detection, characterization and identification of crucifer phytoalexins using high-performance...

12
Journal of Chromatography A, 1133 (2006) 172–183 Detection, characterization and identification of crucifer phytoalexins using high-performance liquid chromatography with diode array detection and electrospray ionization mass spectrometry M. Soledade C. Pedras , Adewale M. Adio, Mojmir Suchy, Denis P.O. Okinyo, Qing-An Zheng, Mukund Jha, Mohammed G. Sarwar Department of Chemistry, University of Saskatchewan, 110 Science Place, Saskatoon, SK S7N 5C9, Canada Received 3 May 2006; received in revised form 11 July 2006; accepted 8 August 2006 Available online 1 September 2006 Abstract We have analyzed 23 crucifer phytoalexins (e.g. brassinin, dioxibrassinin, cyclobrassinin, brassicanals A and C) by HPLC with diode array detection and electrospray ionization mass spectrometry (HPLC-DAD–ESI-MS) using both negative and positive ion modes. Positive ion mode ESI-MS appeared more sensitive than negative ion mode ESI-MS in detecting this group of compounds. A new HPLC separation method, new LC–MS and LC–MS 2 data and proposed fragmentation pathways, LC retention times, and UV spectra for selected compounds are reported. © 2006 Elsevier B.V. All rights reserved. Keywords: Crucifers; Phytoalexins; HPLC-DAD–ESI-MS 2 1. Introduction Phytoalexins are essential secondary metabolites produced de novo by plants in response to diverse forms of stress, includ- ing microbial attack [1]. The accumulation of phytoalexins is one of several induced defense responses associated with plant disease resistance [2]. Most of the crucifer phytoalexins are characterized by a substituted indole ring with nitrogen- and sulfur-containing functional groups. More than 30 indole and indole-related phytoalexins have been isolated from crucifers such as canola, rapeseed, mustard, cabbage, radish, wasabi and turnip since the first report [3] was disclosed [4]. Due to their worldwide cultivation and consumption, whether as vegetables, oilseeds or condiments, crucifers are of enormous economic importance. Therefore, not surprisingly, phytoalexins of cru- cifers have incited a great deal of research of their synthesis, Corresponding author. Tel.: +1 306 966 4772; fax: +1 306 966 4730. E-mail address: [email protected] (M.S.C. Pedras). biosynthesis, detoxification and evaluation of biological activ- ities [5]. Interestingly, a number of epidemiological studies have shown that a diet rich in crucifer vegetables may reduce the risk of various types of cancer by modulating carcinogen metabolism and that phytoalexins appear to have a positive effect [6]. The detection of phytoalexins in extracts of elicited tissues of crucifers has been carried out by TLC with biodetection, e.g. utilizing spores of Cladosporium or Bipolaris species, and HPLC with UV or diode array detection (DAD) [4]. However, the determination of chemical structures by spectroscopic meth- ods still requires the isolation of purified (chromatographically homogeneous material) phytoalexins from substantial amounts of plant material. For example, several kilograms of plant tissue could yield multi-grams of an extract which, after lengthy col- umn separations, could afford 1–5 mg of a compound suitable for spectroscopic characterization by NMR, etc. [4,7]. That is, the whole isolation process is tedious, expensive and time con- suming. Furthermore, considering that rather different crucifer species can produce several of the already known phytoalex- 0021-9673/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2006.08.015

Upload: m-soledade-c-pedras

Post on 26-Jun-2016

220 views

Category:

Documents


4 download

TRANSCRIPT

A

aimr©

K

1

diodcsistwoic

0d

Journal of Chromatography A, 1133 (2006) 172–183

Detection, characterization and identification of cruciferphytoalexins using high-performance liquid

chromatography with diode arraydetection and electrospray

ionization mass spectrometry

M. Soledade C. Pedras ∗, Adewale M. Adio, Mojmir Suchy,Denis P.O. Okinyo, Qing-An Zheng, Mukund Jha, Mohammed G. Sarwar

Department of Chemistry, University of Saskatchewan, 110 Science Place, Saskatoon, SK S7N 5C9, Canada

Received 3 May 2006; received in revised form 11 July 2006; accepted 8 August 2006Available online 1 September 2006

bstract

We have analyzed 23 crucifer phytoalexins (e.g. brassinin, dioxibrassinin, cyclobrassinin, brassicanals A and C) by HPLC with diode

rray detection and electrospray ionization mass spectrometry (HPLC-DAD–ESI-MS) using both negative and positive ion modes. Positiveon mode ESI-MS appeared more sensitive than negative ion mode ESI-MS in detecting this group of compounds. A new HPLC separation

ethod, new LC–MS and LC–MS2 data and proposed fragmentation pathways, LC retention times, and UV spectra for selected compounds areeported.

