Phytochemistry 103 (2014) 76–84

Contents lists available at ScienceDirect

Phytochemistry journal homepage: www.elsevier.com/locate/phytochem

Rapid estimation of the oxidative activities of individual phenolics in crude plant extracts Matti Vihakas a,⇑, Maija Pälijärvi a, Maarit Karonen a, Heikki Roininen b, Juha-Pekka Salminen a a b

Laboratory of Organic Chemistry and Chemical Biology, Department of Chemistry, University of Turku, FI-20014, Finland Department of Biology, University of Eastern Finland, P.O. Box 111, FI-80101 Joensuu, Finland

a r t i c l e

i n f o

Article history: Received 30 September 2013 Received in revised form 17 January 2014 Available online 29 April 2014 Keywords: Oxidative stress Pro-oxidative defence Autoxidation Folivory Midgut Lepidopteran larvae Hydrolysable tannins Myricetin Flavonoid glycosides

a b s t r a c t Previous studies of purified phenolic compounds have revealed that some phenolics, especially ellagitannins, can autoxidise under alkaline conditions, which predominate in the midgut of lepidopteran larvae. To facilitate screening for the pro-oxidant activities of all types of phenolic compounds from crude plant extracts, we developed a method that combined our recent spectrophotometric bioactivity method with an additional chromatographic step via UPLC–DAD–MS. This method allowed us to estimate the total pro-oxidant capacities of crude extracts from 12 plant species and to identify the individual phenolic compounds that were responsible for the detected activities. It was found that the pro-oxidant capacities of the plant species (i.e., the concentrations of the easily-oxidised phenolics) varied from 0 to 57 mg/g dry wt, representing from 0% to 46% of the total phenolics from different species. UPLC–DAD–MS analysis revealed that most flavonol and flavone glycosides were only slightly affected by alkaline conditions, thus indicating their low pro-oxidant activity. Interestingly, myricetin-type compounds differed from the other flavonoids, as their concentrations decreased strongly due to alkaline incubation. The same effect was detected for hydrolysable tannins and prodelphinidins, suggesting that a pyrogallol sub-structure could be a key structural component that partially explains their easy oxidation at high pH. Other types of phenolic compounds, such as hydroxycinnamic acids, were relatively active, as well. These findings demonstrate that this method displays the potential to identify most of the active and inactive pro-oxidant phenolic compounds in various plant species. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Plants produce a wide variety of secondary compounds in their various tissues. One of the many functions of these secondary compounds is defence against herbivores. In some cases, the original secondary compound may require chemical activation or modification to perform its intended activity against herbivores. These compounds include cyanogenic glucosides (Morant et al., 2008; Zagrobelny et al., 2004), glucosinolates (Morant et al., 2008), iridoid glycosides (Pankoke et al., 2012), and some phenolic compounds (Klocke et al., 1986; Salminen and Karonen, 2011; Summers and Felton, 1994). Phenolic compounds are a diverse group of secondary plant compounds, some of which may be oxidatively activated to function against herbivores (Appel, 1993). Phenolics can be oxidised enzymatically or by autoxidation to produce quinones and other oxidation products that can harm the growth of herbivores (Appel, 1993; Felton et al., 1989; ⇑ Corresponding author. Tel.: +358 2 333 6757; fax: +358 2 333 6700. E-mail address: matti.vihakas@utu.fi (M. Vihakas). http://dx.doi.org/10.1016/j.phytochem.2014.02.019 0031-9422/Ó 2014 Elsevier Ltd. All rights reserved.

Summers and Felton, 1994). Autoxidation has been reported for many classes of phenolic compounds, such as flavonoids (Cao et al., 1997) and caffeic acids (Cilliers and Singleton, 1989, 1991). The autoxidation of phenolics is promoted by many factors, such as the alkaline conditions prevalent in the gut of many insect herbivores (Appel, 1993). For example, the midgut of lepidopteran larvae can be highly basic, with a pH as high as 10–12 (Dow, 1984; Gringorten et al., 1993). Previous studies have examined the autoxidation of phenolic compounds by measuring semiquinone radicals that are generated from phenolics in the gut of lepidopteran larvae (Barbehenn et al., 2003, 2005, 2008). The oxidation of phenolics was also studied via in vitro experiments that mimicked the alkaline conditions prevalent in the midgut of lepidopterans (Barbehenn et al., 2006; Moilanen and Salminen, 2008; Tuominen, 2013; Tuominen and Sundman, 2013). The studies described above demonstrated that different classes of phenolics or individual phenolic compounds have variable abilities to function as pro-oxidants in the caterpillar gut. Salminen and Karonen (2011) presented a new method that enabled the comparison of pro-oxidant capacities of different plant

M. Vihakas et al. / Phytochemistry 103 (2014) 76–84

samples. Their method is based on several analytical steps. First, the total phenolics in the plant samples are measured via a modified Folin–Ciocalteu assay. Next, the samples are incubated in an alkaline buffer at pH 10, which oxidises some of the phenolics in the plant samples. The oxidised samples are then measured via the modified Folin–Ciocalteu assay, and the results are compared to those of the non-oxidised samples. The results reveal the quantity of the total phenolics that was oxidised during the incubation. Their method estimates the pro-oxidant capacity of a given plant sample, and it can be used to compare the pro-oxidant capacities of different plant species. This method, however, is limited in that it does not identify which specific phenolic compounds are responsible for the detected oxidative capacities of the plant species. For instance, different samples could produce similar pro-oxidant capacities, even though the samples contain distinct compositions of phenolic compounds. Based on these limitations, we further developed the pro-oxidant method of Salminen and Karonen (2011). We combined the conventional spectrophotometric methods with ultra-high performance liquid chromatography coupled to diode array detection and mass spectrometry (UPLC–DAD–MS) to identify the individual phenolic compounds that are oxidised under alkaline conditions. To do so, we analysed both the initial and oxidised plant extracts from 12 selected plant species via UPLC–DAD–MS. These analyses revealed which types of phenolics were responsible for the low or high prooxidant activities of each tested plant species. Our results emphasise the capacity of the improved methodology to elucidate the prooxidant defence of various plant species. Moreover, these analyses were performed directly from crude plant extracts, which means that the time-consuming purification of secondary compounds performed in earlier studies (Barbehenn et al., 2006; Moilanen and Salminen, 2008; Tuominen and Sundman, 2013) is unnecessary. Our new method could facilitate the detection of the pro-oxidant capacities of multiple plant species and the identification of the types of phenolic compounds responsible for the detected activities. Importantly, this method will help us to understand the types of phenolics that appear to perform anti-herbivore activities in different plant species. 2. Results Different plant species contained varying concentrations of total phenolics (from 20 mg/g to 125 mg/g dry wt). The oxidation measurements revealed that approximately 0 to 57 mg/g dry wt of the initial total phenolics were lost at pH 10 (Figs. 1 and 2). In some cases, the oxidised proportions of the total phenolics represented only 0–5% of the initial total phenolics, but in other cases, up to 46% of the initial total phenolics were lost due to oxidation. We separated the 12 examined plant species in three categories according to their pro-oxidant capacities: plants with weak, moderate or high oxidative capacity. 2.1. Plants with weak oxidative capacity In the weak oxidative capacity category, we placed the plant species whose total phenolic concentrations changed only slightly after incubation at pH 10. This result suggested that the composition of the phenolic compounds remained fairly constant during the incubation in alkaline conditions. UPLC–DAD chromatograms of the initial and oxidised samples are presented in Fig. 1A–C. The chromatographic profiles of the initial and oxidised samples were very similar, and the intensities of the individual peaks typically remained at the same level in both samples. A negligible amount of total phenolics in Celtis durandii Engl. (hackberries) were oxidised under alkaline conditions, and a similar trend was found

