Spatial and Temporal Localization of Flavonoid Metabolites in

Apr 14, 2017 - *(A.C.C.) Phone: +49 (0) 3641 9485 99. Fax: +49 (0) 3641 9482 02. E-mail: [email protected]. Cite this:J. Agric. Food Chem. 65...
0 downloads 0 Views 4MB Size
Article pubs.acs.org/JAFC

Spatial and Temporal Localization of Flavonoid Metabolites in Strawberry Fruit (Fragaria × ananassa) Anna C. Crecelius,*,†,‡ Dirk Hölscher,⊥,# Thomas Hoffmann,§ Bernd Schneider,⊥ Thilo C. Fischer,§ Magda-Viola Hanke,⊗ Henryk Flachowsky,⊗ Wilfried Schwab,§ and Ulrich S. Schubert†,‡ †

Laboratory of Organic and Macromolecular Chemistry (IOMC), Friedrich Schiller University Jena, Humboldtstrasse 10, 07743 Jena, Germany ‡ Jena Center for Soft Matter (JCSM), Friedrich Schiller University Jena, Philosophenweg 7, 07743 Jena, Germany ⊥ Research Group Biosynthesis/NMR, Max Planck Institute for Chemical Ecology, Hans-Knöll-Strasse 8, 07745 Jena, Germany # Organic Plant Production and Agroecosystems Research in the Tropics and Subtropics (OPATS), University of Kassel, Steinstrasse 19, 37213 Witzenhausen, Germany § Biotechnology of Natural Products, Technical University Munich, Liesel-Beckmann-Strasse 1, 85354 Freising, Germany ⊗ Julius Kühn-Institute − Federal Research Centre for Cultivated Plants Institute for Breeding Research on Fruit Crops, Pillnitzer Platz 3a, 01326 Pillnitz, Germany ABSTRACT: Flavonoids are important metabolites in strawberries (Fragaria × ananassa) because they accomplish an extensive collection of physiological functions and are valuable for human health. However, their localization within the fruit tissue has not been extensively explored. Matrix-assisted laser desorption/ionization mass spectrometric imaging (MALDI-MSI) was employed to shed light on the spatial distribution of flavonoids during fruit development. One wild-type (WT) and two transgenic lines were compared, wherein the transgenic enzymes anthocyanidin reductase (ANRi) and flavonol synthase (FLSi), respectively, were down-regulated using an RNAi-based silencing approach. In most cases, fruit development led to a reduction of the investigated flavonoids in the fruit tissue; as a consequence, they were exclusively present in the skin of mature red fruits. In the case of (epi)catechin dimer, both the ANRi and the WT phenotypes revealed low levels in mature red fruits, whereas the ANRi line bore the lowest relative concentration, as analyzed by liquid chromatography−electrospray ionization multiple-step mass spectrometry (LC-ESI-MSn). KEYWORDS: MALDI, imaging, strawberries, flavonoids, foodomics, Fragaria × ananassa



INTRODUCTION Foodomics has been defined as the in-depth, high-throughput approach for the application of food science to enhance humans’ well-being.1 One aim of this “omics” technology is the accurate determination of food composition. Hence, advanced mass spectrometric techniques have been used thanks to their excellent selectivity and sensitivity.2,3 For profiling metabolites in food, typically, modern separation techniques, such as liquid chromatography (LC), are combined with mass spectrometry (MS).4 However, even though very advanced methods have been explored in the past, the spatial distribution of the investigated metabolites is not possible with this approach. To do so, mass spectrometric imaging (MSI) has started to be adapted in foodomics.5 This mass spectrometric technique enables the spatial analysis of, for example, metabolites, proteins, or pesticides in food products by using soft ionization techniques, such as matrix-assisted laser desorption/ionization (MALDI),6 colloidal graphite-assisted laser desorption/ionization (GALDI),7 or desorption electrospray ionization (DESI).8 To optimize human health and well-being, knowledge of the distribution of compounds correlated with health benefits is gaining special attention in the field of foodomics. Besides the nutritional components, such as minerals, folates, and vitamin C, the strawberry (Fragaria × ananassa Duch.) possesses phenolic compounds, which are crucial for human health.9 These © 2017 American Chemical Society

