Metabolic profiling discriminates between strawberry (Fragaria × ananassa Duch.) cultivars grown in Finland or Estonia

    Metabolic profiling discriminates between strawberry (Fragaria Duch.) cultivars grown in Finland or Estonia

ananassa

Anna K˚arlund, Ulvi Moor, Gordon McDougall, Marko Lehtonen, Reijo O. Karjalainen, Kati Hanhineva PII: DOI: Reference:

S0963-9969(16)30404-5 doi: 10.1016/j.foodres.2016.09.013 FRIN 6412

To appear in:

Food Research International

Received date: Revised date: Accepted date:

16 June 2016 5 September 2016 8 September 2016

Please cite this article as: K˚ arlund, A., Moor, U., McDougall, G., Lehtonen, M., Karjalainen, R.O. & Hanhineva, K., Metabolic profiling discriminates between strawberry (Fragaria ananassa Duch.) cultivars grown in Finland or Estonia, Food Research International (2016), doi: 10.1016/j.foodres.2016.09.013

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Metabolic profiling discriminates between strawberry (Fragaria × ananassa Duch.) cultivars grown in Finland or Estonia

T

Anna Kårlund*1, Ulvi Moor2, Gordon McDougall3, Marko Lehtonen4, Reijo O.

SC R

1

IP

Karjalainen1,5, and Kati Hanhineva1

Institute of Public Health and Clinical Nutrition, University of Eastern Finland, P.O.

2

NU

Box 1627, FI-70211 Kuopio, Finland

Estonian University of Life Sciences, Institute of Agricultural and Environmental

MA

Sciences, Kreutzwaldi 1A, EE51014 Tartu, Estonia The James Hutton Institute, Invergowrie, Dundee DD2 5DA, Scotland UK

4

School of Pharmacy, University of Eastern Finland, P.O. Box 1627, FI-70211 Kuopio,

D

3

Department of Environmental and Biological Sciences, University of Eastern Finland,

CE P

5

TE

Finland

P.O. Box 1627, FI-70211 Kuopio, Finland

AC

* Correspondence to: Anna Kårlund E-Mail: [email protected] Tel.: +358 50 371 1421; Fax +358 17 162 131

E-Mail addresses for authors: U. Moor, [email protected]; G. McDougall, [email protected]; M. Lehtonen, [email protected]; R.O. Karjalainen, [email protected]; K. Hanhineva, [email protected] Running title: The cultivar-specific metabolic profiles of strawberry

1

ACCEPTED MANUSCRIPT ABSTRACT Metabolic profiling analysis with LC-ESI-QTOF-MS was utilized to separate and

T

identify 187 putative major metabolites demonstrating significantly different levels in

IP

15 strawberry cultivars grown in Finland or Estonia. Principal component analysis

SC R

showed close clustering of genetically related samples grown in Estonia, and hierarchical cluster analyses highlighted differences and similarities in their metabolic

NU

profiles driving separation between cultivars with specific metabolic phenotypes. Phenolic acids, flavonoids, flavan-3-ol derivatives, terpenes, and many types of

MA

glycosidically bound aroma and flavor precursors showed clear variation between strawberry cultivars. These cultivar-specific differences in the levels of major

D

potentially bioactive phytochemicals in strawberries suggests that cultivar selection is

TE

essential for breeding strawberry cultivars with optimal phytochemical compositions

CE P

contributing to possible functional properties and good cultivation and sensory qualities.

Keywords: Fragaria

× ananassa

Duch., non-targeted metabolite profiling,

AC

metabolomics, cluster analysis, cultivars

2

ACCEPTED MANUSCRIPT 1. INTRODUCTION In traditional strawberry breeding programs, good disease resistance, storage longevity,

T

and high yield have been among the main objectives (Bestfleisch, Möhring, Hanke, Peil

IP

& Flachowsky 2014). From the producers’ point of view, firmness, flavor, and shelf life

SC R

are the most valued traits of strawberry (Yue et al. 2014), while consumers appreciate tastiness and appealing appearance (Rosenfeld & Nes 2000). In fruits and berries, some

NU

classes of phytochemicals influence both the sensory characteristics (e.g. taste, flavor and odor) and the ability to repel biological enemies (Aharoni, Jongsma &

MA

Bouwmeester 2005; Nieuwenhuizen et al. 2013; Tohge, Alseekh & Fernie 2014). For example strawberry metabolism is to some extent regulated in response to

D

environmental stimuli because many types of plant metabolites are produced in

TE

response to biotic and abiotic threats (Crupi, Pichierri, Basile & Antonacci 2013; Kårlund et al. 2014a). However, the extent of fluctuation of the strawberry metabolite

CE P

profile in response to different environmental conditions is often cultivar-dependent (Aaby, Mazur, Nes & Skrede 2012; Kårlund, Hanhineva, Lehtonen, Karjalainen &

AC

Sandell 2015). Strawberry cultivars differ in their qualitative and quantitative metabolite profiles (Aaby et al. 2012; Kårlund et al. 2015), and the metabolic networks determining the development, quality and defense mechanisms of strawberries can be either directly or indirectly modified by breeding (Bestfleisch et al. 2012; Bestfleisch et al. 2014; Ulrich & Olbricht 2013). However, it has been shown that improvement in disease resistance of fruits may reduce their sensory quality (Goff & Klee 2006; Nieuwenhuizen et al. 2013). Strawberries grown in Scandinavia and Baltic states both benefit from the strong image of pure and healthy Arctic berries and berry products owing to the northern climate and low-use of pesticides. The cultivation practices in these countries are generally similar, 3

ACCEPTED MANUSCRIPT and the strawberry cultivars that are grown in Finland and Estonia are suitable for cultivation in northern areas, in general. In this study, we surveyed the metabolite

T

profiles of 15 strawberry cultivars grown on commercial farms in Finland or in Estonia

IP

to define potential differences in phytochemical composition relevant to consumers,

SC R

farmers and breeders. Liquid chromatography-based separation of analytes combined with highly accurate and sensitive mass spectrometry technologies provide a modern,

NU

universal approach for plant secondary metabolite profiling (Allwood & Goodacre 2009; Crupi, Genghi & Antonacci 2014). In comparison to traditional, targeted

MA

approaches, non-targeted metabolite profiling offers a hypothesis-free means for highthroughput examination of small chemical constituents in complex biological samples.