2006 Elsevier B.V. All rights reserved.

bihtm[

oeHtoh

eywords: Crucifers; Phytoalexins; HPLC-DAD–ESI-MS2

. Introduction

Phytoalexins are essential secondary metabolites producede novo by plants in response to diverse forms of stress, includ-ng microbial attack [1]. The accumulation of phytoalexins isne of several induced defense responses associated with plantisease resistance [2]. Most of the crucifer phytoalexins areharacterized by a substituted indole ring with nitrogen- andulfur-containing functional groups. More than 30 indole andndole-related phytoalexins have been isolated from crucifersuch as canola, rapeseed, mustard, cabbage, radish, wasabi andurnip since the first report [3] was disclosed [4]. Due to theirorldwide cultivation and consumption, whether as vegetables,

ilseeds or condiments, crucifers are of enormous economicmportance. Therefore, not surprisingly, phytoalexins of cru-ifers have incited a great deal of research of their synthesis,

∗ Corresponding author. Tel.: +1 306 966 4772; fax: +1 306 966 4730.E-mail address: [email protected] (M.S.C. Pedras).

ocuftss

021-9673/$ – see front matter © 2006 Elsevier B.V. All rights reserved.oi:10.1016/j.chroma.2006.08.015

iosynthesis, detoxification and evaluation of biological activ-ties [5]. Interestingly, a number of epidemiological studiesave shown that a diet rich in crucifer vegetables may reducehe risk of various types of cancer by modulating carcinogenetabolism and that phytoalexins appear to have a positive effect

6].The detection of phytoalexins in extracts of elicited tissues

f crucifers has been carried out by TLC with biodetection,.g. utilizing spores of Cladosporium or Bipolaris species, andPLC with UV or diode array detection (DAD) [4]. However,

he determination of chemical structures by spectroscopic meth-ds still requires the isolation of purified (chromatographicallyomogeneous material) phytoalexins from substantial amountsf plant material. For example, several kilograms of plant tissueould yield multi-grams of an extract which, after lengthy col-mn separations, could afford 1–5 mg of a compound suitable

or spectroscopic characterization by NMR, etc. [4,7]. That is,he whole isolation process is tedious, expensive and time con-uming. Furthermore, considering that rather different cruciferpecies can produce several of the already known phytoalex-

M.S.C

.Pedrasetal./J.C

hromatogr.A

1133(2006)

172–183173

Table 1Characterization of crucifer phytoalexins and related compounds 1–25 by HPLC-DAD–ESI-MS2 analyses

Peak no.,tR (min)

Compound name (molecular formula) [M] ESI-MS: m/z (relative intensity, %)positive mode MS

Positive mode MS2 of [M + H]+a Negative mode MS UV, λmax (nm)

1, 2.8 Isalexin (C9H7NO3) 177 200 [M + Na]+ (100), 178 (93), 132(6), 105 (7)

160 (100), 132 (56), 105 (66) NDb 199, 234, 335

2, 5.4 Indole-3-carboxaldehyde (C9H7NO) 145 168 [M + Na]+ (5), 146 (70), 118(100)

118 (100) 144 (100) 210, 245, 260, 300

3, 6.5 Dioxibrassinin (C11H12N2O2S2) 268 291 [M + Na]+ (80), 269 (48), 251(15), 221 (59), 203 (100), 161 (39)

251 (19), 221 (81), 203 (100), 161 (42) 267 (12), 219 (11), 159 (100), 160 (19) 210, 255

4, 6.8 Cyclobrassinone (C11H8N2O2S) 232 255 [M + Na]+ (6), 233 (100), 176(3)

206 (17), 201 (20), 176 (100) 231 (100), 174 (13) 218, 278, 312

5, 8.8 Brassicanal C (C10H9NO3S) 223 246 [M + Na]+ (28), 192 (100), 174(71), 164 (64), 148 (73), 146 (70)

– 222 (100), 192 (15), 193 (13) 215, 247, 310

6, 9.1 Camalexin (C11H8N2S) 200 201 (100) NDb 199 (100), 190 (6), 172 (7), 142 (3) 215, 278, 3187, 9.3 Brassicanal A (C10H9NOS) 191 214 [M + Na]+ (6), 192 (100), 164