77

in the UPLC data: two predominant compounds in the leaves were unaffected by alkaline conditions (Fig. 1A). These two compounds were classified as apigenin diglycosides with C-glycosyls based on their UV spectra and MS/MS fragmentation patterns (Table 1; Ferreres et al., 2007). Similar types of compounds were previously found in the leaves of Celtis plants (Kaltenhauser et al., 2010; Spitaler et al., 2009). Only 5% of the total phenolics in Aesculus hippocastanum L. (horse chestnut) were oxidised at pH 10 (Fig. 1B). The chromatographic profiles of the initial and oxidised samples were similar, but the intensities of some individual peaks were lower in the chromatogram of the oxidised sample. A hump in the middle of the chromatogram was interpreted as oligo- to polymeric proanthocyanidins (Karonen et al., 2004), and the calculated mass spectra from the hump area in the total ion chromatogram revealed representative peaks of A-type procyanidin oligomers (Table 1; Morimoto et al., 1987). The procyanidin hump was still present after the incubation in alkaline conditions, and the calculated mass spectra of this area did not change, which suggested that the procyanidins in the sample were resistant to oxidation at pH 10. A similar hump containing procyanidins was detected in the chromatograms of Picea abies (L.) Karst (Norway spruce) needles (Fig. 1C), and the chromatograms also contained peaks of catechin and a kaempferol monoglycoside (Table 1; Slimestad and Hostettmann, 1996). The analysis of the total phenolics suggested that 13% of the total phenolics of P. abies were lost during the incubation. Alkaline conditions caused a slight decrease in the size of the proanthocyanidin hump, even though the calculated mass spectra of the hump area still contained representative peaks of oligomeric procyanidins. 2.2. Plants with moderate oxidative capacity The moderate oxidative capacity category of plants contained species whose phenolic compounds appeared to be slightly more prone to autoxidation under alkaline conditions than the plant species in the weak category (Fig. 1D–F). For example, 22% of the total phenolics in the leaves of Prunus africana (Hook.f.) Kalkman (African cherry) were oxidised at pH 10. The chromatographic profile of the P. africana extract changed during the incubation: the height of the peak corresponding to monocaffeoylquinic acid (Nakatani et al., 2000; Ossipov et al., 1996) decreased substantially, and two new peaks appeared on each side of the original peak (Fig. 1D). The new compounds displayed UV and mass spectra similar to those of the original compound (Table 1), which suggested that they were all isomers of monocaffeoylquinic acid. The isomerisation of monocaffeoylquinic acid was also detected in Achillea ptarmica L. (sneezewort; Fig. 1E). In addition to this compound, A. ptarmica contained a dicaffeoylquinic acid, and the height of the corresponding peak decreased substantially, but it did not form isomers under the alkaline conditions. A flavone (an apigenin glucuronide) in the same species remained unaffected by alkaline conditions. The leaves of Oxyanthus speciosus DC. (whipstick-loquat) displayed a medium level of pro-oxidant capacity based on the total phenolics assay, and the predominant compound based on the UPLC chromatogram was unaffected by the incubation (Fig. 1F). This compound was identified as a cyanogenic glycoside, and these types of compounds have previously been detected in the Oxyanthus genus (Rockenbach et al., 1992). The leaves also contained several isomers of mono- and di-caffeoylquinic acids, and one of the isomers of monocaffeoylquinic acid partially co-eluted with the cyanogenic glycoside (as detected on the extracted ion chromatogram of monocaffeoylquinic acid). The heights of the peaks of monocaffeoylquinic acids decreased considerably after the incubation, and the peaks of the dicaffeoylquinic acids disappeared

78

M. Vihakas et al. / Phytochemistry 103 (2014) 76–84 5

Initial 20 mg/g

5

2.5x10

5

2.0x10

AP2

Initial 79 mg/g

5

2.0x10

5

PC1 PC2

5

Intensity (AU)

5

1.0x10

4

5.0x10

0.0 5

2.5x10

pH 10 20 mg/g ~0%

5

2.0x10

5

1.5x10

AP2 AP1

B

QU2 PC3 QU1

1.5x10

AP1

5

1.5x10

Intensity (AU)

2.5x10

A

1.0x10

4

5.0x10

0.0 5

2.5x10

pH 10 75 mg/g −5%

5

2.0x10

5

1.5x10

5

1.0x10

4

5.0x10

QU2 PC3 QU1

5

1.0x10

PC1 PC2

4

5.0x10

0.0

0.0 0

1

2

3

4

5

6

7

0

1

2

C

Initial 45 mg/g

5

1.5x10

3

4

5

6

7

Retention time (min)

Retention time (min) 5

Initial 54 mg/g

5.0x10

KA1

5

4.0x10

D

CA1

5 5

3.0x10

PC4

4

Intensity (AU)

Intensity (AU)

1.0x10 5.0x10

0.0 5

1.5x10

pH 10 39 mg/g − 13 %

5

1.0x10

KA1

5

2.0x10

5

1.0x10

0.0 5 5.0x10

pH 10 42 mg/g − 22 %

5

4.0x10

5

3.0x10

CA1

5

4

2.0x10

PC4

5.0x10

5

1.0x10 0.0

0.0 0

1

2

3

4

5

6

7

0

1

2

R t ti Retention time ti ((min) i ) Initial 27 mg/g

Intensity (AU)

2.0x10

6.0x10

AP3

CA1

5

5

E

CA2

5

pH 10 21 mg/g − 22 %

5

2.0x10

AP3

5

7

CA2

0.0 5 6.0x10

CNG1

pH 10 50 mg/g − 23 %

5

4.0x10

5

6

CA1

5

CA2

F

2.0x10

CA1 1.0x10

5

CA1

4.0x10

1.0x10

0.0 5 3.0x10

4

CNG1

Initial 65 mg/g

5

Intensity (AU)

5

3.0x10

3

R t ti Retention ti time ((min) i )

2.0x10

0.0

0.0 0

1

2

3

4

5

6

7

Retention time (min)

0

1

2

3

4

5

6

7

Retention time (min)

Fig. 1. The UPLC–DAD chromatograms of plants with weak to moderate oxidative capacities (measured at 280 nm). ‘‘Initial’’ shows the chromatogram of the original plant extract and ‘‘pH 10’’ represents the same extract after incubation at pH 10. The amounts of total phenolics in extracts before and after incubation are expressed as mg/g dry wt. The percentage value shows how large portion of the initial total phenolics were lost during the incubation. Studied plant species: A: Celtis durandii, B: Aesculus hippocastanum, C: Picea abies, D: Prunus africana, E: Achillea ptarmica, F: Oxyanthus speciosus. Abbreviated compounds are listed in Table 1.