D

In this study, analysis with liquid chromatography connected with negative electrospray

TE

ionization in quadrupole–time-of-flight–mass spectrometry (LC-ESI-QTOF-MS) was utilized to separate and identify major compounds showing significant variation

CE P

between strawberry sample types. Use of multivariate statistical analyses such as principal component analysis (PCA) and hierarchical cluster analysis was applied to

AC

uncover similarities in metabolite profiles among cultivars. 2. MATERIALS AND METHODS 2.1 Strawberries

Samples from 15 garden strawberry (Fragaria × ananassa Duch.) cultivars were collected from commercial strawberry farms in June-July 2013. Cultivars Capri, Clery, Dely, Ischia, Joly, Linosa, Marmolada, and Sonata were collected from a farm located near Tartu (58.23°N, 26.43°E) in Southeastern Estonia. Finnish samples were collected from four different farms in Northern Savonia, Eastern Finland. Cultivar Bounty was grown on a farm located in Kaavi (62.586°N, 28.295°E), and cultivar Rumba was produced on a farm located in Karttula (62.53°N, 26.58°E). Two farms were located in 4

ACCEPTED MANUSCRIPT Leppävirta (62.30°N, 27.475°E); one farm produced cultivar Honeoye, and the other produced cultivars Florence, Jonsok, Polka, and Salsa. Agronomic practices on the

T

farms were applied according to the typical industry standards in Finland and Estonia.

IP

The collected fruits were in the same, fully ripe stage. Of each cultivar, 200-2000 g

SC R

were collected. After harvesting, pooled whole fruit samples were immediately placed on ice and frozen as soon as possible with liquid nitrogen. Strawberries were stored at -

2.2 Non-targeted metabolite profiling

NU

20 °C until further processing.

MA

Sample preparation. For the metabolite extraction, 10 frozen fruits (61-152 g) for each cultivar were gently thawed in a microwave oven and homogenized with stick blender.

D

Chemical compounds were extracted from pureed strawberries with 80% methanol (30

TE

mL per 10 g of puree). Suspensions were quickly vortexed and extracted in a horizontal shaker (Unimax 2010, Heidolph Instruments, Schwabach, Germany) at 400 rpm for 15

CE P

min at room temperature. Extracts were filtered through a gauze and centrifuged (Centrifuge 5810R, Eppendorf, Hamburg, Germany) at 3220 g for 5 min. The clarified

AC

supernatants were stored at -75 °C until analysis. Prior to the metabolite analysis, samples were again centrifuged and filtered (Acrodisc® 0.22 μm PTFE filter, PALL, Port Washington, NY, USA). LC-MS conditions. The LC-MS conditions have been described in detail earlier (Kårlund et al. 2015). Shortly, the samples were analyzed by LC-ESI-QTOF-MS (Agilent Technologies, Waldbronn, Karlsruhe, Germany) that consisted of a 1290 LC system, a Jetstream electrospray ionization (ESI) source, and a 6540 UHD accuratemass QTOF spectrometer. For each strawberry cultivar, three biological sample replicates were analyzed, one replicate being a pool of 10 individual fruits. Two microliters of the sample solution 5

ACCEPTED MANUSCRIPT were injected onto an RP column (Zorbax Eclipse XDB-C18, 2.1 × 100mm, 1.8 µm) (Agilent Technologies, Palo Alto, CA, USA) kept at 50 °C. Mobile phases, delivered at

T

0.4 mL/min, consisted of water (eluent A) and methanol (eluent B) (Sigma-Aldrich, St.

IP

Louis, MO), both containing 0.1% (v/v) of formic acid (Sigma-Aldrich, St. Louis, MO).

SC R

Data were acquired in negative electrospray ionization (ESI-), 2 GHz extended dynamic range mode was used, and the instrument was set to acquire over the m/z 20–1600. For

NU

automatic data dependent MS/MS analyses, precursor isolation width was 1.3 Da, and from every precursor scan cycle, 4 most abundant ions were selected for fragmentation.

MA

Collision energies were 10, 20, and 40 V in subsequent runs. The sample tray was kept at 4 °C during the analysis.

D

Data processing. The collection of the data matrix was performed using MassHunter

TE

Qualitative Analysis B.05.00 (Agilent Technologies, Palo Alto, CA, USA). Ions were combined to molecular features exhibiting isotopic peaks, dimers, and common adducts.

