White Lupin - ACS Publications - American Chemical Society

Jun 17, 2014 - ABSTRACT: White lupin seed can be used for traditional and functional foods or as animal feed. This study aimed to support lupin breede...
0 downloads 0 Views 512KB Size
Article pubs.acs.org/JAFC

Quality of Lupinus albus L. (White Lupin) Seed: Extent of Genotypic and Environmental Effects Paolo Annicchiarico,*,† Patrizia Manunza,‡ Anna Arnoldi,§ and Giovanna Boschin§ †

Research Centre for Fodder Crops and Dairy Productions, Agricultural Research Council, viale Piacenza 29, 26900 Lodi, Italy Research Unit for Agropastoral Systems in Mediterranean Environments, Agricultural Research Council, Podere Ortigara, 09025 Sanluri, Italy § Department of Pharmaceutical Sciences, University of Milan, via Mangiagalli 25, 20133 Milano, Italy ‡

S Supporting Information *

ABSTRACT: White lupin seed can be used for traditional and functional foods or as animal feed. This study aimed to support lupin breeders and production stakeholders by assessing the extent of genotypic, environmental, and genotype × environment (GE) interaction effects on seed contents of oil, tocopherols (TOC), and quinolizidine alkaloids (QA), grain yield, and seed weight of eight elite genotypes grown in two climatically contrasting Italian locations for two cropping years. On average, plants in the subcontinental climate site exhibited higher grain yield and seed size, about 8% lower oil content, and almost 85% higher QA content than those in the Mediterranean climate site. The range of genotype means was 2.97−5.14 t/ha for yield, 92−110 mg/g for oil, and 0.121−0.133 mg/g for TOC. TOC amount was largely unpredictable and featured large GE interactions that hinder its genetic improvement. Oil and alkaloid contents and seed size are more predictable and offer potential for selection. KEYWORDS: genetic variation, genotype × environment interaction, oil, quinolizidine alkaloids, tocopherols



INTRODUCTION Lupin belongs to the Leguminosae (or Fabaceae) family, which includes over 450 species. Only four Lupinus species are cultivated, namely, Lupinus albus (white lupin), Lupinus luteus (yellow lupin), Lupinus angustifolius (narrow-leafed lupin), and Lupinus mutabilis (Andean lupin). L. albus and L. luteus are mainly cultivated in Europe, L. angustifolius is mostly grown in Australia, and L. mutabilis is essentially grown in South America. The annual world production of lupin seeds exceeds 1 million tons, with Australia being the major producer, followed by Chile, the Russian Federation, France, and Poland.1 L. albus is used for producing traditional2 and functional foods,3,4 or as animal feed. Its seed contains high concentrations of proteins (34−45% of dry matter) and essential amino acids.5 It also has 8−12% oil with excellent nutritional characteristics, such as high levels of unsaturated fatty acids and tocopherols (TOC). 6,7 TOC may prevent the risk of cardiovascular and some eye diseases.8,9 They behave as potent antioxidants, because they are able to interrupt the chain reactions that are responsible for the peroxidation of unsaturated lipids by trapping hydroperoxide intermediates.10 White lupin seed is exploited also in diets for celiac and diabetic people, owing to its extremely low content of starch and its composition of proteins.5,11 The main antinutritional factors in lupin seeds are represented by quinolizidine alkaloids (QA).12,13 They are synthesized by lupin plants as a defense mechanism against pathogens and herbivores. The main QA reported for L. albus are lupanine, albine, multiflorine, and 13α-hydroxylupanine, with minor amounts of α-isolupanine, angustifoline, 11,12-seco12,13-didehydromultiflorine (formerly N-methylalbine), and some esters of 13α-hydroxylupanine.12−15 The majority of © 2014 American Chemical Society

acute toxicity studies were performed on lupanine and sparteine (the latter being one of the major QA in L. luteus), and both showed moderate to acute oral toxicity due to neurological effects leading to the loss of motor coordination and muscular control.16 A maximum limit of 200 mg/kg for the total QA content in lupin flours and foods has been fixed by the health authorities of Great Britain, France, Australia, and New Zealand.16−18 Lupin breeders and lupin production stakeholders (seed processing industries, farmers, etc.) can greatly benefit from knowledge on the relative size of genotypic, environmental, and genotype × environment (GE) interaction sources of variation. This information can support plant breeding strategies (e.g., highlighting opportunities and limitations for the genetic improvement of each trait; optimizing the extent and type of multienvironment testing that is needed to select for a given trait) and provide guidance on the relative importance of the plant variety and the growing environment for enhancing positive traits and reducing negative ones. For example, this could lead food industries to prefer seeds harvested in specific agroclimatic regions from specific cultivars, if location and cultivar had an important impact on seed quality traits. The expected year-to-year variation for these characteristics is also relevant for processing and marketing strategies. In an earlier study, one cultivar grown in different Italian environments exhibited lower grain yield, but higher contents of oil and α-linolenic acid, when grown in a Mediterranean Received: Revised: Accepted: Published: 6539