(14), 117 (24)164 (100), 117 (53) 190 (100), 175 (9) 218, 258, 269, 325

8, 10.6 Brassilexin (C9H6N2S) 174 175 (100), 148 (5) 148 (100) 173 (100) 220, 245, 2649, 10.6 Indolyl-3-acetonitrile (C10H8N2) 156 179 [M + Na]+ (2), 157 (5), 130

(100)130 (100) NDb 220, 272, 285

10, 10.9 Spirobrassinin (C11H12N2S2) 250 273 [M + Na]+ (8), 251 (100), 203(21), 178 (8)

203 (100), 178 (38) 249 (89), 217 (3), 201 (100) 220, 258, 296

11, 11.0 Cyclobrassinin sulfoxide (C11H11N2OS2) 250 273 [M + Na]+ (15), 251 (3), 187(100)

– 249 (3), 201 (4), 160 (100) 213, 226, 278, 330

12, 11.3 Brassitin (C11H12N2OS) 220 243 [M + Na]+ (28), 130 (100) – NDb 220, 270, 27813, 12.3 Arvelexin (C11H10N2O) 186 209 [M + Na]+ (14), 187 (100), 160

(33), 147 (65), 132 (5)160 (61), 147 (100) NDb 220, 266, 280, 290

14, 12.9 Rutalexin (C11H8N2O2S) 232 255 [M + Na]+ (5), 233 (100), 192(17), 148 (17)

176 (5), 148 (100) 231 (100), 174 (4) 213, 242, 275

15, 14.4 Brassicanate A (C11H11NO2S) 221 244 [M + Na]+ (28), 222 (8), 190(100)

– 220 (100), 205 (7)c 220, 238, 268, 300

16, 15.3 Caulilexin A (C10H9NO3S) 223 246 [M + Na]+ 224 (31), 176 (100) 176 (100) 222 (5), 176 (100) 213, 252, 31817, 16.1 1-Methoxyspirobrassinin (C12H14N2OS2) 280 303 [M + Na]+ (3), 281 (100), 250

(21)250 (100), 178 (38) NDb 218, 260, 295

18, 16.3 1-Methoxyindolyl-3-acetonitrile (C11H10N2O) 186 187 (49), 160 (100), 130 (20) 171 (10), 160 (100), 146 (141) NDb 220, 27219, 18.3 Brassinin (C11H12N2S2) 236 259 [M + Na]+ (100), 237 (2), 176

(2), 130 (51)– 236 (100), 190 (9), 172 (7) 220, 269

20, 19.3 Sinalexin (C10H8N2OS) 204 205 (39), 174 (100) 188 (8), 174 (100) NDb 225, 247, 26221, 20.4 Erucalexin (C12H12N2O2S2) 280 303 [M + Na]+ (2), 281 (57), 250

(100), 249 (66), 203 (35)249 (100), 203 (26) NDb 234, 262, 368

22, 20.6 Wasalexin B (C13H14N2O2S2) 294 317 [M + Na]+ (2), 295 (100), 263(3), 247 (15)

263 (23), 247 (100) NDb 205, 248, 285, 368

23, 22.8 Wasalexin A (C13H14N2O2S2) 294 317 [M + Na]+ (6), 295 (100), 263(6), 247 (21)

263 (23), 247 (100) NDb 205, 248, 285, 368

24, 23.1 Cyclobrassinin (C11H11N2S2) 234 235 (37), 187 (15), 162 (100) 187 (20), 162 (100) 233 (8), 190 (37), 172 (9), 161 (100) 205, 229, 285, 29425, 23.6 1-Methoxybrassinin (C12H14N2OS2) 266 289 [M + Na]+ (25), 267 (49), 235

(36), 219 (10), 160 (100), 146 (68),128 (30), 117 (87)

219 (55), 160 (25), 146 (100) 265 (100), 220 (47), 190 (48), 144 (68) 218, 268

a ESI-MS2 not available for [M + H]+ with low intensity (<10%).b ND: not detected.c Negative ESI-MS2 m/z 205 (100), 175 (7).

1 omato

iwat

utchpaiuTonpbwwtan

pt

2

2

vNagpta

2

FE

74 M.S.C. Pedras et al. / J. Chr

ns (e.g. both Eutrema wasabi and Thlaspi arvensis produceasalexins [4]), a simple and rapid analytical method for char-

cterization and identification of currently known crucifer phy-oalexins is of enormous interest.