CE P

Mass Profiler Professional software (version 2.2, Agilent Technologies, Palo Alto, CA, USA) was used for compound alignment. Noise and low abundance features were

AC

removed from the data matrix which was further filtered to qualify only features present in all three sample replicates of at least one sample type with Benjamini-Hochberg corrected p-value of < 0.05 between all samples (analysis of variance, ANOVA). The final dataset contained 1569 molecular features. k-Means clustering (KMC) analysis with heat map representation (TM4 Microarray Software Suite, available at. www.tm4.org/mev.html; algorithm according to Soukas, Cohen, Socci & Friedman 2000) was applied to analyze and display the levels of the differing compounds among all the sample types. The KMC run was conducted in calculated means mode. In the KMC analysis, the molecular features were grouped to 15 clusters representing 44 to 258 molecular features (3-16%) of total 1569. To build a 6

ACCEPTED MANUSCRIPT hierarchical dendrogram of sample replicates and types, hierarchical clustering (HCL) was utilized (TM4 Microarray Software Suite; Eisen, Spellman, Brown & Botstein )

imca 14,

mea , Sweden) was applied

T

1998). Principal component analysis (P

IP

to monitor the relationship between samples and sample replicates based on their

SC R

metabolite profiles.

For the comparison of the clusters and compound abundances, the data matrix was

NU

exported to Excel and sorted based on the maximum peak area abundance value among each cluster. From each cluster, 5 to 27 major compounds were identified on the basis

MA

of the fragmentation patterns in the data-dependent MS/MS acquisition. The MS/MS data and the molecular masses were compared against the METLIN Metabolite

D

Database, SciFinder, Dictionary of Natural Products, NIST Chemistry WebBook, and

TE

Human Metabolome Database or earlier published work describing fragmentation patterns. Altogether, 187 potential metabolites were tentatively annotated.

CE P

3. RESULTS AND DISCUSSION 3.1 Poly- and heterocyclic compounds, acids and lipids predominate in the

AC

metabolite profiles of strawberry In order to determine the cultivar-dependent differences in the metabolic profiles we focused on metabolite signals that were detected in all three replicates of at least one sample type, with significant (p < 0.05) variation between samples. This resulted in dataset of 1569 molecular features, out of which 187 individual compounds were identified. Significant variation between cultivars was found for many distinguishing strawberry metabolite classes, for example flavonoids (e.g. pelargonidin, quercetin, kaempferol and naringenin derivatives, catechin), condensed tannins (e.g. procyanidins and propelargonidins), phenolic acid derivatives (e.g. cinnamic, coumaric, ferulic, ellagic and gallic acids), and hydrolyzable tannins (Table S1). In addition, the non7

ACCEPTED MANUSCRIPT targeted approach detected some lesser-known compounds that drive the separation between strawberry cultivars, for instance glycosidically bound volatile compounds

T

(e.g. carboxylic acid, benzyl alcohol, and terpene glycosides), hetero- and polycyclic

IP

compounds (e.g. lignans, flavanones), and some other miscellaneous compounds (e.g.

SC R

peptides, carbohydrate derivatives, phenylethyl cinnamates, quinones). 3.2 The metabolite profiles of strawberries were discriminated mainly according to

NU

samples’ genetic background

Principal component analysis (PCA) brought out the variation in the metabolic

MA

phenotypes of strawberry samples (Fig.1.). PC1 and PC2 explained 22.6% and 11.3% of the variation, respectively. Notably, all cultivars grown in Estonia except Sonata (i.e.

D

Capri, Clery, Dely, Ischia, Joly, Linosa, and Marmolada) were grouped together in the

TE

PCA. In addition, cultivars Florence, Rumba, Salsa, and Sonata showed close aggregation in the PCA plot. Cultivars Jonsok and Honeoye were clearly distinguished

CE P

from all the other sample types.

All samples grown in Estonia, except Sonata, are derived from the same producer, and

AC

many of them descend from the same parents (Fig.2.) (Leis, Martinelli, & Castagnoli 2012a; Leis, Martinelli, & Castagnoli 2012c; Leis, Martinelli, Castagnoli, Azzolini, Castagnoli & Castagnoli 2012b; Leis, Martinelli, Castagnoli, Azzolini, Castagnoli & Castagnoli 2012d; Leis, Martinelli, & Castagnoli 2013; Meulenbroek 2007; Musacchi & Leis 1994). Hence, the resemblances in the phytochemical contents of the strawberry samples grown in Estonia may be at least partially due to the commonalities in their genetic backgrounds. It is also known that cultivars Rumba, Salsa, and Sonata are produced by Fresh Forward and that both Sonata and Florence are descendants of cultivar Gorella (Fig.2.) (MEIOSIS 2016; Meulenbroek 2007). Presently, there is no further information available on the parents of Rumba or Salsa. 8

ACCEPTED MANUSCRIPT The molecular features that drove the separation of strawberry samples in the PCA were classified into 15 clusters by the k-means clustering (KMC) analysis (Fig.3.).

T

Interestingly, obvious clusters containing molecular features that were particularly

IP

abundant in one cultivar were found for most cultivars demonstrating the large

SC R

variability in the phytochemical content between strawberry genotypes and possible cultivar-specific patterns. Molecular features that were abundant in Jonsok, Bounty,

NU

Polka, Dely, Florence, Honeoye, Ischia, and Clery were accumulated in clusters 1-6, 8 and 10, respectively. In the hierarchical clustering (HCL) analysis, the samples were

MA

clustered primarily according to sample replicates, and some genetically related cultivars were also closely clustered (Fig.3.). For example, cultivars Jonsok, Bounty,

D

and Polka, all descendants of cultivar Senga Sengana (Fig.2.) (Eikemo, Stensvand,

TE

Davik & Tronsmo 2003) were connected in the HCL dendrogram, and molecular features that were on high level in Jonsok and Bounty were highlighted in cluster 15

CE P

(Fig.3.). The molecular features that were accumulated in cluster 15 were on relatively high level in cultivar Polka, as well. In cluster 3, where Polka was the dominant

