Article pubs.acs.org/jpr
Two Traditional Maize Inbred Lines of Contrasting Technological Abilities Are Discriminated by the Seed Flour Proteome Carla Pinheiro,*,† Kjell Sergeant,‡ Cátia M. Machado,† Jenny Renaut,‡ and Cândido P. Ricardo† †
Instituto de Tecnologia Química e Biológica, Universidade Nova de Lisboa, Av. da República-EAN, 2780-157 Oeiras, Portugal Department “Environment and Agro-biotechnologies” (EVA), Centre de Recherche Public-Gabriel Lippmann, 41, rue du Brill, 4422 Belvaux, Luxembourg
‡
S Supporting Information *
ABSTRACT: The seed proteome of two traditional maize inbred lines (pb269 and pb369) contrasting in grain hardness and in preferable use for bread-making was evaluated. The pb269 seeds, of flint type (i.e., hard endosperm), are preferably used by manufacturers, while pb369 (dent, soft endosperm) is rejected. The hypothesis that the content and relative amounts of specific proteins in the maize flour are relevant for such discrimination of the inbred lines was tested. The flour proteins were sequentially extracted following the Osborne fractionation (selective solubilization), and the four Osborne fractions were submitted to twodimensional electrophoresis (2DE). The total amount of protein extracted from the seeds was not significantly different, but pb369 flour exhibited significantly higher proportions of salt-extracted proteins (globulins) and ethanol-extracted proteins (alcohol-soluble prolamins). The proteome analysis allowed discrimination between the two inbred lines, with pb269 demonstrating higher heterogeneity than pb369. From the 967 spots (358 common to both lines, 208 specific to pb269, and 401 specific to pb369), 588 were submitted to mass spectrometry (MS). Through the combined use of trypsin and chymotrypsin it was possible to identify proteins in 436 spots. The functional categorization in combination with multivariate analysis highlighted the most discriminant biological processes (carbohydrate metabolic process, response to stress, chitin catabolic process, oxidation−reduction process) and molecular function (nutrient reservoir activity). The inbred lines exhibited quantitative and qualitative differences in these categories. Differences were also revealed in the amounts, proportions, and distribution of several groups of storage proteins, which can have an impact on the organization of the protein body and endosperm hardness. For some proteins (granule-bound starch synthase-1, cyclophilin, zeamatin), a change in the protein solubility rather than in the total amount extracted was observed, which reveals distinct in vivo associations and/or changes in binding strength between the inbred lines. Our approach produced information that relates protein content, relative protein content, and specific protein types to endosperm hardness and to the preferable use for “broa” bread-making. KEYWORDS: protein polymorphism, Osborne fractionation, maize inbred lines, seed-storage proteins, trypsin, chymotrypsin, semitrypsin
■
for its distinctive sensory characteristics.5 Most of the Portuguese traditional maize varieties (over 95%) are of the flint type, selected over centuries to fit the farmer’s food quality standards. In the process of ‘‘broa’’ manufacturing, the use of the high-yielding dent hybrids designed for feed has been rejected.6 Although flint and dent types are easily distinguishable (in flint types a hard layer encloses the soft endosperm, resulting in a harder endosperm7), the quality properties supporting the choice are unknown. Significant differences in maize dough rheology were detected between the regional flint types and the dent maize varieties.8 Taking into consideration several maize bread-quality traits, it is possible to group the Portuguese maize landraces
INTRODUCTION The nutritional value and technological properties of cerealbased products are largely dependent on the characteristics of their seed-storage proteins.1,2 Although seed proteins represent a minor fraction of the cereal endosperm,1 products based on gluten (a cohesive protein-based mass, e.g., wheat bread and pasta) are particularly relevant as a primary food source. The market for gluten-free cereal-based products using rice, sorghum or maize3 has grown considerably because of the increasing number of consumers diagnosed with celiac disease or other allergic reactions/intolerances to gluten (average worldwide prevalence estimated in 1:266; estimated prevalence in Europe and in the United States 1:1003,4). In Portugal, farmers continue cultivating traditional maize landraces, and maize flour is used to produce a type of cornbread (“broa”) highly appreciated © XXXX American Chemical Society
Received: January 4, 2013
A
dx.doi.org/10.1021/pr400012t | J. Proteome Res. XXXX, XXX, XXX−XXX
Journal of Proteome Research
Article