strawberry phenolic compounds are considered to prevent and improve chronic degenerative diseases, as shown in recent studies.10 Flavonoids have recently been analyzed in strawberry;11 however, their spatial distribution during development is still unknown. The possibility of visualizing flavonoids in fruits, such as rabbiteye blueberry (Vaccinium ashei),12 and ‘Golden Delicious’ apples,13 by employing MALDI-MSI has already been shown, as well as the use of GALDI-MSI for analyzing strawberries.7 Despite the advantage of matrix background reduction, the latter ionization technique shows a lower sensitivity for certain small molecules compared to MALDIMSI.14 Thus, the aim of the present study was to gain a more comprehensive insight into the localization of specialized metabolites, including phenolic acids and flavonoids in strawberry (Fragaria × ananassa Duch.), the economically most important soft fruit worldwide, at two development stages, by employing MALDI-MSI. A wild-type (WT) and two transgenic strawberry lines were analyzed. In the transgenic lines either the enzyme anthocyanidin reductase (ANRi),15 or flavonol synthase (FLSi)16 was down-regulated using an RNAiReceived: Revised: Accepted: Published: 3559

February April 13, April 14, April 14,

7, 2017 2017 2017 2017 DOI: 10.1021/acs.jafc.7b00584 J. Agric. Food Chem. 2017, 65, 3559−3568

Article

Journal of Agricultural and Food Chemistry

Figure 1. Flavonoid biosynthesis in strawberries. Enzyme abbreviations: PAL, phenylalanine ammonia-lyase; C4H, cinnamic acid 4-hydroxylase; 4CL, p-coumarate-CoA ligase; CHS, chalcone synthase; CHI, chalcone isomerase; F3′H, flavonoid-3′-hydroxylase; FHT, flavonone 3-hydroxylase (synonym of F3H); FLS, flavonol synthase; DFR, dihydroflavonol 4-reductase; LAR, leucoanthocyanidin reductase; ANS, anthocyanidin synthase; ANR, anthocyanidin reductase; F3GT, flavonoid-3-O-glycosyltransferase. coated glass slides (Bruker Daltonics, Bremen, Germany), dried at room temperature, and stored in a desiccator for 30 min. Subsequently the MALDI matrix 2′,4′,6′-trihydroxyacetophenone monohydrate (100 mg/mL in acetone, Sigma-Aldrich, Taufkirchen, Germany) was applied onto the strawberry longitudinal sections employing an air-brush device (Revell GmbH, Bünde, Germany). For this procedure, the ITO-coated glass slide with the samples was mounted upright, 25 cm from the airbrush outlet, and 5 s of spraying was alternated with 1 min of airdrying. On average, 20 and 30 cycles of matrix application produced a homogeneous layer of matrix crystals. Mass spectra were immediately recorded using an UltrafleXtreme MALDI TOF/TOF mass spectrometer (Bruker Daltonics), operating in reflectron negative ionization mode with a spatial resolution of 70 μm. A m/z range of 0−1500 was analyzed, accumulating 3000 laser shots per spot. Data sequence preparation, MS acquisition, and visualization were performed using FlexImaging and FlexControl software ver. 4.0 (Bruker Daltonics). Statistical analysis was performed using SCiLS Lab ver. 2015b (SCiLS GmbH, Bremen, Germany). The presented ion images were normalized using the total ion content and weak denoised.17 MALDI-CID Analysis. The tandem MS experiments were also performed on an UltrafleXtreme MALDI TOF/TOF mass spectrometer (Bruker Daltonics), operating in reflectron negative ionization mode employing argon at a pressure of 2.5 bar as collision gas.

based gene silencing approach. Both enzymes are highlighted in red and green, respectively, in Figure 1, depicting the flavonoid biosynthesis. The purpose of this study was to reveal spatiotemporal metabolic alterations, as detected by MALDIMSI, so as to help in understanding the physiological changes and responses that occur during strawberry fruit development and when certain enzymes involved in the flavonoid biosynthesis are silenced.



MATERIALS AND METHODS

Plant Material. Plants of the garden strawberry (Fragaria × ananassa Duch.) cv. ‘Senga Sengana’ were grown in plant chambers (25 °C, 10,000 lx, 80% humidity) and subsequently kept in a greenhouse. Fruits were collected at two development stages: middle-sized green fruits (G2) and mature red fruits (R). Transgenic Strawberry Lines. The transgenic lines of the strawberry cv. ‘Senga Sengana’ expressing the ANRi and FLSi gene construct are identical to lines F15 and F25, respectively, described previously.15,16 MALDI-MSI Analysis. Frozen strawberry sections (thickness = 120 μm) were sliced on a Microm HM 560 cryostat (Thermo Fisher Scientific, Walldorf, Germany) and transferred to conductive ITO3560

DOI: 10.1021/acs.jafc.7b00584 J. Agric. Food Chem. 2017, 65, 3559−3568

Article

Journal of Agricultural and Food Chemistry

Figure 2. Negative MALDI mass spectrum of (A) the synthetic standard kaempferol glucoside and (B) ions originating from the matrix 2′,4′,6′trihydroxyacetophenone monohydrate.