AC

cultivar, Jonsok, Bounty, and Sonata were also slightly distinguished. It is known that Polka is a parent of Sonata (Meulenbroek 2007), and that Polka is a cross between cultivars Sivetta and Induka, which is a descendant of Senga Sengana (Eikemo et al. 2003) (Fig.2.).Cluster 14 accumulated molecular features that were relatively abundant in Jonsok and Honeoye, two cultivars related to cultivar Valentine (Eikemo et al. 2003). All cultivars grown in Estonia except Sonata (i.e. Capri, Clery, Dely, Ischia, Joly, Linosa, and Marmolada) were highlighted in cluster 12 in the KMC analysis (Fig.3.). In addition, cultivars Ischia, Capri, Linosa, Clery, and Joly were all closely situated in the HCL dendrogram (Fig.3.), and some molecular features that were on higher level in Capri and Linosa were accumulated in cluster 9. Cultivars Capri, Ischia and Linosa are 9

ACCEPTED MANUSCRIPT all descendants of strawberry cultivar CIVRI30 (Leis et al. 2012a, 2012c; Leis et al. 2013).

T

3.3 Flavonoids, phenolic acids and terpenes showed universal distribution with

IP

some cluster- or cultivar-specific trends

SC R

Certain flavonoid derivatives and phenolic acids were observed in almost all clusters (Table S1), and their ubiquity may reflect their fundamental role in plant ecology and

NU

secondary metabolism. However, some clusters were characterized by the accumulation of specific classes of flavonoids and phenolic acids. For example, cinnamic acid

MA

hexoside derivatives were highlighted in cluster 4 which accumulated molecular features that were on especially high level in cultivar Dely (Table S1; Fig.3.). Cultivar

D

Ischia had the highest levels of two derivatives of aminobenzoic acid anthranilate

TE

(cluster 8), which were detected in only a limited number of cultivars: Ischia, Capri, Dely, Jonsok, and Sonata (Fig.4a.). The putative anthranilate glycoside (MW 461.1532)

CE P

was detected also in Salsa, Marmolada, and Clery. In addition to anthocyanins, other flavonoid derivatives (quercetin, kaempferol,

AC

dihydrokaempferol, naringenin, and apigenin) were present in all cultivars, as well. However, one of the two identified quercetin glucuronides (cluster 3) was detected only in cultivars Bounty, Florence, Jonsok, and, particularly, Polka (Fig.4b.). Taxifolin pentoside (cluster 15) was present in the samples grown in Finland (i.e. Bounty, Florence, Honeoye, Jonsok, Polka, Rumba, Salsa), and in cultivar Sonata grown in Estonia, with the highest level in Jonsok (Fig.4c.). It must be noted that thawing in microwave oven may affect the phenolic profile of strawberries: while the contents of non-anthocyanin strawberry phenolics may be either decreased or increased during microwave oven treatment depending on compound class, anthocyanins tend to decrease (Holzwarth, Korhummel, Carle & Kammerer 2012). 10

ACCEPTED MANUSCRIPT Although many clusters contained at least one terpene as a major component, clusters 5, 12 and 15 were distinguished for containing several terpene compounds. For example, a

T

triterpene derivative (MW 486.3354) in cluster 5 , present only in Florence, Capri, Dely,

IP

and Sonata, showed highest abundance in cultivar Florence. A triterpene hexoside in

SC R

cluster 12 (MW 680.3786) also demonstrated very limited distribution, being present in only the Estonian-grown cultivars Capri, Clery, Dely, Ischia, Joly, and Linosa (Fig.4d.)

NU

with the highest abundance in cultivar Dely. Several sesquiterpene derivatives were detected in cluster 15.

MA

3.4 Cultivar Jonsok contained particularly high levels of potential aroma and flavor precursors

D

In the KMC analysis, molecular features that were most abundant in Jonsok mainly

TE

belonged to cluster 1, which was the largest cluster representing 258 molecular features out of the total 1569 (16%) (Fig.3.).

CE P

Butyrates, furans, acetophenone glycosides, and free and glycosidic forms of carboxylic and dicarboxylic acids were especially high in Jonsok. These types of metabolites are

AC

typical aroma and flavor precursors of fruits and berries (Beekwilder et al. 2004; Mayorga, Duque, Knapp & Winterhalter 2002; Tohge et al. 2014; Zabetakis & Holden 1997). Jonsok has a characteristic taste being intensely sour and bitter with relatively low sweetness (Kårlund et al. 2015; Rosenfeld & Nes 2000), and these properties may well be associated with the high levels of aromatic and aliphatic carboxylic acids, esters, ketones, and furaneols detected in this study (Kårlund et al. 2015; Perez, Olías, Luaces & Sanz 2002; Zabetakis & Holden 1997). In addition, the particularly high level of a procyanidin tetramer (cluster 1; Fig.S1.) in Jonsok may further increase the bitterness of this cultivar (Soares, Kohl, Thalmann, Mateus, Meyerhof & De Freitas 2013).