December 13, 2013 March 26, 2014 June 17, 2014 June 17, 2014 dx.doi.org/10.1021/jf405615k | J. Agric. Food Chem. 2014, 62, 6539−6545

Journal of Agricultural and Food Chemistry

Article

Figure 1. Total ion chromatogram of GC-MS analyses and QA chemical structure; 1, albine; 2, angustifoline; 3, α-isolupanine; 4, lupanine; 5, 11,12seco-12,13-didehydromultiflorine (formerly N-methylalbine); 6, multiflorine; 7, 13α-hydroxylupanine; 8, 13α-angeloyloxylupanine; 9, 13αtigloyloxylupanine.



climate site compared with a subcontinental climate location (i.e., a location where autumn-sown crops undergo severe lowtemperature stress during winter).19 A second study, including some varieties and one ecotype grown in two climatically contrasting Italian sites, revealed large genotypic variation for various seed quality traits, along with large genotype × environment interaction variation for grain yield.7 Subcontinental climate conditions determined higher QA content than Mediterranean conditions, leading various low-alkaloid varieties to exceed the limit of 200 mg/kg.20 The objective of this study is providing a thorough assessment of the relative contribution of genotypic, environmental, and genotype × environment interaction effects to variation in key production and quality traits of white lupin seeds. Our study included eight elite, phenotypically contrasting genotypes, which were grown in two climatically contrasting Italian locations for two cropping years, and we assessed grain yield, seed weight, and the content of oil, TOC, and QA.

MATERIALS AND METHODS

Chemicals. Hexane, dichloromethane, and 2-propanol were purchased from Baker (Deventer, The Netherlands); 36% hydrochloric acid was from Merck (Darmstadt, Germany); 25% NH4OH was from Carlo Erba (Rodano, Italy); and standards of tocopherols were from Calbiochem (EMD Biosciences, Merck, Darmstadt, Germany). Plant Material and Growing Environments. The test white lupin genotypes were (i) Multitalia, which is the only commercial variety of white lupin registered in Italy; (ii) seven advanced breeding lines produced by a joint selection program of CRA and JouffrayDrillaud, coded hereafter as P2, P3, 4-3, 7-44, 7-7, 8-28, and 8-483. These lines were selected from a large number of breeding lines on the ground of wide or specific adaptation to subcontinental and Mediterranean climate locations of Italy, with the further objective of ensuring the presence of large morphophysiological variation within the set of tested material.21 Modern white lupin varieties can be classified according to their phenological characteristics as spring (S) or autumn (A) types (which recall their suitability for spring or autumn sowing in cool temperate regions such as France); as tall (t), semidwarf (sd), or dwarf (d) types, according to genes affecting internode length and plant stature; and as indeterminate (I) or 6540

dx.doi.org/10.1021/jf405615k | J. Agric. Food Chem. 2014, 62, 6539−6545

Journal of Agricultural and Food Chemistry

Article

30 m × 0.25 mm i.d., 0.25 μm (Supelco, Milan, Italy). The temperature program was as follows: 150 °C for 5 min, from 150 to 300 °C at 5 °C/min, and then 300 °C for 15 min. Analyses were performed in split mode (split ratio 1:25), the injection volume was 1 μL, the injection temperature was 250 °C, the interface temperature was 300 °C, and the acquisition was in the range 50−450 m/z. The source operated in electron ionization (EI) mode. Each analysis was repeated three times. The identification and quantification of QA was performed as previously reported.20,24 The total content, expressed as milligrams per gram of flour, is the sum of nine alkaloids (reported here in elution order): albine, angustifoline, α-isolupanine, lupanine, 11,12-seco-12,13-didehydromultiflorine (formerly N-methylalbine), multiflorine, 13α-hydroxylupanine, 13α-angeloyloxylupanine, and 13α-tigloyloxylupanine. Figure 1 reports a typical GC-MS chromatogram and chemical structures of QA. Each sample was independently analyzed three times. In Supporting Information, the main analytical features of the detected QA (Table S1) and composition of single QA (Table S2) are available. Statistical Analysis. Chemical analysis replications belonging to same field plot were averaged prior to statistical analyses, which were performed on the three field plot replications available for each genotype−environment combination (considered more representative of treatment variation due to environmental effects than chemical replications). An analysis of variance (ANOVA) including the fixed factors genotype, location and year and the random factor experiment block was performed on plot values of grain yield, individual seed weight, oil content, TOC content, and QA content, to assess the significance of all genotypic, environmental, and GE interaction sources of variation. A second ANOVA including the fixed factors genotype and cropping environment (as defined from the four location−year combinations) and the random factor block aimed to compare mean values of genotypes and environments using Duncan’s multiple range test at p < 0.05. The size of genotypic, environmental, and GE interaction effects was assessed by estimating the respective ANOVA components of variance through a restricted maximum likelihood method, assuming in this case that genotypes and environments were a random sample of the genetic and environmental variation relevant to a breeding program. ANOVAs were also carried out for each trait in each experiment, reporting least significance difference (LSD) values for genotype mean comparison at p < 0.05. All analyses were performed with Statistical Analysis System (SAS) software (1999).