Electrospray ionization (ESI) is a soft ionization techniquesed in mass spectrometry applicable to the structural charac-erization and elucidation of non-volatile and thermally labileompounds present in complex biological mixtures [8]. ESIas been used for mass spectral analysis of a variety of naturalroducts such as flavonoids, especially flavonols, flavones,nd their glycosides [9]. In addition, several methodologiesncluding HPLC–ESI-MS2 have been used to quantify individ-al and total intact glucosinolates in cruciferous plants [10].he availability in our plant metabolite libraries of a numberf crucifer phytoalexins [4] prompted the development of aew HPLC method and the analyses of the fragmentationatterns of 23 crucifer phytoalexins and 2 related compoundsy HPLC-DAD–ESI-MS2. The MS, MS2 and UV data togetherith HPLC retention times (tR) of crucifer phytoalexins

ere obtained and compiled to produce a library that allows

he complete characterization of these metabolites. Thus,pplication of this new separation method together with theew data should facilitate the identification of known crucifer

Udla

ig. 1. Typical HPLC-DAD–UV spectra of crucifer phytoalexins: A, brassicanal C (, arvelexin (13) and F, brassinin (19).

gr. A 1133 (2006) 172–183

hytoalexins [4,7,11,12] present in complex plant extracts inhe absence of authentic samples of these compounds.

. Experimental

.1. General

Crucifer phytoalexins (Table 1) were synthesized as pre-iously reported with purities above 98% (HPLC andMR) [4,7,11,12]. Indole-3-carboxaldehyde (2) and indolyl-3-

cetonitrile (9) were purchased from Aldrich, USA and HPLC-rade acetonitrile was purchased from VWR, Canada. Ultra-ure water from a NANOpure Diamond water purification sys-em, purchased from Barnstead Thermolyne, UK, was used forll analyses.

.2. LC-DAD–ESI-MS analyses

An Agilent 1100 series HPLC system (Agilent Technologies,

SA) equipped with autosampler, binary pump, degasser, aiode array detector connected directly to a mass detector (Agi-ent G2440A MSD-Trap-XCT ion trap mass spectrometer) withn ESI source was used. Chromatographic separation was car-

5); B, brassilexin (8); C, spirobrassinin (10); D, cyclobrassinin sulfoxide (11);

omatogr. A 1133 (2006) 172–183 175

r(pa3ipdccHuTptMmtbaa

2

twasi

3

3

ia[MdiattsTucomUtss

o

Table 2Common fragment ions of crucifer phytoalexins obtained ESI-MS2

Phytoalexins/analogs m/z Fragment ions

Brassinin 107 H2NC(S)SCH3

Brassitin 91 H2NC(O)SCH3

Indolyl-3-acetonitrile 27 HCNCyclobrassinin 73, 57 NCSCH3, NCOCH3

Methoxy containing 31 OCH3

Thiomethyl containing 47 SCH3

Sulfoxide containing 63 S(O)CH3

wtcae

3Group I consists of dioxibrassinin (3), brassitin (12),

brassinin (19), wasalexins A (23) and B (22), and 1-methoxybrassinin (25) (Fig. 2). These are indole substitutedcompounds with a thiocarbamate (12), dithiocarbamate (3, 19,

M.S.C. Pedras et al. / J. Chr

ied out at room temperature using an Eclipse XSB C-18 column5 �m particle size silica, 150 mm × 4.6 mm I.D.). The mobilehase consisted of a gradient of 0.2% formic acid in water (A)nd 0.2% formic acid in acetonitrile (B) (75% A to 75% B in5 min, to 100% B in 5 min) and a flow rate of 1.0 mL/min. Theon mode was set as positive and negative. The interface and MSarameters were as follows: nebulizer pressure, 70.0 psi (N2);ry gas, N2 (12.0 L/min); dry gas temperature, 350 ◦C; sprayapillary voltage 3500 V; skimmer voltage, 40.0 V; ion transferapillary exit, 100 V; scan range, m/z 100–500. Ultrahigh puree was used as the collision gas. MS2 analyses were carried outsing the data dependent acquisition capabilities of the LC/MSDrap software data system (Agilent). Data were acquired inositive and negative modes in a single LC run, using the con-inuous polarity switching ability of the mass spectrometer. The

S2 spectra were acquired automatically in a data dependentode that used criteria from the previous MS scan to select

he target precursor peak. All data acquired were processedy Agilent Chemstation software. Each sample dissolved incetonitrile (ca. 0.1–0.3 mg/mL) was injected (10 �L) using anutosampler.