from the chromatograms, which suggested that their partial oxidation most likely comprises the oxidative capacity of this plant species. 2.3. Plants with high oxidative capacity The high oxidative capacity category contained plant species whose phenolics were most affected by alkaline conditions (Fig. 2). The leaves of Iris sp. (flags) and Ribes alpinum L. (alpine currant) contained glycosides of myricetin (Fig. 2A and B; Gluchoff-Fiasson et al., 2001). Interestingly, myricetin glycosides appeared to be very efficiently oxidised at high pH. Their oxidation might have caused the strong reduction in total phenolics in these species. The leaves of Iris sp. also contained flavones and other flavonoids (Williams et al., 1986) that were substantially less

affected (Fig. 2A). The UPLC chromatogram of R. alpinum displayed a distinct hump between the retention times of 2–4 min (Fig. 2B), and the calculated mass spectra from this area contained ions derived from oligomeric prodelphinidins (Table 1; Tits et al., 1992). Interestingly, this hump disappeared after the incubation in alkaline conditions, and the prodelphinidin ions were absent from the calculated mass spectra of the hump area. The leaves of Salix phylicifolia L. (tea-leaved willow) exhibited only one major peak in the UPLC–DAD chromatogram, and this compound, dihydromyricetin (Rank et al., 1998), was almost completely depleted in the oxidised samples (Fig. 2C). The leaves of Fragaria moschata Duchesne (musk strawberry), Betula pubescens Ehrh. (white birch) and Quercus robur L. (pedunculate oak) contained different types of hydrolysable tannins (ellagitannins and galloyl glucoses; Fig. 2 D–F; Table 1). An identical

79

M. Vihakas et al. / Phytochemistry 103 (2014) 76–84

Initial 56 mg/g

5

1.5x10

5

4

6.0x10

QU2

PC4

5.0x10

0.0 5

pH 10 38 mg/g − 32 %

5

1.0x10

PD1

4

4.0x10

Intensity (AU)

4

4

5

1.0x10

pH 10 45 mg/g − 46 %

4

8.0x10

4

4.0x10

QU2

5.0x10

4

2.0x10

0.0

0.0 0

1

2

3

4

5

6

7

0

1

2

6

2.5x10

Initial 83 mg/g

6

2.0x10

C

MY3

6

4

5

6

7

5

8.0x10

6

D

ET1

Initial 105 mg/g

1.0x10

5

1.5x10

6.0x10

6

5

1.0x10

Intensity (AU)

Intensity (AU)

3

Retention time (min)

Retention time (min)

5

5.0x10

0.0 6 2.5x10

pH 10 54 mg/g − 35 %

6

2.0x10

6

1.5x10

4.0x10

5

PC4

2.0x10

0.0 6

1.0x10

pH 10 81 mg/g − 23 %

5

8.0x10

5

6.0x10

6

4.0x10

5

5

2.0x10

1.0x10

EA

5

5.0x10

0.0

0.0 0

1

2

3

4

5

6

7

0

1

2

CA1

Initial 63 mg/g

5

1.6x10

E

1.2x10

GG1

5

5

1.2x10

ET6

Intensity (AU)

5

0.0

pH 10 38 mg/g − 40 %

CA1

4

8.0x10

EA CA3

kaempferol and quercetin glycosides

2.0x10

GG1

5

kaempferol and quercetin glycosides

ET3 ET5

5

EA kaempferol and quercetin glycosides

5

2.0x10

5

0.0

GG1

0.0 3

4

5

6

7

ET8

pH 10 68 mg/g − 46 %

5

4.0x10

1.0x10

2

F

0.0 5.0x10

5

4

1

7

1.0x10

3.0x10

4.0x10

0

6

ET7

5

4

5

5

ET4

3.0x10

4.0x10

1.6x10

4

4.0x10

ET2

4

8.0x10

Initial 125 mg/g

5

5.0x10

kaempferol and CA3 quercetin glycosides GG2 ET2

5

3

R t ti Retention time ti (min) ( i )

R t ti Retention ti time ((min) i )

Intensity (AU)

MY1 PD3

0.0

6.0x10

miscellaneous flavonoids

PD2

2.0x10

4

4

B

MY2

4

miscellaneous flavonoids

1.5x10

Initial 83 mg/g

5

1.0x10 8.0x10

1.0x10

Intensity (AU)

A

MY1

0

1

2

3

4

5

6

7

Retention time (min)

Retention time (min)

Fig. 2. The UPLC–DAD chromatograms of plants with high oxidative capacities. The conditions are the same as in Fig. 1. Studied plant species: A: Iris sp., B: Ribes alpinum, C: Salix phylicifolia, D: Fragaria moschata, E: Betula pubescens, F: Quercus robur. Abbreviated compounds are listed in Table 1.

pattern was detected in their UPLC DAD chromatograms, as the peaks of hydrolysable tannins almost completely disappeared after the incubation in alkaline conditions. A new peak at the retention time of 0.5 min appeared in the chromatograms of the oxidised samples containing hydrolysable tannins. This peak displayed UV absorbance at approximately 200 nm but no clear mass spectrum. This peak was eluted near the peak of dead volume of the system, which suggested that the peak may have consisted of degradation products of hydrolysable tannins that were weakly retained in the column. Another new peak at the retention time of 3.9 min appeared in the chromatograms of the oxidised samples, and the corresponding compound was interpreted as ellagic acid, a hydrolysis product of ellagitannins (Daniel et al., 1991; Tuominen and Sundman, 2013). Monocaffeoylquinic acid in B. pubescens (Fig. 2E) displayed a similar isomerisation pattern at high pH to that which was noted above for other plant species (Fig. 1D and E). Again, flavonol glycosides (kaempferol and quercetin

glycosides) appeared to be resistant to alkaline conditions, and their peaks remained clearly detectable in the chromatograms after the incubation in alkaline conditions (Fig. 2E and F).

3. Discussion The presented new method estimated that the studied 12 plant species display markedly different pro-oxidant capacities under alkaline conditions. For example, our data suggested that the phenolic compounds of some plant species (e.g., C. durandii and A. hippocastanum) possess a low capacity to function as a pro-oxidant defence in the alkaline midgut of lepidopteran larvae, while the phenolics of other plant species (e.g., S. phylicifolia and Q. robur) possess a higher capacity to generate degenerative oxidative products under alkaline gut conditions. The UV and mass spectra of the phenolic compounds enabled the identification of individual

80

M. Vihakas et al. / Phytochemistry 103 (2014) 76–84

Table 1 UV and mass spectral data of compounds marked in the Figs. 1 and 2. Abbreviation

Compound name

Plant species

UV spectrum k max (nm)

[M-H]

m/z values of other ions

References

AP1

Apigenin diglycoside

Cel

268, 336

593

a

Ferreres et al. (2007)

a

Ferreres et al. (2007)

AP2

Apigenin diglycoside

Cel

268, 335

563

AP3 CA1 CA2 CA3 CNG1 EA ET1 ET2 ET3 ET4

Apigenin glucuronide Monocaffeoylquinic acid Dicaffeoylquinic acid Coumaroylquinic acid Cyanogenic glycoside Ellagic acid Agrimoniin Pedunculagin derivative Vescavaloninic acid Vescalagin

Ach Pru, Ach, Oxy, Bet Ach, Oxy Bet Oxy Fra, Bet, Que Fra Bet Que Que

266, 249, 245, 230, 223, 253, 229, 226, 226, 229,

336 300 (sh), 324 300 (sh), 327 310 278 367 254 (sh) 275 260-280 (sh) 260-280 (sh)