determinate (D) types, depending on the presence or absence of genetically determined ability to flower on secondary and higher-order branches.22,23 The breeding lines exhibited the following combinations of traits: (i) lines P2 and P3, S/t/I; (ii) lines 4-3 and 7-44, A/sd/I; (iii) line 7-7, A/t/I; (iv) lines 8-28 and 8-483, A/d/D. Multitalia is t/I/ phenologically intermediate between S and A. The eight genotypes were grown in Lodi (Lombardy, Italy, 45° 19′ N, 9° 03′ E) and Sanluri (Sardinia, Italy, 39° 30′ N, 8° 50′ E) in rainfed autumn-sown field experiments, which were repeated in the seasons 2006−2007 and 2007−2008. Lodi has a subcontinental climate with cold winters and more favorable rainfall pattern than Sanluri, which has a Mediterranean climate (Table 1). The soils of both locations, featuring pH < 7.5 and active lime below 0.4%, were favorable for lupin cropping. Each experiment was designed as a randomized complete block with three replications. Each plot included 54 sown seeds arranged on three rows. Before sowing, the seeds were treated with Germipro UFB (BASF Agro, Ecully, France; 350 g/L iprodione + 177 g/L carbendazim) at the rate of 2.6 mL/kg, to limit possible damage by anthracnose (Colletotrichum gloeosporioides). They were inoculated with NPPL HiStick (Becker Underwood, Toulouse, France), to ensure proper nodulation and N fixation by Bradyrhizobium strains. The experiment in Lodi was fertilized with 40 kg/ha N and 120 kg/ha P2O5 and K2O, whereas soil analyses suggested no fertilization requirement for Sanluri. Chemical weed control was performed in both sites by applying 4 L/ha Stomp 330 (BASF Agro; pendimentalin 32%). Individual plots were harvested individually when they reached crop maturity, which occurred in the first two weeks of July in Lodi and the second and third week of June in Sanluri. Grain yield was expressed at 13% seed moisture, after moisture determination on a random sample of 200 seeds per field replicate. These seed samples were also used for assessing the individual seed weight, expressed on a dry matter basis. Another random sample of 20 g of seed/replicate was used for assessing oil, TOC, and QA contents. Extraction of Oil. Impurities were removed from the seeds, and the cleaned seeds were hand-dehulled, ground in a household mill (Braun, Germany), and sieved through a 60-mesh sieve. The oil was extracted with hexane for 6 h with a Soxhlet apparatus using cellulose extraction thimbles (123 mm × 43 mm i.d.; Whatman International, Brentford, U.K.); its content was gravimetrically determined and expressed as milligrams per gram of flour.7,19 Each sample was independently extracted three times. Extraction and Determination of Tocopherols. Tocopherols (TOC) were determined as previously reported.6 Briefly, the oil was diluted in hexane, filtered through a 0.45 μm filter (Grace, Milan, Italy), and analyzed by a 1100 series HPLC (Agilent Technologies, Santa Clara, CA) equipped with an autosampler and a fluorometric detector (λassn = 295 nm, λexc = 290 nm, λem = 330 nm, frequency 55 Hz, gain 10). The analyses were performed on a Lichrosorb Si60 column with n-hexane/2-propanol (99:1 v/v) as mobile phase; the injection volume was 10 μL and the flow rate was 1 mL/min. The relative amounts of TOC were calculated by comparing their peak areas with standard curves, as previously reported.6 The total TOC content was calculated by summing up the values of each detected isomer. TOC concentration was expressed as milligrams per gram of flour. Each sample was independently analyzed three times. Extraction and Determination of Quinolizidine Alkaloids. Quinolizidine alkaloid (QA) extraction was performed as previously reported.20,24 Briefly, 1 g of defatted flour was suspended in 8 mL of 0.1 N HCl and stirred at room temperature for 17 h; the mixture was centrifuged at 6000g for 50 min at 4 °C, the supernatant was collected, and the solid was washed twice with 5 mL of 0.1 N HCl. The gathered extracts were alkalinized with 5% NH4OH to pH 10−11 and then applied onto an Extrelut NT 20 column (Merck, Darmstadt, Germany). The alkaloids were eluted with CH2Cl2 (4 × 20 mL), and the solvent was evaporated to dryness under vacuum. The residue was diluted in dichloromethane and analyzed by gas chromatography− mass spectrometry (GC-MS). Each sample was independently extracted at least three times. The GC-MS analyses were performed on a Shimadzu QP-5000 GC-MS instrument equipped with an AOC20i autosampler (Shimadzu) using an AT-1 ms capillary column



RESULTS

Climatic information for test years and the long term at the test locations, particularly that with a bearing on the extent of winter cold stress and terminal drought stress, is given in Table 1. As expected, the subcontinental climate site (Lodi) was characterized by colder winters (in terms of number of frost days and absolute minimum temperature) and more rainy springs than the Mediterranean climate site (Sanluri). However, Table 1. Climatic Characteristics of Test Locations of Lupinus albus Genotypes in the Cropping Years and the Long Term frost days

period

winter abs min temp (°C)

mean daily temp October−June (°C)