.3. Plant material and preparation of extracts

Rutabaga (Brassica napus L. ssp. rapifera) tubers werereated as previously reported [11], the ethyl acetate extractsere concentrated to dryness, dissolved in acetonitrile, filtered

nd injected using an autosampler. HPLC-DAD–ESI-MS analy-es of these ethyl acetate extracts were obtained under conditiondentical to those reported above for each phytoalexin.

. Results and discussion

.1. Fragmentation patterns of compounds

As shown in Table 1, 25 compounds including 23 phytoalex-ns [4,7,11], indole-3-carboxaldehyde (2, an internal standard)nd cyclobrassinone (4) (not a naturally occurring phytoalexin)11] were analyzed in both negative and positive mode by ESI-

S. All phytoalexins and the two related compounds 1–25 wereetected as protonated molecular ions, and/or sodium adductons in positive ion mode, whereas only 15 were detected in neg-tive ion mode. Under the conditions used for these analyses,he positive mode was more sensitive than the negative mode,herefore ESI in the positive mode was selected as the ionizationource for further characterization of each compound (Table 1).he fragmentation patterns in positive ion MS and MS2 weresed to characterize the molecular structures of the 23 cru-ifer phytoalexins. The most abundant characteristic fragmentsbserved for each compound are displayed in Table 1. Further-ore, simultaneous DAD monitoring of each sample allowedV characterization of each compound (Table 1). Fig. 1 shows

he UV spectra of six representative crucifer phytoalexins: bras-

icanal C (5), brassilexin (8), spirobrassinin (10), cyclobrassininulfoxide (11), arvelexin (13) and brassinin (19).

To facilitate the discussion of the fragmentation patternsbserved in ESI-MS for each compound, the 25 compounds F

Fig. 2. Structures of group I phytoalexins 3, 12, 22, 19, 23 and 25.

ere distributed into six (I–VI) groups according to their struc-ural relationship and/or fragmentation pattern. Table 2 showsommon fragments observed from retro-Diels–Alder reactionsnd/or direct cleavage of the C X (X N, O, S) bonds under thexperimental conditions used.

.1.1. Fragmentation of group I phytoalexins

ig. 3. Proposed fragmentation patterns of group I phytoalexins 12, 19 and 25.

176 M.S.C. Pedras et al. / J. Chromatogr. A 1133 (2006) 172–183

Fig. 4. Full scan MS (a) and MS2 (b) spectra for the [M + H]+ ions of dioxibrassinin (3).

Fig. 5. Structures of cyclobrassinone (4) and group II phytoalexins.Fig. 6. Proposed fragmentation patterns of group II phytoalexin 14 and com-pound 4.

M.S.C. Pedras et al. / J. Chromato

F

2uni

itsmln2Ctm1p2[

ig. 7. Proposed fragmentation patterns of group II phytoalexins 11 and 24.

5), or dithioimidate (22, 23) group at the 3-position. The molec-lar ions of compounds 3, 19 and 25 were observed in bothegative and positive ion modes using ESI, while the molecularons of compounds 12, 22 and 23 were observed in the positive

3

(

Fig. 8. Full scan MS (a) and MS2 (b) spectra for

gr. A 1133 (2006) 172–183 177

on mode only (Table 1). Fig. 3 shows the proposed fragmenta-ion patterns of 12, 19 and 25 in positive ion mode and Fig. 4hows the MS and MS2 spectra of 3. In compound 3, the for-ation of the base peak at m/z 203 can be rationalized through

oss of H2O and SCH3 ([M + H-H2O-SCH3]+) from the proto-ated molecular ion at m/z 269. Further fragmentation of m/z03 yielded ions at m/z 178, 161 and 145 in the MS2 spectra.ompounds 19 and 25 gave m/z 130 and 160, respectively, due

o loss of m/z 107 [M + H-H2NC(S)SCH3]+. Similarly, loss of/z 91 [M + H-H2NC(O)SCH3]+ in compound 12 gave peak m/z30. For compounds 22 and 23, which were observed only inositive ion mode, the formation of the base peaks m/z 247 and63, respectively, can be rationalized through the loss of m/z 48HSCH3]+ and m/z 32 [HOCH3]+, respectively.

.1.2. Fragmentation of group II phytoalexinsGroup II consists of cyclobrassinin sulfoxide (11), rutalexin

14), and cyclobrassinin (24) (Fig. 5). This group of compounds

the [M + H]+ ions of cyclobrassinin (24).