445 353 515 337 310 301

ET5 ET6 ET7

Castalagin Ellagitannin Cocciferin D2

Que Que Que

227, 260-280 (sh) 222, 260-280 (sh) 228, 260-280 (sh)

933 953

ET8

Ellagitannin

Que

227, 282 (sh)

1083

GG1 GG2 KA1 MY1 MY2 MY3 PC1 PC2 PC3 PC4 PD1 PD2 PD3 QU1 QU2

Monogalloyl glucose Digalloyl glucose Kaempferol hexoside Myricetin deoxyhexoside Myricetin malonyl deoxyhexoside Dihydromyricetin ( )-Epicatechin Procyanidin trimer (A type)b Procyanidin dimer (A type)b (+)-Catechin Prodelphinidin trimer (B type)b Gallocatechin Prodelphinidin tetramer (B type)b Quercetin pentoside Quercetin deoxyhexoside

Bet, Que Bet Pic Iri, Rib Rib Sal Aes Aes Aes Pic, Iri, Fra Rib Rib Rib Aes Aes, Iri

216, 221, 265, 257, 255, 232, 203, 279 279 203, 207, 205, 207, 255, 255,

331 483 447 463 549 319 289 863 575 289 913 305 1217 433 447

277 276 347 349 344 291 279

279 275 275 275 352 349

787 1101 933

413 [M-180] , 293 [api+41-H2O] 413 [M-150] , 293 [api+41-H2O] 269 [M-glyc-H] 191 [M-caffeoyl] 353 [M-caffeoyl]

934 [M-2H]2 769 [M-H2O-H] 1083 [M-H2O-H] 915 [M-H2O-H] , 466 [M-2H]2 466 [M-2H]2 476 [M-2H]2 933 [M-2H]2

Ossipov et al. (1996) Ossipov et al. (1996) Rockenbach et al. (1992) Moilanen and Salminen (2008) Moilanen and Salminen (2008) Salminen et al. (2002) Yarnes et al. (2006) Salminen et al. (2004a) Salminen et al. (2004a) Salminen et al. (2004a), Moilanen and Salminen (2008)

1065 [M-H2O-H] , 541 [M-2H]2

285 [M-glyc-H] 317 [M-glyc-H] 505 [M-COOH]

Salminen et al. (1999) Salminen et al. (1999) Slimestad and Hostettmann (1996) Gluchoff-Fiasson et al. (2001) Gluchoff-Fiasson et al. (2001) Rank et al. (1998) Morimoto et al. (1987) Morimoto et al. (1987) Tits et al. (1992) Tits et al. (1992) Tits et al. (1992)

301 [M-glyc-H] 301 [M-glyc-H]

Abbreviations: api = apigenin aglycone, glyc = a glycosyl group, sh = shoulder; Plant species: Ach = Achillea ptarmica, Aes = Aesculus hippocastanum, Bet = Betula pubescens, Cel = Celtis durandii Fra = Fragaria moschata, Iri = Iris sp., Oxy = Oxyanthus speciosus, Pic = Picea abies, Pru = Prunus africana, Rib = Ribes alpinum, Sal = Salix phylicifolia, Que = Quercus robur. a MS/MS fragments of the precursor ions at m/z 593 or 563. b B type proanthocyanidins have one inter-flavonoid bond between two monomer units, but A type proanthocyanidins have two.

phenolic compounds responsible for the pro-oxidant capacities measured via the total phenolics assay. For example, the total phenolics assay indicated that O. stenocarpus and F. moschata exhibit approximately similar pro-oxidant capacities (Figs. 1F and 2D). However, the chromatographic data further clarified that the prooxidant capacities of these two plant species arose from two different types of phenolic compounds, i.e., caffeoylquinic acids in O. stenocarpus and ellagitannins in F. moschata. Previous studies have used several different analytical methods to examine the oxidation of phenolic compounds. These methods include electron paramagnetic resonance spectrometry (Barbehenn et al., 2006, 2008; Bors et al., 2000; Kuhnle et al., 1969), a spectrophotometric browning assay (Barbehenn et al., 2006; Moilanen and Salminen, 2008; Tuominen, 2013) and enzymatic degradation assays of phenolic compounds (Bors et al., 2000; Koclar Avci et al., 2013). Many of our results using our method, which combines the spectrophotometric total phenolics assay with UPLC–DAD–MS, supported the findings of previous studies, which demonstrates the effectiveness of the new technique. The new method produces similar results but also provides several advantages. Firstly, the use of UPLC shortened the analysis duration and enhanced the peak performance compared to traditional HPLC. Secondly and more importantly, the analyses were performed directly on plant crude extracts, which simplified the analytical procedures compared to previous studies that

performed plant fractionation or purification of phenolic compounds to estimate the pro-oxidant activities of phenolic compounds (Barbehenn et al., 2006; Moilanen and Salminen, 2008). The ability to use crude extracts is especially important for cases in which the isolation of phenolic compounds is not possible due to a limited amount of plant material or a large number of samples. For such cases, our method can be performed using 5 mg or less of dried plant material. Spectrophotometric total sample methods and chromatographic methods function differently, and the results from these two types of analyses should be compared with some caution. The total phenolics assay primarily responds to compounds containing phenolic hydroxyl groups, while UPLC–DAD detects a wide range of compounds that absorb light in the UV/vis spectrum. In many cases, the spectrophotometric and chromatographic data were highly comparable (e.g., Fig. 1A), but occasionally, the similarity between these two types of analyses was not apparent. For example, the UPLC–DAD chromatograms of the initial and oxidised samples of P. abies were rather similar in appearance, but the total phenolics assay suggested that over 10% of the phenolics in this species were prone to oxidation (Fig. 1C). In P. abies, this difference was most likely due to polymeric proanthocyanidins, which were detected as the small chromatographic hump shown in Fig. 1C (3–5 min), and this hump was slightly decreased in the oxidised sample. Because this hump was small and full scan MS does not