Lodi (Subcontinental Climate) 18 −2.5 11.0 38 −6.8 11.0 56 −7.7 9.3 Sanluri (Mediterranean Climate) season 2006−2007 0 0.5 14.6 season 2007−2008 10 −4.0 13.8 long-term value 2 0.1 15.0 season 2006−2007 season 2007−2008 long-term value

6541

rainfall April−June (mm) 202 239 218 151 117 88

dx.doi.org/10.1021/jf405615k | J. Agric. Food Chem. 2014, 62, 6539−6545

Journal of Agricultural and Food Chemistry

Article

Table 2. Mean Values of Grain Yield and Seed Quality Traits for Eight Lupinus albus Genotypes Grown in Four Environmentsa Lodi

Sanluri

genotype

2006−2007

2007−2008

Multitalia P2 P3 4-3 7-44 7-7 8-28 8-483 meanb LSD

8.106 7.563 9.779 5.917 7.715 6.318 5.969 4.314 6.960 a 3.283

6.850 3.583 3.875 6.368 6.828 6.342 3.639 3.766 5.156 b 1.092

Multitalia P2 P3 4-3 7-44 7-7 8-28 8-483 meanb LSD

0.479 0.354 0.349 0.380 0.358 0.381 0.317 0.355 0.372 a 0.027

2006−2007

2007−2008

meanb

Grain Yield (t/ha)

Multitalia P2 P3 4-3 7-44 7-7 8-28 8-483 meanb LSD

117.2 107.5 98.5 92.5 123.0 97.4 88.6 94.0 102.3 ab 6.9

Multitalia P2 P3 4-3 7-44 7-7 8-28 8-483 meanb LSD

0.1180 0.1400 0.1110 0.1360 0.1160 0.1260 0.1130 0.1240 0.1230 bc 0.0117

Multitalia P2 P3 4-3 7-44 7-7 8-28 8-483 meanb LSD

6.392 0.090 0.169 0.329 0.320 0.614 0.175 0.623 1.089 a 0.193

1.764 2.322 4.264 1.211 2.136 2.099 1.225 2.475 2.187 c 1.971 Seed Weight (mg) 0.374 0.351 0.283 0.257 0.271 0.342 0.301 0.286 0.256 0.304 0.322 0.269 0.234 0.269 0.257 0.248 0.287 c 0.291 bc 0.023 0.042 Oil Content (mg/g) 102.1 101.5 103.9 110.2 88.7 109.6 92.3 99.2 92.2 116.5 109.1 110.7 86.1 97.6 87.2 94.7 95.2 b 105.0 ab 13.8 6.2 Tocopherol Content (mg/g) 0.1200 0.1380 0.1240 0.1230 0.1280 0.1380 0.1210 0.1490 0.1520 0.1340 0.1150 0.1450 0.1210 0.1340 0.1320 0.1530 0.1266 b 0.1393 a 0.0172 0.0129 Alkaloid Content (mg/g) 4.820 2.448 0.228 0.121 0.159 0.047 0.291 0.082 0.275 0.108 0.441 0.240 0.300 0.117 0.323 0.146 0.855 b 0.414 d 0.052 0.196

2.760 2.560 2.650 1.777 1.840 1.700 1.230 1.317 1.979 c 0.649

4.870 4.007 5.142 3.818 4.630 4.115 3.016 2.968 4.071 0.941

ab b a b ab b c c

0.361 0.315 0.336 0.273 0.319 0.309 0.301 0.260 0.309 b 0.039

0.391 0.302 0.325 0.310 0.309 0.320 0.280 0.280 0.315 0.015

a c b bc bc b d d

113.3 118.2 112.3 107.8 105.4 114.7 97.5 103.2 109.1 a 10.1

108.5 110.0 102.3 98.0 109.3 108.0 92.4 94.8 102.9 4.2

a a b c a a d cd

0.1070 0.1250 0.1180 0.1170 0.1160 0.1120 0.1160 0.1220 0.1166 c 0.0099

0.1208 0.1280 0.1238 0.1308 0.1295 0.1245 0.1210 0.1328 0.1264 0.0057

3.467 0.130 0.092 0.181 0.225 0.339 0.310 0.389 0.642 c 0.136

4.282 0.142 0.117 0.221 0.232 0.409 0.226 0.370 0.750 0.067

d abc cd ab abc bcd d a

a d d c c b c b

a

Environments were formed by the combination of one subcontinental (Lodi) or Mediterranean (Sanluri) climate location by each of two cropping years (2006−2007 and 2007−2008). Least significant difference (LSD) at p < 0.05 for genotype mean comparison within and across environments is also shown. bEnvironment row means, or genotype column means across environments, that are followed by different letters differ at p < 0.05 according to Duncan’s test. 6542

dx.doi.org/10.1021/jf405615k | J. Agric. Food Chem. 2014, 62, 6539−6545

Journal of Agricultural and Food Chemistry

Article

milder winters and higher mean temperatures than the long term in Lodi (particularly in the first year), and wetter springs (particularly in the first year) and lower winter and mean temperatures (in the second year) than the long term in Sanluri, made the two locations less climatically contrasting over the test years than in the long term (Table 1). Mean values of grain yield and seed quality traits for the eight lupin genotypes grown in the four environments formed by the combination of two climatically contrasting sites (Lodi or Sanluri) and two cropping years (2006−2007 and 2007−2008) are reported in Table 2, whereas ANOVA results for genotype, location and year factors and their interactions are given in Table 3. Significant variation among genotypes was detected