Figure 3. Negative MALDI-CID mass spectrum of the [M − H]− precursor ion at m/z 447 of the synthetic standard kaempferol glucoside using 2′,4′,6′trihydroxyacetophenone monohydrate as matrix.

3561

DOI: 10.1021/acs.jafc.7b00584 J. Agric. Food Chem. 2017, 65, 3559−3568

Article

Journal of Agricultural and Food Chemistry

Figure 4. (A) Optical images and (B) spatial segmentation maps of strawberries (Fragaria × ananassa) considering two fruit development stages: middle-sized green fruits (G2) and mature red fruits (R). One wild-type (WT) and two transgenic lines (ANRi, down-regulation of anthocyanidin reductase; FLSi, down-regulation of flavonol synthase) are presented. Scale bar in the lower left corner corresponds to 1 cm. LC-ESI-MSn Analysis of Fruit Extracts. The analysis of metabolites was performed by LC-ESI-MSn as described previously.15,16 In detail, the high-performance liquid chromatography (HPLC) analysis was conducted on an Agilent 1100 series HPLC system (Agilent Technologies, Waldbronn, Germany) equipped with a quaternary pump and a diode array detector, using a C-18 column (150 mm long × 2.0 mm i.d., particle size = 5 μm; Phenomenex, Aschaffenburg, Germany). The solvent system comprised a binary gradient system: Solvent A consisted of water with 0.1% formic acid and solvent B of 100% methanol with 0.1% formic acid using a flow rate of 0.2 mL/min at a temperature of 28 °C. The gradient program was as follows: 0−30 min, 100% A to 50% A/50% B; 30−35 min, 50% A/50% B to 100% B, hold for 15 min; 100% B to 100% A in 5 min, then hold for 10 min. Mass spectra were recorded using an Esquire 3000plus ESI-ion trap mass spectrometer (Bruker Daltonics). Typically, the metabolites were analyzed in both negative and positive ionization modes, with a capillary voltage of 4000 V, a capillary exit voltage of 121 V, and an end plate voltage of −500 V. Nitrogen was used as drying gas (9 L/min, 330 °C). The acquired mass range (full scan) was between m/z 50 and 800. Helium was used as collision gas (3.56 × 10−6 mbar) with a collision voltage of 1 V. The most abundant signals of [M + H]+, [M − H]−, or [M + HCOO]− of the different metabolites were fragmented in the autotandem MS mode. Individual samples were frozen, lyophilized for 48 h, and, afterward, homogenized to obtain a fine powder employing a mill (Retsch MM 200, Haan, Germany). Three biological replicates were prepared, each using 50 mg of lyophilized powder. The lyophilized powder was dissolved in 250 μL of methanol, vortexed, then sonicated for 10 min, and finally centrifuged at 16000g for 10 min. After removal of the supernatant, the residue was re-extracted with 500 μL of methanol. Both supernatants were combined and dried in a vacuum concentrator, and finally 35 μL of water was added. For the LC-ESI-MSn analysis the clear supernatant was used. This was prepared as described above, by first vortexing the redissolved supernatant (35 μL in water) for 1 min, then, sonication for 10 min, and, finally, centrifugation at 16000g for 10 min.