M. Vihakas et al. / Phytochemistry 103 (2014) 76–84

efficiently detect polymeric proanthocyanidins, this result could not be verified using more specific MS data. Interestingly, many derivatives of flavonols and flavones were only slightly affected by alkaline conditions (e.g., Fig. 1A–C, E). The stability of a quercetin glycoside in alkaline solution has been detected earlier (Tuominen, 2013), but our study revealed that kaempferol and apigenin derivatives were similarly stable. This result suggested that these flavonoids possess rather low pro-oxidant activities under alkaline conditions. Barbehenn et al. (2008) found that rutin and its aglycone quercetin displayed fairly low oxidative activities compared to some other phenolic compounds, such as chlorogenic acid and gallic acid, or the more predominant ellagitannins. Previous studies have noted that flavonoids can act as pro-oxidants under the appropriate conditions (Cao et al., 1997; Tuominen, 2013), and their autoxidation is accelerated in more alkaline solutions (Canada et al., 1990). The pro-oxidant chemistry of apigenin has been studied with respect to human health (Chan et al., 1999; Miyoshi et al., 2007), but to the best of our knowledge, this is the first time that the pro-oxidant activities of flavones have been studied with respect to plant-herbivore interactions. Our data further indicated that substitution (deoxyhexose, glucuronic acid, hexose or pentose) did not alter the high stability of flavonols or flavones under alkaline conditions. Quercetin aglycone autoxidised, but rutin did not, based on the results of Canada et al. (1990), which implied that glycosylation might render some flavonols more resistant to oxidation. All hydroxycinnamic acids (mono- and dicaffeoylquinic acids and coumaroyl quinic acid) displayed moderate oxidative activities based on the incubation assays, and in particular, dicaffeoylquinic acid was efficiently degraded in the oxidised samples (Fig. 1D–F; Fig. 2E). The autoxidative behaviour of other types of hydroxycinnamic acids, such as caffeic acid, has been reported previously (Cilliers and Singleton, 1989, 1991). Chlorogenic acid (monocaffeoylquinic acid; Barbehenn et al., 2008) and galloyl quinic acids (Tuominen, 2013) displayed pro-oxidant activities under alkaline conditions, whereas the oxidative activity of p-coumaric acid was approximately zero (Barbehenn et al., 2008). Monocaffeoylquinic acid isomerised during our incubation assays (Nagels et al., 1980), but we did not detect the isomerisation of other hydroxycinnamic acids. We compared the area of the initial monocaffeoylquinic acid peak to the areas of three isomeric peaks in the oxidised samples (detected at 324 nm). The combined areas of the isomers comprised ca. 60% of the area of the initial peak. This suggested that ca. 40% of the compound had been oxidised during the incubation. Interestingly, Salminen et al. (2004b) demonstrated that approximately 63% of the foliar chlorogenic acid, i.e., monocaffeoylquinic acid, was absent from the larval frass of the lepidopteran Epirrita autumnata. This finding suggested that the in vitro method could in some cases generate comparable results to those of in vivo methods. Moreover, the isomerisation of chlorogenic acid has been found to take place in the gut of E. autumnata as well as various sawfly species (Lahtinen et al., 2005; Salminen et al., 2004b). The plants containing ellagitannins and galloyl glucoses (hydrolysable tannins) displayed high pro-oxidant capacities based on the incubation assays (Fig. 2D–F). These results support those of previous studies, in which ellagitannins or plant extracts containing ellagitannins possessed high pro-oxidant activities (Barbehenn et al., 2006; Tuominen, 2013). Additionally, specific ellagitannins possess distinct oxidative activities under alkaline conditions (Moilanen and Salminen, 2008). Galloyl glucoses were demonstrated to possess lower oxidative activities than ellagitannins (Barbehenn et al., 2006). In our study, mono- and digalloyl glucoses from B. pubescens were efficiently depleted, but a small peak of monogalloyl glucose remained in the oxidised sample of Q. robur (Fig. 2E and F). In addition to probable oxidation reactions, the

81

hydrolysable tannins were also hydrolysed at pH 10, which was indicated by the presence of ellagic acid in the oxidised samples. A clear exception to the low activity of flavonoid derivatives was myricetin glycosides, whose peaks disappeared from the chromatograms of the oxidised samples (Fig. 2A and B). Myricetin glycosides also displayed a similar trend among the various plant species that were examined in this study (data not shown). Additionally, a related compound to myricetin glycosides, a dihydromyricetin, was degraded in the oxidised sample of S. phylicifolia (Fig. 2C). Another detected dihydroflavonol, a dihydroquercetin glycoside in Pinus sylvestris, remained almost completely intact under alkaline conditions, thus displaying an opposite activity to that of dihydromyricetin (data not shown). Many chemical ecology studies examining the pro-oxidant effects of flavonols have primarily focused on quercetin (e.g., Barbehenn et al., 2008; Duffey and Stout, 1996; Pritsos et al., 1988), but until now, the myricetin-type compounds appear to have received less attention. The myricetintype compounds contain three hydroxyl groups in their flavonoid B ring, whereas the more inactive flavonol and flavone glycosides, such as the kaempferol, quercetin and apigenin glycosides, contain one or two hydroxyl groups in their B ring. This led us to consider that the pyrogallol (trihydroxyl) substitution pattern of the myricetin-type compounds might be underlie their pro-oxidant activity. The autoxidative behaviour of myricetin under alkaline conditions has been reported (e.g., Hodnick et al., 1986), and myricetin aglycone was been found to autoxidise more efficiently than quercetin aglycone (Canada et al., 1990; Cao et al., 1997). Pyrogallol (1,2,3-benzenetriol) has been demonstrated to autoxidise under alkaline conditions (Marklund and Marklund, 1974). In addition to the myricetin derivatives, many other phenolic compounds containing a pyrogallol sub-structure were efficiently depleted during our incubation assays. For example, the oxidatively active ellagitannins contain several pyrogallol subunits within their structures, and the intensities of the chromatographic peaks of the mono- and digalloyl glucoses were also notably decreased by alkaline treatment (Fig. 2E and F). Oligomeric prodelphinidins (a pyrogallol sub-structure) were efficiently oxidised under alkaline conditions (Fig. 2B), whereas oligomeric procyanidins (which contain two OH groups in the B ring) were apparently more resistant to alkaline conditions (Fig. 1B and C). Previous studies have demonstrated that prodelphinidins possess higher prooxidant activities than procyanidins (Tuominen, 2013), and prodelphinidin-rich proanthocyanidins displayed a higher browning rate (i.e., formation of quinones or polymeric pigments) than procyanidins under alkaline conditions (Barbehenn et al., 2006). We also found preliminary data using Chrysophyllum albidum G. Don (white star apple), which contains a gallocatechin gallate (two pyrogallol groups), and this compound was also depleted by alkaline treatment (data not shown). Our data altogether suggest that a pyrogallol sub-structure in many phenolic compounds could be a key structural component that helps to explain the high oxidative activities of many different types of phenolic compounds. The oxidation assays described above were performed in a laboratory environment and can thus provide only suggestions regarding the actual chemical reactions that occur inside a living caterpillar. Several factors likely affect the fate of phenolics in vivo. For example, the pH of the midgut in lepidopteran larvae varies between species from mildly alkaline to up to pH 12 (Berenbaum, 1980; Dow, 1984; Gross et al., 2008). The autoxidation of phenolic compounds is often faster in more alkaline environments (Cilliers and Singleton, 1989; Hodnick et al., 1986; Tuominen and Sundman, 2013). This environmental difference most likely causes the rates of autoxidation to vary between different lepidopteran species. In addition, our oxidation assays were performed at an ambient pressure of oxygen, whereas the midgut of lepidopteran larvae are often nearly anoxic (Johnson and

