by Tsialtas et al. (Reference Tsialtas, Irakli and Lazaridou2018).
A sample of 100 intact seeds per plot was dried at 75°C till constant weight. Dry weights were estimated and TSW was calculated by multiplying the dry weights by 10.
From the samples used for TSW calculations, 10 randomly selected seeds per sample were used to assess ST, the essential energy (J/g) for seed breaking by compression test. For the ST measurements, a texture analyser (TA-XT2i, Stable Micro Systems, Godalming, Surrey, UK) was used and the procedure followed Tsialtas et al. (Reference Tsialtas, Irakli and Lazaridou2018). The reported value for each plot represented the average of the 10 measurements. The lower the fracture energy, the softer the seeds were.
All the 100 dried seeds used for TSW calculations were ground to fine powder using an ultra-centrifugal mill ZM-1000 (Retsch GmbH, Haan, Germany) equipped with a 0.5 mm sieve. Per plot, a subsample of ca. 1.0 mg was weighed in order to determine both seed carbon (C) and nitrogen (N) concentrations on an elemental analyser (EuroEA 300, EuroVector SpA, Milan, Italy) and the concentrations were expressed in per cent (%). Seed PC was estimated as the product %N × 6.25 and the carbon to nitrogen ratio (C/N) was also calculated.
To measure total phenols (TPhe) and tannins (TTan), ca. 0.2 g of grounded seed material per plot was extracted with 4 ml aqueous acetone (70% v/v) using ultrasonic treatment for 15 min and then, the samples were centrifuged at 2200 rpm for 10 min. After the collection of supernatants, the pellets were treated once more applying the abovementioned procedure. Folin–Ciocalteu reagent was used for the detection of total phenols and tannins. The evaluation of total phenols in the samples is a result of their absorption at 725 nm and total tannins were calculated after subtraction of non-tannin phenolics from total phenolics. Results were expressed in mg of gallic acid equivalents (mg GAE) per g sample (Tsialtas et al., Reference Tsialtas, Irakli and Lazaridou2018).
Statistical analysis
Data analysis was performed based on the over-year analysis of variance as RCB design with pea phenotypes, testa colour and hilum colour as main factors and three replications. Data were analysed using M-STAT statistical package (M-STAT C, version 1.41, Crop and Soil Science Department, Michigan State University, USA). The least significant difference (LSD) test was used for means comparisons, considering P < 0.05 as significance level.
The Microsoft Excel software (v. 2010) was used to construct graphs and estimate correlation coefficients (r) between the traits for which statistically significant differences were found among phenotypes and their interaction with growing seasons. In order to avoid spurious self-correlations, coefficients were not calculated for interconnected traits, e.g. BD-BI, PD-BI, TPhe-TTan.
Results
The main factors (phenotypes, testa colour, hilum colour and growing seasons) explained the largest part of the variation of the determined traits and thus, the mean comparisons are presented for these factors.
Percentages of damaged seeds, morphological, mechanical and chemical traits
The four phenotypes differed in the percentage of bruchid infected seeds (BI) and seed phenolic and tannin concentration (Table 1). The TSHW (spotted testa, white hilum) showed to be more susceptible to bruchids (62.5%) compared to the rest three phenotypes (44.50–54.50%). It also had the highest TTan (4.51 mg GAE/g) and along with the TSHB (spotted testa, black hilum) showed the highest TPhe (7.46–7.49 mg GAE/g).
Table 1. Mean comparisons of the number of seeds per pod (SpP), percentages of damaged seeds [bruchid-damaged (BD), parasitoid-damaged seeds (PD) and percentage of bruchid infected seeds (BI)], 1000-seed weight (TSW), seed toughness (ST), total phenolic compounds (TPhe) and total tannins (TTan)
For the same column and factor (phenotypes, testa colour, hilum colour, growing seasons), means followed by the same letter did not differ significantly using LSD at P < 0.05. Where CV is coefficient of variation.
On one hand, non-spotted phenotypes (TGHW and TGHB combined) had smaller seeds (TSW: 71.53 g versus 77.60 g) compared to spotted phenotypes (TSHB and TSHW combined), but lower TPhe (6.55 versus 7.47 mg GAE/g). On the other hand, the phenotypes with black hilum (TGHB and TSHB combined) had lower TTan (3.77 versus 4.54 mg GAE/g) compared to the phenotypes with white hilum (TGHW and TSHW combined).