into distinct bread-quality groups.5 One of the traits that contribute to the discrimination is protein content,5,8 but proteins or protein classes that can contribute more significantly to such discrimination are still unknown. Maize seeds accumulate proteins of the globulin type (7S) in the aleurone cell layer and of the prolamin type in the starchy endosperm.2 More recently it was shown that some prolamins can also accumulate in the aleurone cells.9 In maize, prolamins are known as zeins, a heterogeneous group of proteins, designated as α-, β-, γ-, or δ-zeins on the basis of their structure.2,10 Zeins are described as being determinant in endosperm development and in the resulting kernel hardness,11,12 which in turn is relevant for the milling performance, a quality parameter for industry.13 In addition to the storage proteins, the nonstorage protein pool is also considered as relevant for rheological behavior at least in wheat.14,15 In wheat inbred lines, nongluten factors can contribute as much as 50% to the variation in the bread-making quality.15 A detailed study on the seed flour proteins of two Portuguese maize inbred lines, pb269 (flint) and pb369 (dent) is presented here. These lines differ in the grain hardness and also belong to two contrasting groups concerning the “broa” bread-making ability (M. Carlota Vaz Patto and Carla Brites, personal communication). A proteome study of the maize seed, which can accumulate up to 10% of protein in the endosperm,16 deals with a very high dynamic range of protein abundance. The endosperm accumulates seed-storage proteins (SSP) that, because of their abundance, will predominate in the obtained protein profile. In order to increase the dynamic range of the proteomic analysis (being able to analyze both highly abundant proteins and low abundance proteins), the Osborne fractionation was performed. This method makes use of the distinct solubility and extractability of proteins in several solvents, e.g., ionic strength and pH.17 The sequential extraction allows the fractioning of the sample into several parts. The process significantly reduces the complexity within each protein fraction while concentrating protein species, which results in a larger number of visible proteins by 2-DE. This approach allowed the evaluation of the zein and nonzein content (total amount and specific classes) in relation to endosperm hardness. It also allowed an evaluation of how such protein classes contribute to the discrimination of the inbred lines. This highlights the fact that relative storage protein content (the proportions of each zein-type and globulin-type) can be determinant in the maize flour technological ability for bread making. This study furthermore provides a working hypothesis for breeders, as the genes that have been identified can be used for marker-assisted breeding or gene-transformation programs.
Table 1. Characteristics of the Maize Germplasm Used, Three Commercial Varities (Arzano, DKC6575, PR32R43), and Two Traditional Maize Inbred Lines (pb269, pb369) Developed by the Portuguese Plant Germplasm Bank (PBGV) name pb269 pb369 Arzano
type of germplasm/owner or company traditional/germplasm bank traditional/germplasm bank commercial/Maisadour Semences commercial/Monsanto Company
untransformed DKC6575 (MON810 control) PR32R43 commercial (genetically modified)/Pioneer Hi-Breed Seeds
grain type
flour color
preferred use for “broa”
FLINT DENT FLINT
yellow white yellow
yes no no
DENT
yellow
no
DENT
yellow
no
from the flour (150 μm) were sequentially extracted according to their solubility,18,19 and the four Osborne fractions were obtained: water-soluble (albumins), salt-soluble (globulins), alcohol-soluble prolamins (ethanol extracted), and alcohol-insoluble prolamins (NaOH extracted). The protein concentration in each fraction was determined using a modified Bradford procedure.20 Two-Dimensional Electrophoresis of the Flour Proteins of pb269 and pb369
With the exception of the alcohol-soluble prolamins, the protein fractions were dialyzed (SnakeSkin, 3.5 kDa cutoff, ThermoScientific) prior to lyophilization. Samples were resuspended in a 2DE solution [8 M urea, 2% (w/v) CHAPS, and 65 mM DTT], and after two hours at room temperature, the protein concentration was determined.20 For each protein fraction the electrophoretic conditions were optimized, and prestained molecular weight markers were used to monitor the SDS-PAGE run (161− 0372, Bio-Rad Laboratories). Each protein fraction, from each inbred line (4−5 biological replicates), was run in 13 cm gels (Table S1, Supporting Information). Isoelectric focusing (IEF) was carried out with (i) pH 3−10NL immobilized strips loaded with 350 μg (albumins and globulins) or 100 μg (alcohol-insoluble prolamins) and run to a cumulative 32 120 Vh; or with (ii) pH 6−11 immobilized strips loaded with 10 μg (alcohol-soluble prolamins) and run to a cumulative 46 000 Vh. For the albumin, globulin and alcohol-insoluble prolamin fractions, the second dimension was carried out in SDS-PAGE [12.5% T (g of acrylamide plus bisacrylamide per 100 mL); 2.5% C (g of bisacrylamide per g of acrylamide plus bisacrylamide)]. For the alcohol-soluble prolamins, the SDS-PAGE was carried out with 15−20% T, and the electrophoresis was run until the 15 kDa molecular weight marker got out the gel. After SDS-PAGE the gels were stained with colloidal coomassie blue.21 The gels were scanned with an ImageScanner II (Amersham Biosciences), and the images were analyzed with the REDFIN software (Ludesi AB) or the ImageMaster-Platinum, v.5.0, software (GE Healthcare). In each Osborne fraction a spot was considered present when detected in ≥80% of the gels (i.e., of the replicates). Relative spot volumes were used for the subsequent statistical analysis.