82

M. Vihakas et al. / Phytochemistry 103 (2014) 76–84

Barbehenn, 2000). The autoxidation of phenolics appears to progress more slowly when the quantity of oxygen is reduced (Cilliers and Singleton, 1989; Gross et al., 2008; Hodnick et al., 1986). However, these anoxic conditions may only occur in the centre of the midgut, and oxygen can permeate into the perimeter of the midgut (Gross et al., 2008), which would enable faster oxidation reactions in this portion of the midgut. In addition to phenolic compounds, the insect gut contains numerous other chemicals that can affect the stability of phenolics. These compounds include plant derived enzymes (e.g., polyphenol oxidases and peroxidases), which oxidise phenolics inside the larvae (Felton et al., 1989; Barbehenn et al., 2010), and antioxidants (e.g., ascorbates), which can prevent the oxidation of phenolic compounds (Barbehenn et al., 2001, 2003). 4. Conclusions The oxidation method presented here elucidated the oxidative capacities of plant crude extracts and identified the individual phenolic compounds that are most likely responsible for the detected oxidative activities. This new method was time-saving because fractionation of plant extracts or isolation of phenolic compounds was unnecessary to perform the analyses. The use of UPLC enhanced the peak performance, shortened the analysis duration, and reduced solvent consumption compared to traditional HPLC. New information was found regarding the oxidative activities of different types of phenolic compounds, such as flavones, hydroxycinnamic acids, and pyrogallol-containing phenolics. In particular, ellagitannins have been identified as compounds that produce oxidative stress in lepidopteran larvae (Barbehenn et al., 2008; Salminen and Karonen, 2011), and plants containing these compounds produced high levels of semiquinone radicals in the midgut of a generalist caterpillar (Barbehenn et al., 2008). It would be interesting to determine whether myricetin-type compounds could also produce similar effects in these larvae. The low pro-oxidant capacities of some plant species suggest that these plant species might utilise other types of chemical defence against herbivores. In the future, the oxidation method described here could be utilised to screen numerous plant crude extracts for their potent pro-oxidant phenolic compounds. This information would be crucial to better understand the distribution of pro-oxidant phenolics in the plant kingdom. Moreover, it would be important to reveal the potent pro-oxidant compounds in plant species so that their concentration rather than that of total phenolics could be correlated to insect performance. 5. Experimental 5.1. General experimental procedures The UPLC system (Acquity UPLCÒ, Waters Corporation, Milford, USA) consisted of a sample manager, a binary solvent manager, and a photodiode array detector. The column was an Acquity UPLCÒ BEH Phenyl column (2.1  100 mm, 1.7 lm, Waters Corporation, Wexford, Ireland). Two solvents, 0.1% formic acid (solvent A) and acetonitrile (solvent B), were used in a gradient program that was 11.5 min long: 0–0.5 min, 0.1% B in A; 0.5–5.0 min, 0.1–30.0% B in A (linear gradient); 5.0–8.0 min, 30.0–45.0% B in A (linear gradient); 8.0–8.1 min, 45.0–90.0% B in A (linear gradient); and 8.1–11.5 min, column wash and stabilisation. The flow rate was 0.5 ml/min, and the injection volume was 5 ll. The photodiode array detector was operating between 190 and 500 nm, and phenolic compounds were detected at 280 nm. All samples were filtered using a syringe filter (4 mm, 0.2 lm PTFE, Thermo Fisher Scientific Inc., Waltham, USA) prior to the UPLC–MS analyses.

The UPLC system was combined with a triple quadrupole mass spectrometer (XevoÒ TQ, Waters Corporation, Milford, USA). The mass spectrometer was operated in a negative ionisation mode, and ions were scanned between 250 and 1400 Da. The capillary voltage was 3.4 kV, the desolvation temperature was 650 °C, and the source temperature was 150 °C. The desolvation and cone gas (N2) flow rates were 1000 l/h and 100 l/h, respectively. Individual phenolic compounds were identified by comparing their mass and UV spectra to values in the literature and the spectral libraries of our laboratory. The following chemicals were used: A pH 10 carbonate buffer (50 mM; sodium carbonate/sodium hydrogen carbonate, J.T. Baker, Deventer, the Netherlands), formic acid (for the oxidation assays, J.T. Baker, Deventer, the Netherlands), formic acid (for UPLC–MS, VWR International Ltd., Poole, England), sodium carbonate (Oy FF-Chemicals Ab, Yli-Ii, Finland), Folin & Ciocalteu’s phenol reagent, gallic acid, and acetonitrile (LC–MS grade) (Sigma–Aldrich Co., St. Louis, USA). 5.2. Plant samples and reproducibility assays The oxidation method was performed using phenolic extracts from twelve plant species. Voucher specimens from all twelve plant species are deposited in the University of Turku Herbarium (TUR) (Turku, Finland, voucher numbers TUR-599519 – TUR599530). The plant extracts came from two separate projects (from Finland and Uganda), which explains why two different extraction protocols were used for the samples. The first set of samples contained leaves of C. durandii, O. speciosus and P. africana, which were collected within a 1-km radius around the Makerere University Biological Station (MUBFS) in Kibale National Park in western Uganda. Six trees of each species were used, and one leaf sample was collected from each tree. The leaf samples were analysed individually to assess the reproducibility of the oxidation method. The lyophilised and ground leaves were first extracted using 70% aq. acetone (3  4 h) followed by MeOH:MeCl2 (50:50, v:v, 3  4 h). The acetone and MeOH:MeCl2 extracts were combined, after which the organic solvents were evaporated from the extracts. The extracts were frozen and lyophilised. Finally, the extracts were liquid–liquid extracted using water and hexane. A sample was collected from the water phase, and this sample, containing the water-soluble phenolic compounds, was frozen and lyophilised for the oxidation assays. All of the test samples of C. durandii, O. speciosus and P. africana displayed similar trends based on the total phenolics assay and UPLC–MS analyses, which confirmed that individual phenolic compounds displayed a reproducible activity level under alkaline conditions. The second set of samples consisted of nine species (A. ptarmica, A. hippocastanum, B. pubescens, F. moschata, Iris sp., P. abies, R. alpinum, S. phylicifolia and Q. robur), which were collected from the Botanical Garden of the University of Turku and nearby forests in Turku, Finland. One pooled sample was collected from all plant species so that the sample contained leaves (the sample of A. ptarmica contained flowers) from 5 to 10 different individuals. All pooled plant samples were lyophilised and ground into powder. The plant powder was extracted using 80% aq. acetone for 2 h, and the extraction was repeated two times. The acetone was evaporated from the extracts, and the extracts were frozen and lyophilised prior to the oxidation assays. 5.3. The oxidation assay The oxidation of the phenolic compounds was performed according to the method of Salminen and Karonen (2011) with some modifications. All microplate reader (Thermo Multiskan Ascent, Thermo Electron Corporation, Shanghai, China) analyses

M. Vihakas et al. / Phytochemistry 103 (2014) 76–84

were performed in triplicate, and the average values of these analyses were used for the calculations. For the oxidation of phenolic compounds, 20 ll of the diluted plant extract was pipetted into one well of the 96-well microplate reader. After that, 180 ll of a buffer at pH 10 was applied to each well, and the samples were incubated at 25 °C for 60 min with shaking for 10 s per minute. After 60 min, 100 ll of 0.6% formic acid was applied to each well, which neutralised the samples to pH 6. For the initial plant samples, 20 ll of the untreated plant extract was mixed with 280 ll of a buffer mixture that contained 9 volumes of buffer at pH 10 and 5 volumes of 0.6% formic acid (overall pH 6). This sample represented the plant extract in its initial, nonoxidised state, but it contained the same amounts of buffers and had the same pH as the neutralised sample of the oxidation assay. This allowed for direct comparisons between the initial and oxidised plant samples in the following analyses. The Folin–Ciocalteu assay was used for the analysis of the total phenolics. Samples of 50 ll of the initial and oxidised extracts were placed in separate wells of the 96-well plate. Into one column of the plate, samples including 50 ll of water (blank) and gallic acid standard soln. (10; 25; 100 lg/ml) were placed. Then, 50 ll of 1 N Folin–Ciocalteu reagent was applied to each well. The plate was shaken for 1 min, after which 100 ll of 20% sodium carbonate (m/v) was added to each well. The plate was shaken for 10 s per min, and the absorbance at 730 nm was measured immediately after shaking. The measurement duration was 60 min. The concentrations of the total phenolics were calculated as gallic acid equivalents. The concentrations of the total phenolics in the oxidised samples were compared to the corresponding concentrations in the initial samples. This calculation provided the proportion of the total phenolics that had been oxidised during the incubation. The initial and oxidised samples were analysed via UPLC–DAD–MS. The oxidised sample subjected to UPLC–DAD–MS was a pooled sample of three replicates from the 96-well plate. One initial and one oxidised sample was analysed per plant species, except for C. durandii, O. speciosus. and P. africana, all of which contained six samples from six individual trees.