Growing seasons had a significant effect on all the determined traits. In 2017 growing season, bruchid and endoparasitoid infections (BD, PD, BI), TPhe and TTan were higher, but SpP, seed weight (TSW) and ST were lower compared to 2018 (Table 1).
Testa colour and nutrient concentrations
The four phenotypes differed significantly in testa colour and seed nutrient parameters with the exception of seed PC (Table 2). L* (lightness) was highest in TGHW (36.11), moderate in TGHB and TSHW (32.00 and 32.31, respectively) and lowest in TSHB (30.53). The TSHB phenotype had the highest a* (−0.27), but the lowest b* value (12.31), while the adverse was found for TGHW. The highest carbon concentration was measured in TSHB (44.58%), while the lowest in TSHW (42.11%), which also had the lowest C/N ratio (9.46).
Table 2. Mean comparisons of seed coat colour parameters (L*, a*, b*), protein concentration (PC), seed carbon concentration (C) and carbon to nitrogen ratio (C/N)
For the same column and factor (phenotypes, testa colour, hilum colour, growing seasons), means followed by the same letter did not differ significantly using LSD at P < 0.05. Where CV is coefficient of variation.
No differences were found between non-spotted and spotted phenotypes for seed nutrient composition (Table 2). The non-spotted phenotypes had higher L* (34.27 versus 31.42) and b* values (16.62 versus 13.16), but lower a* (−2.44 versus −0.80) compared to spotted ones.
Hilum colour had no effect on PC and C/N ratio (Table 2). On the other hand, phenotypes with white hilum had higher L* (34.21 versus 31.49) and b* values (16.16 versus 13.62), but lower a* (−2.28 versus −0.95) and carbon concentration (42.51 versus 43.90%).
With the exception of a* and PC, growing seasons affected the testa colour and seed nutrient traits, all of which were higher in 2018 growing season (Table 2).
Relationships between the traits differed significantly among phenotypes
Among the four phenotypes, percentage of bruchid infection (BI) was negatively correlated with seed carbon concentration (r = −0.98, P < 0.05, n = 4) and C/N ratio (r = −0.98, P < 0.05, n = 4) (Fig. 1).
Figure 1. Correlations between percentage (%) of bruchid infection (BI) with seed carbon concentration (a) and C/N ratio (b).
Testa colour parameters (L*, a*, b*) were significantly inter-correlated positively (L*–b*: r = −0.98, P < 0.05, n = 4) or negatively (L*–a*: r = −0.98, P < 0.05, n = 4 and a*–b*: r = −1.00, P < 0.05, n = 4) (Fig. 2).
Figure 2. Inter-correlations between testa colour parameters (L*, a*, b*).
For the phenotypes × growing seasons interaction, a negative correlation was found between total tannin concentration and seed C/N ratio (Fig. 3).
Figure 3. Total tannin concentration in seeds was correlated with C/N ratio for the phenotypes × growing seasons interaction.
Discussion
Pea (P. sativum L.) landraces hold considerable genetic diversity, which could be exploitable via breeding/selection (Keneni et al., Reference Keneni, Jarso, Wolabu and Dino2005; Gixhari et al., Reference Gixhari, Pavelková, Ismaili, Vrapi, Jaupi and Smýkal2014; Hagenblad et al., Reference Hagenblad, Boström, Nygårds and Leino2014). This was confirmed in the present study by the differentiation of the four phenotypes in seed chemical (total phenols, total tannins, C concentration and C/N ratio) and testa colour traits, but more interesting in bruchid infection (BI) susceptibility.