■
MATERIALS AND METHODS Seeds from the Portuguese traditional maize inbred lines pb269 (flint) and pb369 (dent) were used. These lines derived from traditional germplasm and were developed by the Portuguese Plant Germplasm Bank (PBGV). For this study, seeds multiplied at ESAC in 2006/07 (Escola Superior Agrária de Coimbra) were used. For comparison purposes, the seed protein content and solubility from three commercial maize varieties were also evaluated. Table 1 summarizes some of the characteristics of the germplasm used.
Protein Identification, Database Search, and Functional Categorization
Solubility-Based Protein Fractionation (Osborne Fractionation)
For the MS-based identification, spots were cut from the gels and were processed as previously described.19 Trypsin was used in the processing of spots from the water and salt-soluble
Seeds were ground in a Retsch ZM100 sample mill using 0.5 mm aperture and sieved through a 150 μm mesh. The proteins B
dx.doi.org/10.1021/pr400012t | J. Proteome Res. XXXX, XXX, XXX−XXX
Journal of Proteome Research
Article
secondary searches against the NCBI EST viridiplantae database were carried out. The result of these additional searches is shown in Table S4.1−4.4 (Supporting Information) in red. Identified proteins were grouped in functional categories using the Gramene Ontologies database (www.gramene.org/ plant_ontology/).22 When no result was obtained with Gramene, quickGO (www.ebi.ac.uk/QuickGO/)23,24 was used.
protein fractions, while for the alcohol-soluble and insoluble prolamins chymotrypsin was used in addition to trypsin.19 Briefly, the excised gel plugs were washed twice with 100 μL of 50 mM ammonium bicarbonate in 50% MeOH followed by two additional washes with 75% acetonitrile (ACN) to remove residual stain. After drying, the destained spots were incubated at 37 °C in 8 μL of 20 mM ammonium bicarbonate containing 5 ng/μL of trypsin or chymotrypsin. In the later case, 2 mM of calcium chloride were added to the chymotrypsin for the digestion. After incubation, the resulting peptides were extracted by washing the gel plugs twice with 50% ACN/0.1% trifluoroacetic acid (TFA). The dried peptides were dissolved in 50% ACN/0.1% TFA, and 0.7 μL was mixed with an equal volume of matrix solution (5 mg of α-cyano-4-hydroxycinnamic acid in 700 μL of 50% ACN/0.1% TFA) and applied to stainless steel target plates. The Applied Biosystems 4800 (Foster City, CA) MALDI TOF/TOF was used in this study for all MS and MS/MS analyses. The mass spectrometer was calibrated internally with autocleavage products of trypsin, or chymotrypsin for the spots digested with this protease. When these products were absent, an updated default calibration file using the 4700 calibration mixture (Applied Biosystems) was used for external calibration. For MS/MS analysis an external calibration based on Glu-fibrinopeptide fragments was used. For each spot one MS-spectrum was acquired, accumulating 1500 laser shots, and the 8 highest peaks from this spectrum (with a signal-to-noise above 30, excluding autocleavage products and known contaminants) were automatically selected for MS/MS analysis in 1 kV CID-on fragmentation mode. For each MS/MS spectrum 3000 single shots were accumulated. All database searches were done using the in-house Mascot server on a GPS