11

ACCEPTED MANUSCRIPT In the very beginning of the domestication and breeding of Fragaria, strawberries with pleasant sensory characteristics were favored (Ulrich & Olbricht 2013). Since then, the

T

strawberry market has grown and new environmental challenges and consumer demands

IP

have emerged (Bestfleisch et al. 2014). Jonsok, launched in 1969 in Norway (Rosenfeld

SC R

& Nes 2000), is a relatively old strawberry cultivar bred to meet the requirements of commercial strawberry farming in northern areas, in respect of disease resistance

NU

(Koehler et al. 2012), productivity and cold tolerance (Laugale & Lapse 2007). Despite the well appreciated cultivation qualities, its sensory traits (sourness and bitterness) can

MA

make Jonsok less attractive to today’s food industry and consumers. 3.5 Relatively high concentrations of several pelargonidin derivatives were

D

detected in cultivar Polka

TE

The levels of anthocyanin derivatives varied significantly between strawberry cultivars. Polka had the highest content for most of the identified pelargonidin derivatives,

CE P

whereas cultivar Bounty had the highest content of a putative cyandin glycoside derivative, the other main anthocyanidin component in strawberries.

AC

Anthocyanins have many putative health-promoting effects (Del Rio et al. 2013; Carrieri et al. 2013). The supply of anthocyanins from the every-day diet is usually good, even though they are absorbed from the gastrointestinal tract with varying efficiencies (Del Rio et al. 2013). In addition, anthocyanins are rather sensitive to the typical steps of strawberry processing in food production (Kårlund, Moor, Sandell & Karjalainen 2014b). Hence, the relatively higher anthocyanin content of Polka-based raw material

could

mitigate

against

processing-induced

anthocyanin

losses.

Furthermore, the chemical composition of Polka is known to be relatively stable under different cultivation conditions (Kårlund et al. 2015), and this may further ease the quality control and standardization of strawberry products. The fact that cyanidin has 12

ACCEPTED MANUSCRIPT demonstrated higher antioxidant capacity in comparison to pelargonidin (Wang, Cao & Prior 1997) may increase the breeding value of Bounty, which had the highest level of

T

the cyanidin derivative.

IP

3.6 Cultivar Rumba showed exceptionally high content of several condensed

SC R

tannins

Catechin and some dimeric and trimeric proanthocyanidins were at high levels in

NU

cultivar Rumba (Fig.S1.). Additionally, cultivar Sonata showed the highest concentrations of a propelargonidin proanthocyanidin derivative. Rumba and Sonata

MA

were the representative cultivars in cluster 13, where many of the most abundant compounds were proanthocyanidins (Table S1). Rumba and Sonata were also tightly

D

clustered in the HCL dendrogram (Fig.3.).

TE

Flavan-3-ol derivatives are usually accumulated in earlier stages of strawberry fruit ripening (Fait et al. 2008). Hence, the relatively high catechin and proanthocyanidin

CE P

contents of fully ripe Rumba may indicate rather distinctive regulation of flavan-3-ol biosynthesis in late ripening stage. However, it is also possible that Rumba samples

AC

were slightly less ripe and hence richer in flavan-3-ol derivatives but all cultivars were picked when deemed fully ripe. Flavan-3-ols have demonstrated protective effects against cardiovascular diseases (Del Rio et al. 2013), and in this context, cultivars Rumba and Sonata may serve as interesting research and breeding targets. 4. CONCLUSION The present work highlights the complexity of fruit secondary metabolite profiles in different strawberry cultivars, which can distinguish between cultivars and suggest their genetic inter-relationships. The diversity of polyphenol components makes strawberries good sources of potentially bioactive constituents. However, the metabolomic approach also highlighted the presence of many other phytochemicals (including unknown 13

ACCEPTED MANUSCRIPT components) which may also contribute to the physiology, quality and potential bioactivities of strawberries. We conclude that careful cultivar selection is essential to

T

the breeding of modern strawberry cultivars with optimal and stable phytochemical

IP

compositions contributing to the potential functionalities and sensory qualities. Because

SC R

the constancy of strawberries’ chemical composition in different farming and processing conditions is also cultivar-dependent it would be useful to survey the stability of the

NU

chemical composition of strawberry genotypes cultivated and processed with different

AC

CE P

TE

D

MA

methods.

14

ACCEPTED MANUSCRIPT ACKNOWLEDGEMENTS his study was financially supported y the Finnish ultural Foundation, the

Rikala Horticultural Foundation, the Northern Savonia ELY-center, Biocenter

T

ro

ai u and

IP

Finland, and Academy of Finland.

AC

CE P

TE

D

MA

NU

SC R

The authors declare no conflict of interest.

15

ACCEPTED MANUSCRIPT REFERENCES

T

Aaby, K., Mazur, S., Nes, A., & Skrede, G. (2012). Phenolic compounds in strawberry (Fragaria x ananassa Duch.) fruits: composition in 27 cultivars and changes during ripening. Food Chemistry, 132(1), 86-97.

IP

Aharoni, A., Ric de Vos, C., Verhoeven, H. A., Maliepaard, C. A., Kruppa, G., Bino, R., & Goodenowe, D. B. (2002). Nontargeted metabolome analysis by use of Fourier transform ion cyclotron mass spectrometry. Omics: a Journal of Integrative Biology, 6(3), 217-234.

SC R

Aharoni, A., Jongsma M. A., & Bouwmeester H. J. (2005). Volatile science? Metabolic engineering of terpenoids in plants. Trends in Plant Science, 10(12), 594-602.

NU

Allwood, J. W., & Goodacre, R. (2009). An introduction to liquid chromatography–mass spectrometry instrumentation applied in plant metabolomic analyses. Phytochemical Analysis 21(1), 33-47.

MA

lvare -Fern nde , M. ., ere o, . ., a ete- odr ue , A. M., Troncoso, A. M., & García-Parrilla, M. C. (2015). Composition of nonanthocyanin polyphenols in alcoholic-fermented strawberry products using LC–MS (QTRAP), high-resolution MS (UHPLC-Orbitrap-MS), LC-DAD, and antioxidant activity. Journal of Agricultural and Food Chemistry, 63(7), 2041-2051.