Testa colour parameters (L*, a*, b*) showed significant inter-correlations and in accordance to Zhao et al. (Reference Zhao, Su, Qin, Wang and Kang2020), a positive correlation between L* (lightness) and b* (yellowness) was evident. These researchers reported negative correlations of seed phenolic concentrations with L* and b*, which were not confirmed for the four phenotypes in our work. The findings of Zhao et al. (Reference Zhao, Su, Qin, Wang and Kang2020) were in agreement with previous reports that pea seeds with darker testa had more phenolics and higher antioxidant capacity (Troszyńska and Ciska, Reference Troszyńska and Ciska2002; Xu et al., Reference Xu, Yuan and Chang2007; Marles et al., Reference Marles, Warketin and Bett2013; Jha et al., Reference Jha, Purves, Elessawy, Zhang, Vandenberg and Warkentin2019). However, when the four phenotypes grouped according to testa and hilum colour, our findings were consistent with Zhao et al. (Reference Zhao, Su, Qin, Wang and Kang2020); the larger, spotted seeds contained more phenolics and had lower L* and b* values. On the other hand, the white-hilumed seeds contained more tannins, had higher L* and b* values and higher C concentrations.
Seed weevil is the major insect pest of dry peas with very high levels of infection, especially in cream-coloured peas where infection percentages can reach 98% (Teshome et al., Reference Teshome, Mendesil, Geleta, Andargie, Anderson, Rämert, Hillbur, Dagne and Bryngelsson2015). The two spotted phenotypes had the lowest (44.50%) and the highest levels (66.50%) of BI, with white-hilumed phenotype (TSHW) to be the most susceptible. This finding comes to confirm previous reports for genotypic differences in bruchid tolerance in grain legumes like peas (Teshome et al., Reference Teshome, Mendesil, Geleta, Andargie, Anderson, Rämert, Hillbur, Dagne and Bryngelsson2015; Nikolova, Reference Nikolova2016), vetches (Tsialtas et al., Reference Tsialtas, Irakli and Lazaridou2018; Boulata et al., Reference Boulata, Irakli and Tsialtas2022), faba beans (Seidenglanz and Huňady, Reference Seidenglanz and Huňady2016; Dell'Aglio and Tayeh, Reference Dell'Aglio and Tayeh2023) and red pea (Lathyrus cicera L.) (Tsialtas et al., Reference Tsialtas, Theologidou, Bilias, Irakli and Lazaridou2020). Finding easily assessable phenotypic traits like testa and hilum colour, related to bruchid tolerance, could be a useful and handy tool in respective selection/breeding efforts. However, in our case, such generalization cannot be done since when the data combined over testa or hilum colour, no difference was evident.
Previously, tolerance to bruchids has been associated with small seed size (Fricke and Wright, Reference Fricke and Wright2016; Tsialtas et al., Reference Tsialtas, Irakli and Lazaridou2018), low seed N and P concentrations (Nikolova, Reference Nikolova2016), high ST (Tsialtas et al., Reference Tsialtas, Irakli and Lazaridou2018) and toxic effects of high phenolic or metal (e.g. Fe) concentrations (Barbehenn and Constable, Reference Barbehenn and Constabel2011; Tsialtas et al., Reference Tsialtas, Theologidou, Bilias, Irakli and Lazaridou2020). In our case, the tolerance of the four phenotypes to bruchids was related with higher seed C concentration and C/N ratio meaning that seeds with relatively higher N concentration (lower C/N) were more preferable by bruchids, which was in line with the findings of Nikolova (Reference Nikolova2016). The inverse relationship between seed C/N ratio and total tannins for the phenotypes × growing seasons interaction inferred that seeds with higher tannin concentration were more susceptible to bruchids, which contradicts the inhibitory role of tannins to feeding insects (Barbehenn and Constable, Reference Barbehenn and Constabel2011). However, the synthesis of the costly tannins as a chemical defence means is not always an option for the plants, but it depends on abiotic factors (e.g. soil nutrient and water availability), the growth type of the plant (fast- or slow-growing) and is combined with other chemical defensive substances (Seigler, Reference Seigler2002).
Conclusions
The pea landrace ‘Katsouni’ holds phenotypic variation in testa and hilum colour, which was associated with seed trait variation. Spotted seeds with darker testa had more phenolics, while white-eyed seeds contained more tannins. Seeds with spotted testa and black hilum were less susceptible to bruchids.
Acknowledgements
We thank Mr. Μ. Kovaios, Schinoussa, Greece for providing the seed sample and Professor A. Lazaridou, Aristotle University of Thessaloniki, Faculty of Agriculture, Department of Food Chemistry and Biochemistry, Thermi, Greece, for kindly offering access to instrumentation. We also thank the anonymous reviewers for their constructive comments that contributed to upgrade our initial submission.
Competing interest
None.
Open access