178 M.S.C. Pedras et al. / J. Chromatogr. A 1133 (2006) 172–183

hcuoHoibmi

Fig. 9. Structures of group III phytoalexins.

as a tricyclic structure consisting of a six membered hetero-yclic ring fused to an indole ring. Typically, these compoundsndergo a retro-Diels–Alder fission of ring C and/or cleavagef a weak C X bond (X O, S, N) of groups such as HSCH3,OCH3, or HS(O)CH3, corresponding to loss of m/z 48, 32,r 64, respectively. Compounds 11, 14 and 24 were observed

n both negative and positive modes by ESI-MS analyses. Thease peak m/z 187 of compound 11, detected in the positive ionode, corresponds to direct cleavage of the protonated sulfox-

de group m/z 64 [HS(O)CH3]+; however, the base peak m/z

F1

Fig. 11. Full scan MS (a) and MS2 (b) spectra

ig. 10. Proposed fragmentation patterns of group IV phytoalexins 9, 13 and8.

for the [M + H]+ ions of arvelexin (13).

M.S.C. Pedras et al. / J. Chromatogr. A 1133 (2006) 172–183 179

F

1fimaa

Twn1

ig. 12. Structures of indole-3-carboxaldehyde (2) and group IV phytoalexins.

60 [M-1-NCS(O)CH3]− resulting from a retro-Diels–Alder

ssion was detected in negative ion mode. In the positive ionode the ESI-MS2 spectrum of 14 showed the molecular ion

t m/z 233 [M + H]+ and base peak at m/z 148, as a result ofretro-Diels–Alder reaction followed by loss of CO, Fig. 6.

[nfn

Fig. 14. Full scan MS (a) and MS2 (b) spectra f

Fig. 13. Proposed fragmentation patterns of group IV phytoalexin 16.

he fragmentation pattern of rutalexin (14), a phytoalexin [11],as similar to the closely related synthetic isomer cyclobrassi-one (4). In the case of cyclobrassinone (4), the base peak m/z76 ([M + H-C2H3NO]+) results from elimination of m/z 57

NCOCH3] through a retro-Diels–Alder fission of the promi-ent [M + H]+ m/z 233 (Table 1). Fig. 6 shows the proposedragmentation pathway for both rutalexin (14) and cyclobrassi-one (4). In case of cyclobrassinin (24), the base peak m/z 162

or the [M + H]+ ions of caulilexin A (16).

180 M.S.C. Pedras et al. / J. Chromatogr. A 1133 (2006) 172–183

cmTcM

3

m

aato

+ +

Fig. 15. Structures of group V phytoalexins.

orresponds to [M + H-C2H3NS]+ formed by the elimination of/z 73 (MeSCN) through a retro-Diels–Alder fission (Fig. 7).he fragment m/z 187 [M + H-HSCH3]+ was formed by directleavage of HSCH3 group corresponding to m/z 48. The MS andS2 spectra of 24 is shown in Fig. 8.

.1.3. Fragmentation of group III phytoalexinsGroup III consists of indolyl-3-acetonitrile (9), arvelexin (4-

ethoxyindolyl-3-acetonitrile) (13), and 1-methoxyindolyl-3-

Fig. 16. Full scan MS (a) and MS2 (b) spectra fo

baoa

Fig. 17. Proposed fragmentation patterns of group V phytoalexin 10.

cetonitrile (18) (Fig. 9). Compounds 9, 13 and 18 were detecteds protonated molecular ions, and/or sodium adduct ions in posi-ive ion mode. In general, the fragmentation of HCN (m/z 27) wasbserved for this group of compounds (Fig. 10). Nonetheless,

r the [M + H]+ ions of spirobrassinin (10).

oth the [M + H] and [M + Na] ions of 9 had very low intensitynd might be inconspicuous in a complex mixture. The existencef an abundant fragment ion at m/z 160 in the MS2 spectra of 13nd 18 is consistent with the presence of a OCH3 in the indole

M.S.C. Pedras et al. / J. Chromatogr. A 1133 (2006) 172–183 181

Fig. 18. Structures of group VI phytoalexins.

Fig. 20. Full scan MS (a) and MS2 (b) sp

Fig. 19. Proposed fragmentation patterns of group VI phytoalexin 1.

ectra for the [M + H]+ ions of 20.