Acknowledgements This study was financially supported by grants from the Kone Foundation (M.V.) and the Academy of Finland (Grant 258992 to JPS). The authors would like to thank the following people for their assistance: Marica Engström, Anne Koivuniemi and Jorma Kim collected the plant samples in Finland, and Vilma Lehtovaara collected the samples in Uganda; Simo Laine helped to identify the Finnish plant samples; Kalle Virta assisted in the oxidation assays of P. africana; Jeff Ahern kindly reviewed the language of the manuscript. The chemical analyses using the UPLC–MS system were made possible by a Strategic Research Grant of the University of Turku (Ecological Interactions).

References Appel, H.M., 1993. Phenolics in ecological interactions: the importance of oxidation. J. Chem. Ecol. 19, 1521–1552. Barbehenn, R.V., Bumgarner, S.L., Roosen, E.F., Martin, M.M., 2001. Antioxidant defenses in caterpillars: role of the ascorbate-recycling system in the midgut lumen. J. Insect Physiol. 47, 349–357. Barbehenn, R.V., Poopat, U., Spencer, B., 2003. Semiquinone and ascorbyl radicals in the gut fluids of caterpillars measured with EPR spectrometry. Insect Biochem. Mol. Biol. 33, 125–130. Barbehenn, R., Cheek, S., Gasperut, A., Lister, E., Maben, R., 2005. Phenolic compounds in red oak and sugar maple leaves have prooxidant activities in the midgut fluids of Malacosoma disstria and Orgyia leucostigma caterpillars. J. Chem. Ecol. 31, 969–988. Barbehenn, R.V., Jones, C.P., Hagerman, A.E., Karonen, M., Salminen, J.-P., 2006. Ellagitannins have greater oxidative activities than condensed tannins and

83

galloyl glucoses at high pH: potential impact on caterpillars. J. Chem. Ecol. 32, 2253–2267. Barbehenn, R., Weir, Q., Salminen, J.-P., 2008. Oxidation of ingested phenolics in the tree-feeding caterpillar Orgyia leucostigma depends on foliar chemical composition. J. Chem. Ecol. 34, 748–756. Barbehenn, R., Dukatz, C., Holt, C., Reese, A., Martiskainen, O., Salminen, J.-P., Yip, L., Tran, L., Constabel, P., 2010. Feeding on poplar leaves by caterpillars potentiates foliar peroxidase action in their guts and increases plant resistance. Oecologia 164, 993–1004. Berenbaum, M., 1980. Adaptive significance of midgut pH in larval Lepidoptera. Am. Nat. 115, 138–146. Bors, W., Michel, C., Stettmaier, K., 2000. Electron paramagnetic resonance studies of radical species of proanthocyanidins and gallate esters. Arch. Biochem. Biophys. 374, 347–355. Canada, A.T., Giannella, E., Nguyen, T.D., Mason, R.P., 1990. The production of reactive oxygen species by dietary flavonols. Free Radical Biol. Med. 9, 441–449. Cao, G., Sofic, E., Prior, R.L., 1997. Antioxidant and prooxidant behavior of flavonoids: structure–activity relationships. Free Radical Biol. Med. 22, 749– 760. Chan, T., Galati, G., O’Brien, P.J., 1999. Oxygen activation during peroxidase catalysed metabolism of flavones or flavanones. Chem. Biol. Interact. 122, 15– 25. Cilliers, J.J.L., Singleton, V.L., 1989. Nonenzymic autoxidative phenolic browning reactions in a caffeic acid model system. J. Agric. Food Chem. 37, 890–896. Cilliers, J.J.L., Singleton, V.L., 1991. Characterization of the products of nonenzymic autoxidative phenolic reactions in a caffeic acid model system. J. Agric. Food Chem. 39, 1298–1303. Daniel, E.M., Ratnayake, S., Kinstle, T., Stoner, G.D., 1991. The effects of pH and rat intestinal contents on the liberation of ellagic acid from purified and crude ellagitannins. J. Nat. Prod. 54, 946–952. Dow, J.A.T., 1984. Extremely high pH in biological systems: a model for carbonate transport. Am. J. Physiol. 246, 633–636. Duffey, S.S., Stout, M.J., 1996. Antinutritive and toxic components of plant defense against insects. Arch. Insect Biochem. Physiol. 32, 3–37. Felton, G.W., Donato, K., Del Vecchio, R.J., Duffey, S.S., 1989. Activation of plant foliar oxidases by insect feeding reduces nutritive quality of foliage for noctuid herbivores. J. Chem. Ecol. 15, 2667–2694. Ferreres, F., Gil-Izquierdo, A., Andrade, P.B., Valentão, P., Tomás-Barberán, F.A., 2007. Characterization of C-glycosyl flavones O-glycosylated by liquid chromatography–tandem mass spectrometry. J. Chromatogr. A 1161, 214–223. Gluchoff-Fiasson, K., Fenet, B., Leclerc, J.-C., Reynaud, J., Lussignol, M., Jay, M., 2001. Three new flavonol malonylrhamnosides from Ribes alpinum. Chem. Pharm. Bull. 49, 768–770. Gringorten, J.L., Crawford, D.N., Harvey, W.R., 1993. High pH in the ectoperitrophic space of the larval lepidopteran midgut. J. Exp. Biol. 183, 353–359. Gross, E.M., Brune, A., Walenciak, O., 2008. Gut pH, redox conditions and oxygen levels in an aquatic caterpillar: potential effects on the fate of ingested tannins. J. Insect Physiol. 54, 462–471. Hodnick, W.F., Kung, F.S., Roettger, W.J., Bohmont, C.W., Pardini, R.S., 1986. Inhibition of mitochondrial respiration and production of toxic oxygen radicals by flavonoids: a structure–activity study. Biochem. Pharmacol. 35, 2345–2357. Johnson, K.S., Barbehenn, R.V., 2000. Oxygen levels in the gut lumens of herbivorous insects. J. Insect Physiol. 46, 897–903. Kaltenhauser, M., Ellmerer, E.P., Zidorn, C., 2010. Rhamnopyranosylvitexin derivatives from Celtis australis. J. Serb. Chem. Soc. 75, 733–738. Karonen, M., Loponen, J., Ossipov, V., Pihlaja, K., 2004. Analysis of procyanidins in pine bark with reversed phase and normal-phase high-performance liquid chromatography–electrospray ionization mass spectrometry. Anal. Chim. Acta 522, 105–112. Klocke, J.A., van Wagenen, B., Balandrin, M.F., 1986. The ellagitannin geraniin and its hydrolysis products isolated as insect growth inhibitors from semi-arid land plants. Phytochemistry 25, 85–91. Koclar Avci, G., Coruh, N., Bolukbasi, U., Ogel, Z.B., 2013. Oxidation of phenolic compounds by the bifunctional catalase-phenol oxidase (CATPO) from Scytalidium thermophilum. Appl. Microbiol. Biotechnol. 97, 661–672. Kuhnle, J.A., Windle, J.J., Waiss, A.C., 1969. Electron paramagnetic resonance spectra of flavonoid anion-radicals. J. Chem. Soc. B, 613–616. Lahtinen, M., Kapari, L., Ossipov, V., Salminen, J.-P., Haukioja, E., Pihlaja, K., 2005. Biochemical transformation of birch leaf phenolics in larvae of six species of sawflies. Chemoecology 15, 153–159. Marklund, S., Marklund, G., 1974. Involvement of the superoxide anion radical in the autoxidation of pyrogallol and a convenient assay for superoxide dismutase. Eur. J. Biochem. 47, 469–474. Miyoshi, N., Naniwa, K., Yamada, T., Osawa, T., Nakamura, Y., 2007. Dietary flavonoid apigenin is a potential inducer of intracellular oxidative stress: the role in the interruptive apoptotic signal. Arch. Biochem. Biophys. 466, 274–282. Moilanen, J., Salminen, J.-P., 2008. Ecologically neglected tannins and their biologically relevant activity: chemical structures of plant ellagitannins reveal their in vitro oxidative activity at high pH. Chemoecology 18, 73–83. Morant, A.V., Jørgensen, K., Jørgensen, C., Paquette, S.M., Sánchez-Pérez, R., Møller, B.L., Bak, S., 2008. b-Glucosidases as detonators of plant chemical defense. Phytochemistry 69, 1795–1813. Morimoto, S., Nonaka, G.-I., Nishioka, I., 1987. Tannins and related compounds. LIX. Aesculitannins, novel proanthocyanidins with doubly-bonded structures from Aesculus hippocastanum L. Chem. Pharm. Bull. 35, 4717–4729.