Table 4. Estimates of Genotypic and Environmental Components of Variance for Grain Yield and Grain Quality Traits of Eight Lupinus albus Genotypesa

Table 3. Analysis of Variance F-Test Results for Grain Yield and Grain Quality Traits of Eight Lupinus albus Genotypesa

variance component relative to the genotypic and environmental components of variance). Oil content variation was mainly affected by genotypic differences, which were partly inconsistent across environments (as indicated by comparable genotype and GE interaction variance). Purely genetic variation for TOC content had negligible size (although being statistically significant in Table 3) when compared with the outstanding variance of cropping environment and GE interaction (Table 4). Finally, QA content variation in this data set was largely affected by the genotype (Table 4), as further discussed later on. Grain yield responses of the cultivars were heavily influenced by their phenological characteristics, as highlighted in greater detail in an analysis of adaptation that is reported elsewhere.21 Spring types (lines P2 and P3) were top-yielding in the Mediterranean site, performed fairly well in the mild winter year of the subcontinental climate site, and yielded very poorly in the cold winter year of the same site (Table 2) because of severe winter mortality. Autumn types, including the remaining breeding lines, displayed the opposite adaptation pattern, whereas Multitalia (phenologically intermediate) was moderately adapted to any test environment (Table 2). Multitalia also featured larger seed than any breeding line (Table 2). The genetic variation for mean value of oil content was sizable (range of genotype means across environments 9.2− 11%), whereas that for TOC was modest (range of genotype means across environments: 0.121−0.133 mg/g) (Table 2). In both cases, they were unrelated to phenological types (Table 2). In absolute terms, the highest recorded values for these quality traits were 12.3% for oil content (line 7-44 in the first cropping year at Lodi) and 0.153 mg/g for TOC content (line 8-483 in the first year at Sanluri) (Table 2). Multitalia exhibited extremely high QA content, exceeding by at least 12-fold the limit of 200 mg/kg even in the most suitable environment for producing low-QA lupins (Table 2). All breeding lines except for line P3 exceeded the 200 mg/kg limit in at least one test environment (Table 2).

source of variation genotype location year genotype × location genotype × year location × year genotype × location × year

grain yield

seed weight

oil content

total TOC

total QA

** ** * ** ** * *

** ** ** ** * ** NS

** * NS ** ** NS *

** NS * ** ** ** **

** ** NS ** ** ** **

source of variation

grain yield (t/ ha)2

seed weight (mg)2

oil content (mg/g)2

total TOC (mg/g)2 × 1000

total QA (mg/g)2

genotype environment genotype × environment

0.263 5.608 0.937

0.00106 0.00143 0.00051

0.384 0.211 0.366

0.0039 0.0759 0.0788

1.963 0.046 0.334

a

Environments are combinations of two climatically contrasting locations by two cropping years. All variance components different from zero at p < 0.01.

a

Grown in one subcontinental and one Mediterranean climate location for two cropping years. **Significant at p < 0.01; *significant at p < 0.05; NS, not significant.

both across and within environments for all traits. Genotype responses, however, interacted significantly with location and year in all cases (Table 3). Averaged across genotypes, the lupin crop in the subcontinental climate site exhibited, in comparison with the Mediterranean location, almost 3-fold higher grain yield, larger seed (actually only in the first year), about 8% lower oil content, and definitely higher QA content (Table 2). The grain yield advantage of the former site over the latter was about 2fold when the genotypes capable of maximizing the crop yield in each location are considered, that is, 7-44 in Lodi (site mean yield 7.27 t/ha), and P3 in Sanluri (site mean yield 3.46 t/ha) (Table 2). The mean crop response across environments tended to be more consistent and predictable for oil content (location factor was significant at p < 0.05; year, and location × year interaction were not significant) than for the other variables (Table 3). In contrast, site variation for TOC was inconsistent across years due to large location × year interaction (Table 3), since the Mediterranean location exhibited both the highest (first year) and the lowest (second year) environment mean value for this variable (Table 2). The relative size of genotypic, environmental and GE interaction components of variance differed depending on the trait (Table 4). Grain yield variation was outstandingly affected by the cropping environment on the basis of the much greater size of the environmental component of variance relative to the genotypic and GE interaction variance components. Genotype yield responses were more affected by GE interaction than by purely genetic effects (as indicated by the almost 4-fold greater size of GE interaction variance relative to genotypic variance). Seed weight was affected by the environment and, to a lesser extent, by the genotype, in a fairly consistent manner (as indicated by the relatively smaller size of the GE interaction