D

Beekwilder, J., Alvarez-Huerta, M., Neef, E., Verstappen, F. W., Bouwmeester, H. J., & Aharoni, A. (2004). Functional characterization of enzymes forming volatile esters from strawberry and banana. Plant Physiology, 135(4), 1865-1878.

CE P

TE

Bestfleisch, M., Höfer, M., Richter, K., Hanke, M., Schulte, E., Peil, A., & Flachowsky, H. (2012). Breeding of resistant strawberry cultivars for organic fruit production – Diallel crossing strategies and resistance tests for Botrytis cinerea and Xanthomonas fragariae. In Proceedings of the 15th international conference on organic fruit production (pp. 221-227). Bestfleisch, M., Möhring, J., Hanke, M., Peil, A., & Flachowsky, H. (2014). A diallel crossing approach aimed on selection for ripening time and yield in breeding of new strawberry (Fragaria× ananassa Duch.) cultivars. Plant Breeding, 133(1), 115-120.

AC

Binh, N. T., Ky, P. T., Thao, N. P., Minh, C. V., Cuong, N. X., & Kiem, P. V. (2014). Triterpenes and triterpene-glycoside from the leaves of Lawsonia inermis. Vietnam Journal of Chemistry, 47(4), 511-517. Bondia-Pons, I., Barri, T., Hanhineva, K., Juntunen, K., Dragsted, L. O., Mykkänen, H., & Poutanen, K. (2013). UPLC-QTOF/MS metabolic profiling unveils urinary changes in humans after a whole grain rye versus refined wheat bread intervention. Molecular Nutrition and Food Research, 57(3), 412-422. alani, L., e hè, D., ena, P., Del io, D., runi, ., Fa ri, ., Dall’ sta, C., & Galaverna, G. (2013). Ultra-HPLC–MSn (Poly)phenolic profiling and chemometric analysis of juices from ancient Punica granatum L. cultivars: a nontargeted approach. Journal of Agricultural and Food Chemistry, 61(23), 5600-5609. Carrieri, C., Milella, R.A., Incampo, F., Crupi, P., Antonacci, D., Semeraro, N., & Colucci, M. (2013). Antithrombotic activity of twelve table grape varieties. Relationship with polyphenolic profile. Food Chemistry, 140(4), 647-653 Crupi, P., Pichierri, A., Basile, T.,& Antonacci, D. (2013). Postharvest stilbenes and flavonoids enrichment of table grape cv Redglobe (Vitis vinifera L.) as affected by interactive UV-C exposure and storage conditions. Food Chemistry, 141(2), 802-808.

16

ACCEPTED MANUSCRIPT Crupi, P., Genghi, R., & Antonacci, D. (2014). In-time and in-space tandem mass spectrometry to determine the metabolic profiling of flavonoids in a typical sweet cherry (Prunus avium L.) cultivar from Southern Italy. Journal of Mass Spectrometry 49(10), 1025-1034.

IP

T

Del Rio, D., Rodriguez-Mateos, A., Spencer, J. P., Tognolini, M., Borges, G., & Crozier, A. (2013). Dietary (poly)phenolics in human health: structures, bioavailability, and evidence of protective effects against chronic diseases. Antioxidants & redox signaling, 18(14), 1818-1892. Dictionary of Natural Products. URL http://dnp.chemnetbase.com/. Accessed 9.5.2016

SC R

Eikemo, H., Stensvand, A., Davik, J. & Tronsmo, A. M. (2003). Resistance to crown rot (Phytophthora cactorum) in strawberry cultivars and in offspring from crosses between cultivars differing in susceptibility to the disease. Annals of Applied Biology, 142(1), 83-89.

NU

Eisen, M. B., Spellman, P. T., Brown, P. O., & Botstein, D. (1998). Cluster analysis and display of genome-wide expression patterns. Proceedings of the National Academy of Sciences of the United States of America, 95(25), 14863-14868.

MA

Engström, K. (1998). Sesquiterpenoid spiro compounds from potato tubers infected with Phoma foveata and Fusarium spp. Phytochemistry, 47(6), 985-990.

D

Fait, A., Hanhineva, K., Beleggia, R., Dai, N., Rogachev, I., Nikiforova, V. J., Fernie, A. R., & Aharoni, A. (2008). Reconfiguration of the achene and receptacle metabolic networks during strawberry fruit development. Plant Physiology, 148(2), 730-750.

TE

Goossens Flevoplant. Varieties. URL http://www.flevoplant.nl/eng/rassen.html. Accessed 21.5.2016

CE P

Goossens Flevoplant. Rumba, a shiny start to the summer. URL: http://www.flevoplant.nl/eng/vroeg_2_3427574760.pdf. Accessed 21.5.2016 in ordona a, J., & Terry, L. A. (2009). Development of a glucose biosensor for rapid assessment of strawberry quality: relationship between biosensor response and fruit composition. Journal of Agricultural and Food Chemistry, 57(18), 8220-8226.