84

M. Vihakas et al. / Phytochemistry 103 (2014) 76–84

Nagels, L., van Dongen, W., de Brucker, J., de Pooter, H., 1980. High-performance liquid chromatographic separation of naturally occurring esters of phenolic acids. J. Chromatogr. 187, 181–187. Nakatani, N., Kayano, S., Kikuzaki, H., Sumino, K., Katagiri, K., Mitani, T., 2000. Identification, quantitative determination, and antioxidative activities of chlorogenic acid isomers in prune (Prunus domestica L.). J. Agric. Food Chem. 48, 5512–5516. Ossipov, V., Nurmi, K., Loponen, J., Haukioja, E., Pihlaja, K., 1996. High-performance liquid chromatographic separation and identification of phenolic compounds from leaves of Betula pubescens and Betula pendula. J. Chromatogr. A 721, 59–68. Pankoke, H., Bowers, M.D., Dobler, S., 2012. The interplay between toxin-releasing b-glucosidase and plant iridoid glycosides impairs larval development in a generalist caterpillar, Grammia incorrupta (Arctiidae). Insect Biochem. Mol. Biol. 42, 426–434. Pritsos, C.A., Ahmad, S., Bowen, S.M., Elliot, A.J., Blomquist, G.J., Pardini, R.S., 1988. Antioxidant enzymes of the black swallowtail butterfly, Papilio polyxenes, and their response to the prooxidant allelochemical, quercetin. Arch. Insect Biochem. Physiol. 8, 101–112. Rank, N.E., Köpf, A., Julkunen-Tiitto, R., Tahvanainen, J., 1998. Host preference and larval performance of the salicylate-using leaf beetle Phratora vitellinae. Ecology 79, 618–631. Rockenbach, J., Nahrstedt, A., Wray, V., 1992. Cyanogenic glycosides from Psydrax and Oxyanthus species. Phytochemistry 31, 567–570. Salminen, J.-P., Ossipov, V., Loponen, J., Haukioja, E., Pihlaja, K., 1999. Characterisation of hydrolysable tannins from leaves of Betula pubescens by high-performance liquid chromatography mass spectrometry. J. Chromatogr. A 864, 283–291. Salminen, J.-P., Ossipov, V., Pihlaja, K., 2002. Distribution of hydrolysable tannins in the foliage of Finnish birch species. Z. Naturforsch. 57c, 248–256. Salminen, J.-P., Roslin, T., Karonen, M., Sinkkonen, J., Pihlaja, K., Pulkkinen, P., 2004a. Seasonal variation in the content of hydrolysable tannins, flavonoid glycosides, and proanthocyanidins in oak leaves. J. Chem. Ecol. 30, 1693–1711.

Salminen, J.-P., Lahtinen, M., Lempa, K., Kapari, L., Haukioja, E., Pihlaja, K., 2004b. Metabolic modifications of birch leaf phenolics by an herbivorous insect: detoxification of flavonoid aglycones via glycosylation. Z. Naturforsch. 59c, 437–444. Salminen, J.-P., Karonen, M., 2011. Chemical ecology of tannins and other phenolics: we need a change in approach. Funct. Ecol. 25, 325–338. Slimestad, R., Hostettmann, K., 1996. Characterisation of phenolic constituents from juvenile and mature needles of Norway spruce by means of high performance liquid chromatography-mass spectrometry. Phytochem. Anal. 7, 42–48. Spitaler, R., Gurschler, S., Ellmerer, E., Schubert, B., Sgarbossa, M., Zidorn, C., 2009. Flavonoids from Celtis australis (Cannabaceae). Biochem. Syst. Ecol. 37, 120– 121. Summers, C.B., Felton, G.W., 1994. Prooxidant effects of phenolic acids on the generalist herbivore Helicoverpa zea (Lepidoptera:Noctuidae): potential mode of action for phenolic compounds in plant anti-herbivore chemistry. Insect Biochem. Mol. Biol. 24, 943–953. Tits, M., Angenot, L., Poukens, P., Warin, R., Dierckxsens, Y., 1992. Prodelphinidins from Ribes nigrum. Phytochemistry 3, 971–973. Tuominen, A., 2013. Defensive strategies in Geranium sylvaticum, Part 2: roles of water-soluble tannins, flavonoids and phenolic acids against natural enemies. Phytochemistry 95, 408–420. Tuominen, A., Sundman, T., 2013. Stability and oxidation products of hydrolysable tannins in basic conditions detected by HPLC/DAD–ESI/QTOF/MS. Phytochem. Anal. 24, 424–435. Williams, C.A., Harborne, J.B., Goldblatt, P., 1986. Correlations between phenolic patterns and tribal classification in the family Iridaceae. Phytochemistry 25, 2135–2154. Yarnes, C.T., Boecklen, W.J., Tuominen, K., Salminen, J.-P., 2006. Defining phytochemical phenotypes: size and shape analysis of phenolic compounds in oaks (Fagaceae, Quercus) of the Chihuahan desert. Can. J. Bot. 84, 1233–1248. Zagrobelny, M., Bak, S., Vinther Rasmussen, A., Jørgensen, B., Naumann, C.M., Lindberg Møller, B., 2004. Cyanogenic glucosides and plant–insect interactions. Phytochemistry 65, 293–306.