1 omato

rs

3

A1aaao(cH

F(n

2ir

3

b(a1

+

82 M.S.C. Pedras et al. / J. Chr

ing (Fig. 10). The MS and MS2 spectra of compound 13 ishown in Fig. 11.

.1.4. Fragmentation of group IV phytoalexinsGroup IV includes brassicanals C (5) and A (7), brassicanate(15) and caulilexin A (16) (Fig. 12). Compounds 5, 7, 15,

6 and indole-3-carboxaldehyde (2) were observed in both neg-tive and positive mode LC–ESI-MS analyses. Compounds 5nd 15 gave in the positive ion mode a base peak m/z 192nd 190 [M + H-HOCH3]+

, respectively, corresponding to loss

f HOCH3 (m/z 32), while 7 gave the molecular ion m/z 192[M + H]+) as the base peak. The formation of ion m/z 176 forompound 16 can be explained by a loss of HSCH3 [M + H-SCH3]+, whereas the base peak m/z 118 observed for aldehyde

ig. 21. HPLC chromatograms using MS and DAD. (A–C) Mixture containing compoA) positive ion mode ESI-MS-TIC (total ion current) detection; (B) negative ion mapus L. ssp. rapifera) extracts; (D) positive ion mode ESI-MS-TIC detection; (E) ne

ciwg

gr. A 1133 (2006) 172–183

is due to loss of CO (Table 1). Figs. 13 and 14 show a rational-zation for the fragmentation pathways and MS spectra of 16,espectively.

.1.5. Fragmentation of group V phytoalexinsGroup V consists of “spiro” phytoalexins such as spiro-

rassinin (10), 1-methoxy spirobrassinin (17), and erucalexin21) (Fig. 15). Compound 10 was observed in both negativend positive ion modes in ESI-MS analysis, while compounds7 and 21 were detected only in positive ion mode (Table 1). In

unds 1–25 (numbering of peaks refer to their identification as shown in Table 1):ode ESI-MS-TIC detection; (C) DAD at 220 nm. (D–F) Rutabaga (Brassicagative ion mode ESI-MS-TIC detection; (F) DAD at 220 nm.

ompound 10, the base peak m/z 203 [M + H-HSCH3] observedn ESI-MS corresponds to direct cleavage of HSCH3 (m/z 48),hereas in compounds 17 and 21 the cleavage of the N O bondave the [M + H-HOCH3]+ (m/z 250, Table 1). Figs. 16 and 17

omato

sm

3

(s(adpTM(ba(is(

3a

pmus8tpaec(p

Uls1Mas(e([

4

o

tcipmEwtthlcamdstaucp

A

SedvGKm

R

[10] Q. Tian, R.A. Rosselot, J.S. Schwartz, Anal. Biochem. 343 (2005) 93.

M.S.C. Pedras et al. / J. Chr

how the MS spectra for compound 10 and the proposed frag-entation pathway, respectively.

.1.6. Fragmentation of group VI phytoalexinsGroup VI consists of relatively stable compounds, isalexin

1), camalexin (6), brassilexin (8) and sinalexin (20), whosetructures and fragmentation behaviors differ from groups I–VFig. 18). Compounds 6 and 8 were observed in both negativend positive ion modes by ESI-MS, while sinalexin (20) could beetected only in the positive ion mode. The MS spectrum of 1 dis-layed the sodium adduct as the base peak (m/z 200, [M + Na]+,able 1). The base peak m/z 160 [M + H-H2O]+ observed in theS2 spectrum of compound 1 can be attributed to loss of H2O

Fig. 19). Compounds 6 and 8 gave their molecular ion as thease peak, but 6 appeared to be relatively more stable than 8 sincepeak at m/z 148 was observed in its MS2 spectrum. Sinalexin

20) gave a base peak at m/z 174 [M + H-OCH3]+ correspond-ng to loss of OCH3 (m/z 31). Fig. 20 shows the MS and MS2

pectra of 20. No sodium adducts were observed for camalexin6), brassilexin (8) and sinalexin (20) (Table 1).

.2. HPLC-DAD–ESI-MS analyses of phytoalexin mixturend rutabaga extracts

A solution of phytoalexins and two standards, com-ounds 1–25, was analyzed using HPLC-DAD–ESI-MS. Chro-atograms were obtained under conditions identical to those

sed for single compounds. As shown in Fig. 21, threeets of peaks did not resolve adequately: (i) peaks with tR.8–9.3 min corresponding to compounds 5–7, (ii) peaks with

R 10.6–11.3 min corresponding to compounds 8–12 and (iii)eaks with tR 23.1–23.6 min corresponding to compounds 24nd 25. However, compounds 5–9, 12, 24 and 25 could be differ-ntiated and characterized by their [M + H]+ (Table 1) whereasompounds 10 and 11 were differentiated by their fragment ions10: m/z 203, 178; 11: m/z 187, Table 1). Therefore, all com-ounds 1–25 could be identified in the test mixture.