DISCUSSION Our results can contribute to choices and strategies of product supply for seed lupin processing industries, as well as to variety selection strategies, for production areas of southern Europe. This region includes Mediterranean climate areas (represented here by Sanluri) that are widespread in central and southern Italy and in coastal areas of the Iberian and Balkan peninsulas and southern France, as well as subcontinental climate areas (represented by Lodi) placed in northern Italy and in inland areas of the Iberian and Balkan peninsulas and southwestern France. However, conclusive indications for these climatic 6543

dx.doi.org/10.1021/jf405615k | J. Agric. Food Chem. 2014, 62, 6539−6545

Journal of Agricultural and Food Chemistry

Article

rainfall were reported for narrow-leafed lupin in Australia.29 The outstanding QA seed content of Multitalia is probably due to extensive contamination, which may have arisen at first from pollen of neighboring bitter-seed plants and then built up in the cultivar stock across the many cycles of multiplication of this 25-year old cultivar. The average QA value of Multitalia in Lodi (5.6 mg/g) may imply over 30% bitter seeds, considering that the most bitter-seed Italian ecotypes displayed around 15 mg/g QA in the same location.20 A contamination of 3% bitter seeds may lead to 25% bitter seeds after eight generations of seed multiplication, owing to the advantage of high QA plants under natural selection.22 Although much lower, the QA content of most current breeding lines occasionally exceeded the limit of 200 mg/kg, probably because of slight contamination that arose at some stage of the selection process. Tocopherols exist in four different congeners (vitamers) called α, β, γ, and δ. The most abundant are α, which is the main contributor to vitamin E activity, and γ, which shows the highest antioxidant activity in foods.10 White lupin seeds contains α-, γ-, and δ-TOC (unlike narrow-leafed lupin, which lacks the third one).6 The occurrence of fairly wide variation for TOC values had already been reported.34,35 While confirming that (absolute range 0.107−0.153 mg/g; Table 2), our study also shows that this variation is substantially unpredictable and not exploitable, being largely inconsistent both across locations in different years and across genotypes in different environments. Large seed is important essentially for direct lupin utilization as a traditional food. No tested material featured seed large enough for that use. Our results suggest, however, that genetically based differences for seed weight are largely repeatable across environments and, hence, selectable through breeding. The environmental effect is larger than the genetic one but is partly predictable, as somewhat larger seed can be obtained from more favorable growing environments. In conclusion, the different white lupin traits object of our study offer different challenges and opportunities for breeders and crop production stakeholders. Some traits, such as oil content, alkaloid content, and to a lesser extent seed size, are affected by the agroclimatic area in a fairly consistent manner across cropping years and could be selected for on the basis of the relative size of purely genetic effects. The reverse is true for the total content of TOC, despite its fairly wide range of variation across genotypes and cropping environments. Grain yield is largely affected by the cropping site, and it would require the selection of different cultivars for climatically contrasting regions.

regions, especially for areas outside Italy, would require verification from experiments in a larger sample of test environments. Assessing also site crop production, as in the current study, has great practical importance for the seed market, particularly with respect to plant material with locally highest yielding ability and in case a trade-off existed between grain yield and quality. The Mediterranean climate growing site offered two seed quality advantages, namely, consistently higher oil content and much lower risk of exceeding the safety QA level of 200 mg/kg, compared with the subcontinental climate site. These results confirm earlier findings relative to other white lupin cultivars.7,19,20 These advantages may justify the attribution of somewhat higher market price to seeds produced in Mediterranean environments, which could partly counterbalance the much lower grain yield of these environments relative to subcontinental climate ones. Besides its use as a traditional or functional food, white lupin could be used as a dual-purpose crop for oil extraction and for use of the residual from extraction as a high-protein feed (as ordinarily done for soybean). This option would be reinforced by the excellent food quality of its oil.6,7 This study confirmed that Mediterranean environments are more suitable than subcontinental climate ones to produce oil-rich seeds. However, the oil content of locally adapted material (lines P3 and P2 and Multitalia) never exceeded 12% (Table 2). Although their oil contents were higher than those reported for other white lupin material in the United States (highest value 8.2%)25,26 and Turkey (highest value 10.7%),27,28 they might be too low to justify the oil extraction. Higher oil content (up to 13.6%) was previously observed in another Mediterranean climate Italian site, which featured, however, extremely low grain yield.19 On the whole, these results emphasize the potential interest of screening large germplasm collections with the aim to identify material with outstanding oil content. The selection from such material of improved varieties is encouraged by the current finding that about half the genetic variation for oil content is not subject to GE interaction and can, therefore, be expressed consistently across cropping environments. This finding agrees with previous results for different cultivars of white lupin,7 as well as with results for narrow-leafed lupin (L. angustifolius) in Australia.29 This study confirms the large GE interaction for grain yield across south-European environments that had already emerged for other sweet-seeded cultivars30 and for a world collection of ecotypes of white lupin.31 This GE interaction depends to a large extent on the site-specific importance of flowering time, as early material is more susceptible to winter frost but more tolerant to terminal drought than late material, and vice versa.31 More generally, white lupin breeding for European regions needs to cope with particularly large specific-adaptation effects.32 The amount of QA in white lupin cultivars is genetically controlled by one major recessive gene and various minor genes.22 The current content of QA referred to the sum of albine, angustifoline, multiflorine, 11,12-seco-12,13-didehydromultiflorine, lupanine, α-isolupanine, 13α-hydroxylupanine, 13α-angeloyloxylupanine, and 13α-tigloyloxylupanine. The lower QA content exhibited by seed produced in the Mediterranean environments agrees with the finding by Christiansen et al.33 that terminal drought stress tends to decrease the white lupin seed content of QA. Large environmental effects that were substantially unrelated to