AC

Hanhineva, K., Rogachev, I., Kokko, H., Mintz-Oron, S., Venger, I., Kärenlampi, S., & Aharoni, A. (2008). Non-targeted analysis of spatial metabolite composition in strawberry (Fragaria × ananassa) flowers. Phytochemistry, 69(13), 2463-2481. Hanhineva, K., Rogachev, I., Aura, A., Aharoni, A., Poutanen, K., & Mykkänen, H. (2012). Identification of novel lignans in the whole grain rye bran by non-targeted LC–MS metabolite profiling. Metabolomics, 8(3), 399-409. Herderich, M., Richling, E., Roscher, R., Schneider, C., Schwab, W., Humpf, H., & Schreier, P. (1997). Application of atmospheric pressure ionization HPLC-MS-MS for the analysis of natural products. Chromatographia, 45(1), 127-132. Holzwarth, M., Korhummel, S., Carle, R., & Kammerer, D. R. (2012). Evaluation of the effects of different freezing and thawing methods on color, polyphenol and ascorbic acid retention in strawberries (Fragaria×ananassa Duch.). Food Research International 48(1), 241-248. The Human Metabolome Database. URL http://www.hmdb.ca/. Accessed 1.9.2016. Kårlund, A., Moor, U., Sandell, M., & Karjalainen, R. O. (2014b). The impact of harvesting, storage and processing factors on health-promoting phytochemicals in berries and fruits. Processes, 2(3), 596-624.

17

ACCEPTED MANUSCRIPT a rlund, A., Hanhineva, K., Lehtonen, M., Karjalainen, R. O., & Sandell, M. (2015). Nontargeted metabolite profiles and sensory properties of strawberry cultivars grown both organically and conventionally. Journal of Agricultural and Food Chemistry, 63(3), 1010-1019.

IP

T

Kårlund, A., Salminen, J.-P., Koskinen, P., Ahern, J., Karonen, M., Tiilikkala, K., & Karjalainen, R. (2014a). Bioactive polyphenols in strawberry (Fragaria x ananassa) leaves induced by plant activators. Journal of Agricultural and Food Chemistry, 62(20), 4592-4600

SC R

Kerwin, J. L., Wiens, A. M., & Ericsson, L. H. (1996). Identification of fatty acids by electrospray mass spectrometry and tandem mass spectrometry. Journal of Mass Spectrometry, 31(2), 184-192. Koehler, G., Wilson, R. C., Goodpaster, J. V., Sonsteby, A., Lai, X., Witzmann, F. A., You, J. S., Rohloff, J., Randall, S. K., & Alsheikh, M. (2012). Proteomic study of low-temperature responses in strawberry cultivars (Fragaria x ananassa) that differ in cold tolerance. Plant Physiology, 159(4), 17871805.

NU

Laugale, V., & Lepse, L. (2007). Research trials on strawberry cultivars in Pūre Horticultural esearch Station (Latvia) during the last 10 years. Sodininkystė Ir Daržininkystė, 26(3), 81-92.

MA

Leis, M., Martinelli, A., & Castagnoli, G. (2012a). Strawberry plant named ‘LINOSA’. Plant Patent Application Publication.

D

Leis, M., Martinelli, A., & Castagnoli, G. (2012b). Strawberry plant named ‘JOLY’. Plant Patent Application Publication.

TE

Leis, M., Martinelli, A., & Castagnoli, G. (2013). Strawberry plant named ‘CAPRI’. Plant Patent Application Publication.

CE P

Leis, M., Martinelli, A., Castagnoli, G., Azzolini, D., Castagnoli, P., & Castagnoli, A. (2012c). Strawberry plant named 'Ischia'. Plant Patent Application Publication. Leis, M., Martinelli, A., Castagnoli, G., Azzolini, D., Castagnoli, P., & Castagnoli, A. (2012d). Strawberry plant named 'Dely'. Plant Patent Application Publication.

AC

Li, E., Luo, J., & Kong, L. (2009). Qualitative and quantitative determination of seven triterpene acids in Eriobotrya japonica Lindl. by high‐performance liquid chromatography with photodiode array detection and mass spectrometry. Phytochemical Analysis, 20(4), 338-343. Mayorga, H., Duque, C., Knapp, H., & Winterhalter, P. (2002). Hydroxyester disaccharides from fruits of cape gooseberry (Physalis peruviana). Phytochemistry, 59(4), 439-445. MEIOSIS. Junebearer strawberries, Florence. URL http://www.meiosis.co.uk/fruit/florence.htm. Accessed 21.5.2106 Mena, P., Calani, L., Dall'Asta, C., Galaverna, G., García-Viguera, C., Bruni, R., Crozier, A., & Del Rio, D. (2012). Rapid and comprehensive evaluation of (poly) phenolic compounds in pomegranate (Punica granatum L.) juice by UHPLC-MSn. Molecules, 17(12), 14821-14840. Meulenbroek, E. J. (2007). Strawberry plant named ‘Sonata’. Plant Patent Application Publication. The METLIN Metabolite Database. URL http://metlin.scripps.edu/. Accessed 1.9.2016. Musacchi, D., & Leis, M. (1994). Strawberry plant named Onebor. Plant Patent Application Publication. NIST (National Institute of Standards and Technology) Chemistry WebBook. URL http://webbook.nist.gov/chemistry/. Accessed 1.9.2016.

18

ACCEPTED MANUSCRIPT Ornelas-Paz, J. d. J., Yahia, E. M., Ramírez-Bustamante, N., Pérez-Martínez, J. D., Escalante-Minakata, M. d. P., Ibarra-Junquera, V., Acosta-Muñiz, C., Guerrero-Prieto, V., & Ochoa-Reyes, E. (2013). Physical attributes and chemical composition of organic strawberry fruit (Fragaria x ananassa Duch Cv. Albion) at six stages of ripening. Food Chemistry, 138(1), 372-381.

IP

T

Perez A. G., Olías, R., Luaces, P. & Sanz, C. (2002). Biosynthesis of strawberry aroma compounds through amino acid metabolism. Journal of Agricultural and Food Chemistry, 50(14), 4037-4042.