platform (Applied Biosystems), combining the MS-spectrum with a maximum of 8 MS/MS spectra in a single search. Although not all searches were done at the same moment, the database used was the one previously used in the published preliminary study.19 The data were searched against the maize protein database downloaded on the 27th of November, 2007, containing 172 137 sequences. A sequence composed of aligning several maize ESTs with the EST 108706671 was often identified during the preliminary study as main contributor; this sequence was added to the database and was designated as cupin sequence. All searches were done using a peptide mass and fragment mass tolerance of 100 ppm and 0.5 Da, respectively. For trypsin two missed cleavages were allowed (four for chymotrypsin) and carbamidomethyl cysteine as fixed modification. As variable modifications the oxidation of methionine and the oxidation of tryptophan to kynurenine and N-formylkynurenine were allowed. MS and MS/MS peak lists were composed by extracting the peaks with a minimal signal-to-noise of 10 from the raw spectrum, and a maximal number of 60 masses, representing the 60 spectral features with the highest signal-to-noise, were retained. Furthermore, known contaminants such as peaks corresponding to keratin and the used protease were excluded from the peptide mass fingerprint analysis. For MS/MS analysis low mass peaks (m/z < 60) and peaks closer than 20 to the precursor were also excluded. Identifications solely based on PMF were rejected; a protein was considered to be identified when at least 2 peptides attained an ion-score above the threshold calculated by MASCOT (>35) or when the significant score for 1 peptide was combined with a protein expect-value of 0.050 >0.050
>0.050 >0.050
0.050
25.5 ± 2.8 28.8 ± 2.6 23.6 ± 1.8 26.7 ± 1.4 25.7 ± 1.4 >0.050 not applicable
10.9 ± 1.0 9.2 ± 0.8 11.6 ± 2.1 11.2 ± 2.0 12.9 ± 0.9 >0.050 not applicable
9.9 ± 0.8 15.4 ± 1.7 15.6 ± 2.1 13.1 ± 0.4 12.3 ± 1.4 0.029 pb269−pb369
1.5 ± 0.6 0.9 ± 0.1 0.7 ± 0.1 3.8 ± 0.3 2.6 ± 0.2 0.012 pb369−MON Arz−MON
77.7 ± 1.1 74.6 ± 2.2 72.0 ± 4.0 71.9 ± 2.1 72.2 ± 0.5 >0.050 not applicable
>0.050 >0.050
>0.050 >0.050
0.050 >0.050
>0.050 >0.050
alcohol-soluble prolamins 0.1 0.1 0.2 0.1 0.1
alcohol-insoluble prolamins 57.9 ± 51.5 ± 55.3 ± 46.5 ± 48.9 ± 0.016 ns
3.8 1.2 2.0 0.2 1.4
a
The Kruskal−Wallis test (nonparametric alternative to one-way ANOVA) was used to assess the existence of significant differences between maize cultivars and/or inbred lines. The two Portuguese inbred lines (pb269 and pb369) are compared with the nonparametric tests Klomogorov− Smirnov and Mann−Whitney U. bpb269 (n = 6); pb369 (n = 7); Arz, Mon810control, PR32R43 (n = 4); flour (< 150 mm); bran (> 150 mm).
than 50% (but below 55%). Considering the % of volume (relative to each protein fraction), spot volumes as low as 0.01% (in the globulin fraction) and as high as 36% (in the alcohol-soluble prolamin fraction) are registered. The average spot volume in albumins is 0.29%, in globulins it is 0.44%, and in prolamins it is 2.65% (alcohol-soluble fraction) and 0.92% (alcohol-insoluble fraction). However, since the relative protein abundance in each fraction is very distinct (Table 2), the actual abundance of each protein spot is lower (Table S3, Supporting Information).