ASSOCIATED CONTENT

* Supporting Information S

Two tables showing main analytical features of the detected QA and composition of single QA in white lupin seeds. This material is available free of charge via the Internet at http:// pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Telephone +39-0371-404751; fax +39-0371-31853; e-mail [email protected]. 6544

dx.doi.org/10.1021/jf405615k | J. Agric. Food Chem. 2014, 62, 6539−6545

Journal of Agricultural and Food Chemistry

Article

Funding

(15) Erbas, M. The effects of different debittering methods on the production of lupin bean snack from bitter Lupinus albus L. seeds. J. Food Qual. 2010, 33, 742−757. (16) Lupin Alkaloids in Food. A Toxicological Review and Risk Assessment; Technical Report Series 3; Australia New Zealand Food Authority: Canberra, Australia, 2001; pp 1−21. (17) Advisory Committee on Novel Foods and Processes. ACNFP Report on Seeds from Narrow Leafed Lupin, Appendix IX; MAFF Publications: London, 1996; p 107. (18) CSHPF. Bulletin Officiel 98/27; Conseil Superieur d’Hygiene Publique de France: Paris, 1998. (19) Boschin, G.; D’Agostina, A.; Annicchiarico, P.; Arnoldi, A. The fatty acid composition of the oil from Lupinus albus cv. Luxe as affected by environmental and agricultural factors. Eur. Food Res. Technol. 2007, 225, 769−776. (20) Boschin, G.; Annicchiarico, P.; Resta, D.; D’Agostina, A.; Arnoldi, A. Quinolizidine alkaloids in seeds of lupin genotypes of different origins. J. Agric. Food Chem. 2008, 56, 3657−3663. (21) Annicchiarico, P.; Manunza, P.; Proietti, S. White lupin tolerance to winter cold, terminal drought and soil lime: patterns of genetic variation and their exploitation in breeding for Southern Europe. In Lupin crops: An Opportunity for Today, a Promise for the Future; Naganowska, B., Kachlicki, P., Wolko, B., Eds.; Proceedings of the XIII International Lupin Conference, International Lupin Association, Canterbury, New Zealand, 2011; pp 99−103. (22) Papineau, J.; Huyghe, C. Le Lupin doux Protéagineux. Editions France Agricole: Paris, 2004. (23) Huyghe, C. White lupin (Lupinus albus L.). Field Crops Res. 1997, 53, 147−160. (24) Resta, D.; Boschin, G.; D’Agostina, A.; Arnoldi, A. Evaluation of total quinolizidine alkaloids content in lupin flours, lupin-based ingredients, and foods. Mol. Nutr. Food Res. 2008, 52, 490−495. (25) Bhardwaj, H. L.; Hamama, A. A.; Merrick, L. C. Genotypic and environmental effects on lupin seed composition. Plant Foods Hum. Nutr. 1998, 53, 1−13. (26) Bhardwaj, H. L.; Hamama, A. A.; van Santen, E. Fatty acids and oil content in white lupin seed as affected by production practices. J. Am. Oil Chem. Soc. 2004, 81, 1035−1038. (27) Erbas, M.; Certel, M.; Uslu, M. K. Some chemical properties of white lupin seeds (Lupinus albus L.). Food Chem. 2004, 89, 341−345. (28) Uzun, B.; Arslan, C.; Karhan, M.; Toker, C. Fat and fatty acids of white lupin (Lupinus albus L.) in comparison to sesame (Sesamum indicum L.). Food Chem. 2007, 102, 45−49. (29) Cowling, W. A.; Tarr, A. Effect of genotype and environment on seed quality in sweet narrow-leafed lupin (Lupinus angustifolius L.). Aust. J. Agric. Res. 2004, 55, 745−751. (30) Annicchiarico, P.; Carroni, A. M. Diversity of white and narrowleafed lupin genotype adaptive response across climatically-contrasting Italian environments and implications for selection. Euphytica 2009, 166, 71−81. (31) Annicchiarico, P.; Harzic, N.; Carroni, A. M. Adaptation, diversity, and exploitation of global white lupin (Lupinus albus L.) landrace genetic resources. Field Crops Res. 2010, 119, 114−124. (32) Annicchiarico, P.; Harzic, N.; Huyghe, C.; Carroni, A. M. Ecological classification of white lupin landrace genetic resources. Euphytica 2011, 180, 17−25. (33) Christiansen, J. L.; Jornsgard, B.; Buskov, S.; Olsen, C. E. Effect of drought stress on content and composition of seed alkaloids in narrow-leafed lupine, Lupinus angustifolius L. Eur. J. Agron. 1997, 7, 307−314. (34) Martinez-Villaluenga, C.; Frias, J.; Vidal-Valverde, C. Functional lupin seeds (Lupinus albus L. and Lupinus luteus L.) after extraction of α-galactosides. Food Chem. 2006, 98, 291−299. (35) Frias, J.; Miranda, M. L.; Doblado, R.; Vidal-Valverde, C. Effect of germination and fermentation on the antioxidant vitamin content and antioxidant capacity of Lupinus albus L. var. Multolupa. Food Chem. 2005, 92, 211−220.