MALDI matrix 1,8-bis(dimethylamino)naphthalene19 was tested, because Ye et al.20 described its successful application for the MALDI-MSI measurements of flavonoids in roots and root nodules of Medicago truncatula during nitrogen fixation. Unfortunately, this matrix also did not result in useful data. Finally, the matrix 2′,4′,6′-trihydroxyacetophenone monohydrate, as described by Kuhnert and colleagues,21 was tested using an air-brush device to enable the use of acetone, which is not approved for the use of the ImagePrep, because it can affect the performance of the optical sensor. To further confirm the suitability of this matrix, reference compounds of flavonoids were measured to validate the method. Figure 2A shows the MALDI ion spectrum in the negative mode of a synthetic standard of kaempferol 3-O-glucoside. The diagnostic ion of kaempferol 3O-glucoside is visible at m/z 447. However, additional signals are visible in the mass spectrum, which could be assigned to matrix signals, as presented in Figure 2B. The signals at m/z 1066 and 1300 in Figure 2A are presumably clusters of the standard and the matrix 2′,4′,6′-trihydroxyacetophenone monohydrate. To prove that the signal at m/z 447 in fact belongs to kaempferol 3-Oglucoside, collision-induced dissociation (CID) was performed on the deprotonated flavonoid glucoside. The resulting CID mass spectrum is presented in Figure 3. The CID fragmentation pattern shows two signals, at m/z 284 and 285, which are formed through a homolytic or heterolytic cleavage of the O-glycoside bond.22 The chemical structure of the resulting product ions is also displayed in Figure 3. The successful identification of the synthetic flavonoid standards motivated us to use 2′,4′,6′trihydroxyacetophenone monohydrate for the following MALDI-MSI analysis of strawberries. Visualizing the Spatial Regions of Plant Metabolites in the Strawberry Fruit (Nontargeted Analysis). To gain information as to whether major spatiotemporal alterations are induced by silencing events and fruit development, the large MALDI-MSI data sets acquired were mined by computational means of spectral segmentation.23 Subsequently, the MALDIMSI data sets of the WT and the two transgenic lines, ANRi and FLSi, at two development stages, middle-sized green fruits (G2) and mature red fruits (R) (in total six data sets), were clustered



RESULTS AND DISCUSSION Matrix Selection for MALDI-MSI Analysis. To reduce the sample preparation time and to allow a spatial resolution of 10 μm, strawberry sections were initially analyzed without employing a matrix according to the method of Hölscher et al.18 However, an almost complete fragmentation of the specialized metabolites occurred, as described by Franceschi et al.,13 so the 3562

DOI: 10.1021/acs.jafc.7b00584 J. Agric. Food Chem. 2017, 65, 3559−3568

Article

Journal of Agricultural and Food Chemistry

Figure 5. MALDI-MSI analysis of strawberries (Fragaria × ananassa) considering two fruit development stages: middle-sized green fruits (G2) and mature red fruits (R). One wild-type (WT) and two transgenic lines (ANRi, down-regulation of anthocyanidin reductase; FLSi, down-regulation of flavonol synthase) are presented. The ion images are normalized using TIC. Scale bar in the upper left corner corresponds to 1 cm.

the WT strawberry at the fruit development stage R, all four segments appeared as well: Segment 1 (dark red) made up the skin, and the remaining three segments, 2 (red), 3 (pink), and 4 (green), were present in the cortex and pith. In the transgenic line ANRi at stage G2, only three segments were observed: segment 1 (dark red) at the skin, segment 2 (red) at the cortex, and segment 3 (pink) at the centeral area of the flesh, close to the pith. Only two segments could be differentiated in the ANRi line at stage R: Segment 1 (dark red) is the skin, whereas segment 3 (pink) includes the flesh from the cortex and pith. The transgenic line FLSi has only two segments in the fruit development stages G2 and R: Flesh, including skin, cortex, and pith, make up segment 1 (dark red) at stage G2, whereas only the veins of the strawberry fruit form segment 3 (pink). At fruit development stage R, segment 4 (green) forms the skin and segment 3 (pink) is made up of the flesh, including the cortex and pith. It can therefore be concluded that the localization of plant metabolites in the fruit tissue depends on the fruit development stage as well on the silencing event.

according to their spectral similarities (Figure 4B). Only these two development stages (G2 and R) were selected because they are distinguishable in their phenotype in the WT fruits and can clearly be separated from the ANRi transgenic fruits, which already had a slightly red color at stage G2 (Figure 4A). Hence, we postulated that when the phenotype is distinguishable, the spatial distribution of the considered specialized metabolites should be discriminative. Additionally, we added the FLSi transgenic fruits, which show the same phenotype as the WT (Figure 4A), to see if a variation in the spatial distribution of the considered metabolites is visible, even though the phenotype is similar to the WT. In total, four color-coded segments could be defined (Figure 4B), assisting in the interpretation of prominent spatial regions in the strawberry fruit, such as cortex and pith (Figure 4A). The depth at which the strawberries were cut is approximately 1 cm. In the WT strawberry fruit at development stage G2, four segments could be specified: Segment 1 (dark red) was made up of the skin and segment 4 (green) of the pith. Segment 2 (red) and segment 3 (pink) were from the cortex. In 3563

DOI: 10.1021/acs.jafc.7b00584 J. Agric. Food Chem. 2017, 65, 3559−3568

Article

Journal of Agricultural and Food Chemistry Table 1. Metabolites Tentatively Identified in Strawberry Tissue (Fragaria × ananassa) observed m/z [M − H]− 191 211 280 397 421 459 509 577 583 727 765 1033 a