differences were noted in the relative amount of the globulin and the alcohol-soluble prolamin fraction. Pairwise comparisons (posthoc test) show that differences can be assigned to the Portuguese inbred lines in respect to the globulin fraction, but not to the alcohol-soluble fraction. A single comparison using nonparametric tests (Kolmogorov−Smirnov and Mann− Whitney U tests, Table 2) confirms the differences between the two inbred lines concerning the relative amount of globulins. Seed Protein Polymorphism Evaluated by Osborne Fractionation Followed by 2DE
Mass Spectrometry Analysis
Flour proteins of the four Osborne protein fractions from the two inbred lines were separated by two-dimensional electrophoresis (Figure 1). The adopted methodology allows detecting 967 spots (358 common to both lines; 208 specific to pb269 and 401 specific to pb369; Table S1, Supporting Information). Separate analysis of alcohol-soluble and alcoholinsoluble prolamins shows the enrichment in the alcoholsoluble fraction of highly abundant spots concentrated in a very narrow MW range and with basic pH (Figure 1C; Figure S1, Supporting Information). Data on relative spot volume (average ± standard error), molecular weight (gel-based calculated), pI (gel-based calculated) and statistical analysis (univariate and multivariate) are available in Table S2 (Supporting Information). Despite the high variability in spot volume usually reported in 2D experiments,34,35 our data set shows a high consistency of quantification (Table S1, Supporting Information). The coefficient of variance is lower than 25% for 792 spots (i.e., 81.9% of the considered spots), the average coefficient of variance being 20.0% ± 0.5% for pb269 and 16.5% ± 0.4% for pb369. Only for 1.4% of the spots (14 spots) is the coefficient of variance higher
For MS identification we used two proteases, trypsin and chymotrypsin. Trypsin was used for 498 spots, while chymotrypsin was used for 199 spots, and an identical set of 115 spots was digested with the 2 proteases. Of the 582 spots submitted to MS analysis (not counting for duplications), positive and unambiguous identification was obtained for 436 spots. Data on protein identification by MS analysis, theoretical MW, theoretical pI and functional categorization are available (Table S2, Supporting Information). Data on MS analysis are available (Table S4.1−S4.4, Supporting Information). From the spots analyzed only with trypsin (n = 384), 298 were identified as a single protein, while in 14 spots two distinct proteins were identified (these spots were not considered for further quantitative analysis). For the spots analyzed only with chymotrypsin (n = 90), positive and unambiguous identification was achieved for 67 spots. When considering the spots that were treated with trypsin and chymotrypsin (n = 115), 11 spots only resulted in identification with trypsin and 12 only with chymotrypsin. For 15 spots it was possible to get an identical identification with D
dx.doi.org/10.1021/pr400012t | J. Proteome Res. XXXX, XXX, XXX−XXX
Journal of Proteome Research
Article
Figure 1. Representative gel images of the Osborne albumins (A), globulins (B), alcohol-soluble prolamins (C), and alcohol-insoluble prolamins (D) fractions extracted from the seed flour of the pb269 inbred line. The proteins selected for identification are numbered in the gels (Table S2, Supporting Information). IEF was carried out with pH 3−10NL immobilized strips loaded with 350 μg (A, B), 100 μg (C) or with pH 6−11 immobilized strips loaded with 10 μg (D). Gels were stained with colloidal coomassie blue.
identification was obtained with chymotrypsin, the inclusion of semitryptic cleavages was beneficial.19 This change in the search parameters allowed the identification of 49 out of these 356 spots. For 9 and 25 spots, the identification respectively with chymotrypsin and with trypsin was solely based on PMF analysis, and these spots were not considered for further analysis. In summary, 41 spots with protein identification were not considered for further quantitative analysis since the identification was not of high confidence (peptide mass fingerprinting for 15 spots) or unambiguous (multiple proteins
both proteases (same UniProtKB accession), while in nine spots the protein identified was of identical molecular function (storage protein), but with different accession numbers. By comparing the amino acid composition of these accessions it was found that in two cases there is no significant similarity; in three spots the protein share less than 70% identity, and in 5 spots amino acid identity is >85%. Only these five spots were considered for further analysis. In seven cases two different proteins were identified within the same spot (not considered for further analysis). Considering the spots for which no E
dx.doi.org/10.1021/pr400012t | J. Proteome Res. XXXX, XXX, XXX−XXX
Journal of Proteome Research
Article
Table 3. Compilation of the 2DE Generated Data (No. Of Spots Per Osborne Fraction) and the Summary of the Statistical Evaluation Using Several Univariate Methodsa prolamins albumins globulins (pH 3−10NL) (pH 3−10NL) no. of spots present no. specific spots % of spots that are differentially expressed
% of specific spots that are differentially expressed no. of common spots % of common spots that are differentially expressed
alcohol-soluble (pH 6−11)
pb269 pb369 KS (p-level