The field experiments were carried out within the project Plant Genetic Resources/FAO Treaty, funded by the Ministry of Agricultural, Food and Forestry Policies of Italy. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We are grateful to Nathalie Harzic for the initial multiplication of plant material; to Antonio Carroni and Sandro Proietti for scientific and technical contributions, respectively, to the field trials; to Marta Caglioni, Cristina Rutigliano, and Laura Vaccaro for valuable help in laboratory work; and to Donatella Resta for precious discussions.



ABBREVIATIONS USED GE, genotype × environment; MW, molecular weight; QA, quinolizidine alkaloids; TOC, tocopherols



REFERENCES

(1) About lupin, global production & trade. http://www.lupins.org/ lupins (Accessed 25 March 2014). (2) D’Agostina, A.; Antonioni, C.; Resta, D.; Arnoldi, A.; Bez, J.; Knauf, U.; Waesche, A. Optimization of a pilot-scale process for producing lupin protein isolates with valuable technological properties and minimum thermal damage. J. Agric. Food Chem. 2006, 54, 92−98. (3) Sirtori, C. R.; Triolo, M.; Bosisio, R.; Bondioli, A.; Calabresi, L.; De Vergori, V.; Gomaraschi, M.; Mombelli, G.; Pazzucconi, F.; Zacherl, C.; Arnoldi, A. Hypocholesterolaemic effects of lupin protein and pea protein/fibre combinations in moderately hypercholesterolaemic individuals. Br. J. Nutr. 2012, 107, 1176−1183. (4) Boschin, G.; Scigliuolo, G. M.; Resta, D.; Arnoldi, A. Optimization of the enzymatic hydrolysis of lupin (Lupinus) proteins for producing ACE-inhibitory peptides. J. Agric. Food Chem. 2014, 62, 1846−1851. (5) Sujak, A.; Kotlarz, A.; Strobel, W. Compositional and nutritional evaluation of several lupin seeds. Food Chem. 2006, 98, 711−719. (6) Boschin, G.; Arnoldi, A. Legumes are valuable sources of tocopherols. Food Chem. 2011, 127, 1199−1203. (7) Boschin, G.; D’Agostina, A.; Annicchiarico, P.; Arnoldi, A. Effect of genotype and environment on fatty acid composition of Lupinus albus L. seed. Food Chem. 2008, 108, 600−606. (8) Dillard, C. J.; German, J. B. Phytochemicals: Nutraceuticals and human health. J. Sci. Food Agric. 2000, 80, 1744−1756. (9) Cho, Y.-S.; Yeum, K.-J.; Chen, C.-Y.; Beretta, G.; Tang, G.; Krinsky, N. I.; Yoon, S.; Lee-Kim, Y. C.; Blumberg, J. B.; Russell, R. M. Phytonutrients affecting hydrophilic and lipophilic antioxidant activities in fruits, vegetables and legumes. J. Sci. Food Agric. 2007, 87, 1096−1107. (10) Bramley, P. M.; Elmadfa, I.; Kafatos, A.; Kelly, F. J.; Manios, Y.; Roxborough, H. E.; Schuch, W.; Sheehy, P. J. A.; Wagner, K. H. Vitamin E. J. Sci. Food Agric. 2000, 80, 913−938. (11) Duranti, M. Grain legume proteins and nutraceutical properties. Fitoterapia 2006, 77, 67−82. (12) Wink, M.; Meissner, C.; Witte, L. Patterns of quinolizidine alkaloids in 56 species of the genus Lupinus. Phytochemistry 1995, 38, 139−153. (13) Boschin, G.; Resta, D., Alkaloids derived from lysine: Quinolizidine, a focus on lupin alkaloids. In Natural Products: Phytochemistry, Botany and Metabolism of Alkaloids, Phenolics and Terpenes; Ramawat, K. G., Merillon, J. M., Eds.; Springer: Berlin and Heidelberg, Germany, 2013; pp 381−403. (14) Muzquiz, M.; Cuadrado, C.; Ayet, G.; de la Cuadra, C.; Burbano, C.; Osagie, A. Variation of alkaloid components of lupin seeds in 49 genotypes of Lupinus albus from different countries and locations. J. Agric. Food Chem. 1994, 42, 1447−1450. 6545

dx.doi.org/10.1021/jf405615k | J. Agric. Food Chem. 2014, 62, 6539−6545