Next, the ethyl acetate extracts of rutabaga tubers elicited byV irradiation and non-elicited tissues (controls) [11] were ana-

yzed using HPLC-DAD–ESI-MS. Typical chromatograms arehown in Fig. 21D–F. Peaks 10, 14, 15, 24 and 25 at tR, 10.9,2.9, 14.4, 23.1 and 23.6 min, respectively, showed UV spectra,S and MS2 fragmentation patterns consistent with those of

uthentic standards of spirobrassinin (10), rutalexin (14), bras-icanate A (15), cyclobrassinin (24) and 1-methoxybrassinin25), respectively. The presence of these compounds in rutabagaxtracts was also confirmed by additional spectroscopic analysesNMR, FTIR) after isolation and purification of each phytoalexin11].

. Conclusions

The fragmentation patterns of 23 crucifer phytoalexinsbserved in an ESI ion trap mass spectrometer were analyzed

[

[

gr. A 1133 (2006) 172–183 183

o further characterize these structures. Under the experimentalonditions used for these HPLC–ESI-MS analyses, the positiveon mode was more sensitive in detecting this group of com-ounds than the negative ion mode (Table 1; Fig. 21). The frag-entation data obtained from negative or positive modes usingSI-MS2 indicated that the fragmentation patterns observedere similar to those obtained in ESI-MS. Therefore, it appears

hat spectrometer capabilities such as MS2 do not provide addi-ional information to characterize these molecules. On the otherand, UV data obtained in parallel with ESI-MS using an on-ine DAD provided data that facilitated the identification of theserucifer phytoalexins. It is concluded that HPLC-DAD–MSnalysis is a powerful tool that allows the identification ofost of the crucifer phytoalexins currently known using the

ata reported in Tables 1 and 2. These analyses can decreaseubstantially the time required for identification of known phy-oalexins present in crude or processed crucifer plant extractss isolation and direct comparison with authentic samples isnnecessary. Furthermore, the data provided in Tables 1 and 2an assist in the detection of new or unknown cruciferhytoalexins.

cknowledgements

Support for the authors’ work was obtained from the Naturalciences and Engineering Research Council of Canada (Discov-ry and AGENO Research Grants to M.S.C.P.), Canada Foun-ation for Innovation, Canada Research Chairs Program, Uni-ersity of Saskatchewan (graduate assistantships to D.P.O.O.,.M.S. and M.J.). We acknowledge the technical assistance of. Thoms, Department of Chemistry in mass spectrometry deter-inations.

eferences

[1] J.A. Bailey, J.W. Mansfield (Eds.), Phytoalexins, Blackie, UK, 1982,p. 1.

[2] R.A. Dixon, C.J. Lamb, Ann. Rev. Plant Physiol. Plant Mol. Biol. 41 (1990)339.

[3] M. Takasugi, N. Katsui, A. Shirata, J. Chem. Soc. Chem. Commun. (1986)1077.

[4] M.S.C. Pedras, M. Jha, P.W.K. Ahiahonu, Curr. Org. Chem. 7 (2003)1635.

[5] M.S.C. Pedras, P.W.K. Ahiahonu, Phytochemistry 66 (2005) 391.[6] M. Pilatova, M. Sarissky, P. Kutschy, A. Mirossay, R. Mezencev, Z.

Curillova, M. Suchy, K. Monde, L. Mirossay, J. Mojzis, Leukemia Res.29 (2005) 415.

[7] M.S.C. Pedras, M.G. Sarwar, M. Suchy, A.M. Adio, Phytochemistry 67(2006) 1503.

[8] M. Przybylski, M.O. Glocker, Angew. Chem. Int. Ed. Engl. 35 (1996)806.

[9] F. Cuyckens, M. Claeys, J. Mass Spectrom. 39 (2004) 1.

11] M.S.C. Pedras, S. Montaut, M. Suchy, J. Org. Chem. 69 (2004)4471.

12] M.S.C. Pedras, M. Suchy, P.W.K. Ahiahonu, Org. Biomol. Chem. 4 (2006)691.