MS2 main fragments from MALDI-CID

tentative assignment

class

nd nd nd 80, 131, 165, 208, 218, 228, 294, 330, 353, 379 80, 190, 254, 319, 379, 405 166, 190, 279, 292, 332, 356, 370, 442 166, 172, 191, 278, 479, 495 165, 189, 289, 332, 346, 371, 409, 479, 533, 560 nd 167, 191, 279, 319, 333, 371, 393, 459 167, 191, 357, 422, 497, 561, 599, 661 167, 191, 280, 334, 497, 527, 765, 803, 841

citric acidb 3-hydroxy-4-methylphenyl lactic acidc unknown compound unknown compound octacosanoic acidc dicaffeoylthreonic acidd,31 unknown compound (epi)catechin dimerb unknown compound unknown compound unknown compound unknown compound

organic acid phenolic acid

molecular formula C6H7O7 C10H12O5

C28H54O2 C22H20O11 C30H25O12

a

phenolic acid phenolic acid flavan-3-ol

nd, not detected. bIdentification by LC-ESI-MSn analysis. cIdentification by METLIN database search. dIdentification by comparison with literature.

Visualizing the Spatial Distribution of Primary and Specialized Metabolites Involved in the Flavonoid Biosynthesis of Strawberries. The flavonoid biosynthesis has been thoroughly investigated,24,25 and major enzymes have been described in strawberry fruits.26 The pathway known for strawberry fruit is depicted in Figure 1. In this study, besides a WT line, two transgenic lines, ANRi and FLSi, were investigated. In the ANRi line, silencing of the enzyme ANR causes a reduced production of (epi)catechin and, subsequently, of proanthocyanidins, whereas in the FLSi line, the enzyme FLS is downregulated, causing a reduced production of kaempferol and quercetin and its respective derivatives.16 Ion images of the individual primary and specialized metabolites involved in the flavonoid biosynthesis were generated (Figure 5) to analyze changes in their spatiotemporal location during fruit development and when the enzyme ANR or FLS was silenced. The ion images in a green color scale (normalized to the total ion content and weak denoised)17 were superimposed onto the optical images of the strawberry fruit acquired prior to the matrix deposition. In Figure 5, the m/z values as measured by MALDIMSI of the recorded metabolites are presented. To putatively identify each recorded m/z value, MALDI-CID experiments were carried out, and the results are displayed in Table 1. Additionally, some of the metabolites were tentatively identified by METLIN database search (Table 1). Of the determined flavonoids, flavon-3-ol (epi)catechin dimer (both isoforms) was selected as an example for typically CID experiments. Figure 6 represents the CID results obtained from the two kinds of ionization techniques commonly applied: Figure 6A MALDITOF/TOF (collision gas: argon) and Figure 6B ESI-ion trap analysis (collision gas: helium). The diagnostic ions present in both mass spectra are denoted with an asterisk, and the corresponding putative fragmentation scheme is highlighted in Figure 6C. The different ionization techniques and collision gases (argon or helium) used can have caused the differences between both spectra. Targeted Analysis of Selected Flavonoids Involved in the Flavonoid Biosynthesis of Strawberries. To discuss the spatiotemporal metabolic alterations revealed by MALDI-MSI and their relative concentrations in fruit extracts as determined by LC-ESI-MSn in more detail, two metabolites were selected, one from the citric acid cycle and one from the flavonoid pathway. Because quantitation by MALDI-MSI27,28 is challenging, a second MS technique, LC-ESI-MSn, was selected to gain information on the relative concentrations of the selected metabolites in strawberry fruits. The primary metabolite selected

is citric acid, a precursor of 2-oxoglutarate, which serves as a cofactor for the flavonoid enzyme flavonol synthase (FLS) (Figure 1). FLS is down-regulated in the FLSi transgenic line. The WT phenotype shows an alteration in the localization of citric acid from the cortex to the skin region going from G2 to R (Figure 7B, green color). The ANRi line shows the same trend, whereas in the FLSi line, the change from the cortex to the skin through fruit development is not as pronounced as for the WT and ANRi lines. The corresponding LC-ESI-MSn data for citric acid are also presented (Figure 7A, green color). To obtain relative quantitation by LC-ESI-MSn, the equivalent ‰ dry weight (DW) in relation to the flavonoid standard biochanin A was measured. In Table 2, the calculated p values using the Wilcoxon−Mann−Whitney U test for the nonparametric analysis of intergroup comparison are additionally presented. p values of