SC R

Rosenfeld, H. J., & Nes, A. (2000). Prediction of sensory quality of strawberry jam by means of sensory quality attributes of fresh fruits. Journal of the Science of Food and Agriculture, 80(13), 1895-1902. SciFinder. URL https://scifinder.cas.org. Accessed 9.5.2016

NU

Soares, S., Kohl, S., Thalmann, S., Mateus, N., Meyerhof, W., & De Freitas, V. (2013). Different phenolic compounds activate distinct human bitter taste receptors. Journal of Agricultural and Food Chemistry, 61(7), 1525-1533.

MA

Soukas, A., Cohen, P., Socci, N. D., & Friedman, J. M. (2000). Leptin-specific patterns of gene expression in white adipose tissue. Genes & Development, 14(8), 963-980. Stevenson, D. E., Parkar, S. G., Zhang, J., Stanley, R. A., Jensen, D. J., & Cooney, J. M. (2007). Combinatorial enzymic synthesis for functional testing of phenolic acid esters catalysed by Candida antarctica lipase B (Novozym 435®). Enzyme and Microbial Technology, 40(5), 1078-1086.

TE

D

Tohge, T., Alseekh, S., & Fernie, A. R. (2014). On the regulation and function of secondary metabolism during fruit development and ripening. Journal of Experimental Botany, 65(16), 3599-4611.

CE P

Ulrich, D., & Olbricht, K. (2013). Diversity of volatile patterns in sixteen Fragaria vesca L. accessions in comparison to cultivars of Fragaria× ananassa. Journal of Applied Botany and Food Quality, 86(1), 3746. Wang, H., Cao, G., & Prior, R. L. (1997). Oxygen radical absorbing capacity of anthocyanins. Journal of Agricultural and Food Chemistry, 45(2), 304-309.

AC

Yue, C., Gallardo, R. K., Luby, J., Rihn, A., McFerson, J. R., McCracken, V., Whitaker, V. M., Finn, C. E., Hancock, J. F., & Weebadde, C. (2014). An evaluation of US strawberry producers trait prioritization: evidence from audience surveys. HortScience, 49(2), 188-193. Zeng, G., Xiao, H., Liu, J., & Liang, X. (2006). Identification of phenolic constituents in Radix Salvia miltiorrhizae by liquid chromatography/electrospray ionization mass spectrometry. Rapid Communications in Mass Spectrometry, 20(3), 499-506.

19

ACCEPTED MANUSCRIPT Figure Captions

T

Figure 1. Principal component analysis (PCA) of the metabolite contents of 15 strawberry

IP

cultivars grown either in Finland or Estonia. The PCA plot shows differences between

SC R

strawberry sample replicates according to their metabolite profiles based on metabolite-specific signal abundances. Each replicate is a pool of 10 strawberry fruits. Closely clustered samples

NU

grown in Estonia are indicated with a dashed line circle. t[1], PC1; t[2], PC2

Figure 2. A tentative pedigree of strawberry cultivars Capri (Ca), Clery (Cl), Dely (De),

MA

Florence (Fl), Honeoye (Ho), Ischia (Is), Joly (Jo), Jonsok (Jn), Linosa (Li), Marmolada (Ma), Polka (Po), Rumba (Ru), Salsa (Sa), and Sonata (So). Cultivars that were studied are presented

D

in white and blue squares and arranged according to hierarchical clustering (HCL) analysis

TE

based on the normalized signal abundances of 1569 strawberry molecular features across all analyzed samples. Parent cultivars Valentine (Val), Senga Sengana (Sen), Induka (Ind), Vibrant

CE P

(Vib), Gorella (Gor), Holiday (Hol), CIVRI30 (Civ), selection A20-17 (A20-17), selection T2-6 (T2-6), and Elsanta (Els) are presented in circles. Arrows indicate the relatedness between

AC

cultivars. Producers of the cultivars that were studied are presented in black-bordered squares on top of the HCL dendrogram. AAFC, Agriculture and Agri-Food Canada; NYSAES, New York State Agricultural Experiment Station; NMBU, Norwegian University of Life Sciences

Figure 3. Heat map representation of the k-means clustering analysis of strawberry sample replicates. Samples (in columns) represent 15 strawberry cultivars. The dendrogram was produced with hierarchical clustering analysis based on the normalized signal abundances of 1569 strawberry molecular features across all analyzed samples. Potential metabolites having similar accumulation patterns are classified in clusters (1-15). The color-coding scale indicates the relative abundance within each molecular feature: blue: low abundance, red: high

20

ACCEPTED MANUSCRIPT abundance, green/yellow: average abundance. Three biological sample replicates for each

T

sample type were analyzed. A replicate consists of 10 strawberry fruit.

IP

Figure 4. The mean peak area abundance values (+SD) of A) anthranilate derivatives, B)

SC R

quercetin glucuronide, C) taxifolin pentoside, and D) terpenoid derivatives demonstrating

AC

CE P

TE

D

MA

NU

cultivar-dependent distribution in 15 strawberry cultivars.

21

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

Figure 1

22

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

Figure 2

23

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

Figure 3

24

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

Figure 4

25

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

Graphical abstract

26

ACCEPTED MANUSCRIPT Highlights

T

IP

SC R NU MA D



TE

 

CE P



strawberry cultivars were distinguished on the basis of their metabolic phenotypes genetically related strawberries showed resemblances in their metabolite profiles cultivar Jonsok contained high levels of potential aroma precursor compounds cultivar Polka was relatively rich in pelargonidin derivatives cultivar Rumba showed high levels of flavan-3-ol derivatives

AC



27