Báo cáo khoa học: Brassinin oxidase, a fungal detoxifying enzyme to overcome a plant defense – purification, characterization and inhibition potx

The enzyme was purified by monitoring BO activity using brassinin as substrate.. Fractions with BO activity obtained in the last chromatography column were pooled, concentrated and used f

Brassinin oxidase, a fungal detoxifying enzyme to

overcome a plant defense – purification, characterization and inhibition

M S C Pedras, Zoran Minic and Mukund Jha

Department of Chemistry, University of Saskatchewan, Canada

Microbial plant pathogens display a variety of

succe-ssful strategies to invade plant tissues and obtain the

necessary nutrients that allow growth and

reproduc-tion Plants fight back with no smaller a variety of

weapons, including the synthesis of small to very large

molecules to inhibit specific metabolic processes in the

pathogen [1–3] In general, plants under microbial

attack produce de novo a blend of antimicrobial

defenses known as phytoalexins, the specific

compo-nents of which appear to depend on the type of stress

[4,5] Despite such an arsenal, fungal pathogens can

disarm the plant by counterattacking with enzymes

that detoxify promptly these phytoalexins [6–8] The

outcome of this ‘arms race’ [3] frequently favors the

pathogen, causing great crop devastation and

substan-tial yield losses Brassinin is a phytoalexin of great importance to crucifer plants, due to its dual role both

as an antimicrobial defense and a biosynthetic precur-sor of several other phytoalexins The toxophore group

of brassinin is a dithiocarbamate, with an interesting resemblance to the potent fungicides used in the 1960s [9] From a plant’s perspective, it is highly desirable to prevent brassinin detoxification by any pathogen Crucifers include a wide variety of crops cultivated across the world; for example, the oilseeds rapeseed and canola (Brassica napus and Brassica rapa) and vegetables such as cabbage (Brassica oleraceae var capitata), cauliflower (Brassica oleraceae var botrytis)

or broccoli (Brassica oleraceae var italica) In addi-tion, both wild and cultivated crucifers are known to

Keywords

brassinin oxidase; camalexin; detoxifying

enzyme; Leptosphaeria maculans;

phytoalexin

Correspondence

M S C Pedras, Department of Chemistry,

University of Saskatchewan, 110 Science

Place, Saskatoon, Saskatchewan S7N 5C9,

Canada

Fax: +1 306 966 4730

Tel: +1 306 966 4772

E-mail: s.pedras@usask.ca

(Recived 24 April 2008, revised 17 May

2008, accepted 21 May 2008)

doi:10.1111/j.1742-4658.2008.06513.x

Blackleg fungi [Leptosphaeria maculans (asexual stage Phoma lingam) and Leptosphaeria biglobosa] are devastating plant pathogens with well-estab-lished stratagems to invade crucifers, including the production of enzymes that detoxify plant defenses such as phytoalexins The significant roles of brassinin, both as a potent crucifer phytoalexin and a biosynthetic precur-sor of several other plant defenses, make it critical to plant fitness Brassi-nin oxidase, a detoxifying enzyme produced by L maculans both in vitro and in planta, catalyzes the detoxification of brassinin by the unusual oxi-dative transformation of a dithiocarbamate to an aldehyde Purified brassi-nin oxidase has an apparent molecular mass of 57 kDa, is approximately 20% glycosylated, and accepts a wide range of cofactors, including

quinon-es and flavins Purified brassinin oxidase was used to screen a library of brassinin analogues and crucifer phytoalexins for potential inhibitory activ-ity Unexpectedly, it was determined that the crucifer phytoalexins cama-lexin and cyclobrassinin are competitive inhibitors of brassinin oxidase This discovery suggests that camalexin could protect crucifers from attacks

by L maculans because camalexin is not metabolized by this pathogen and

is a strong mycelial growth inhibitor

Abbreviations

BO, brassinin oxidase; CKX, cytokinin oxidase ⁄ dehydrogenase; DEA, diethanolamine; FCC, flash column chromatography; PMS, phenazine methosulfate; PNGase, N-glycosidase; Q0,2,3-dimethoxy-5-methyl-1,4-benzoquinone; Q10,2,3-methoxy-5-methyl-6-geranyl-1,4-benzoquinone.

have positive effects on human health (e.g a high

intake of crucifers is associated with a reduced risk of

cancer) [10] Economically significant diseases of

cru-cifer oilseeds and vegetables caused by fungi such as

the ‘blackleg’ fungi [Leptosphaeria maculans (asexual

stage Phoma lingam) and Leptosphaeria biglobosa] are

a global issue [11] L maculans is a pathogen with

well-established stratagems to invade crucifers,

includ-ing the production of enzymes that detoxify essential

phytoalexins [7] For example, the phytoalexin

brassi-nin is detoxified via oxidation to

indole-3-carboxalde-hyde [7] or hydrolysis to indolyl-3-methanamine

(Fig 1) [12]

Considering the apparent specificity of the enzyme

involved in the oxidative detoxification of brassinin,

brassinin oxidase (BO), we suggested that BO

inhi-bitors could prevent detoxification of brassinin by

L maculans and thus avoid its depletion in infected

plants [13,14] The concomitant accumulation of

brass-inin and related phytoalexins might prompt a recovery

in which the infected plant would be able to ward off

the sensitive pathogen(s) To better understand the role

of BO and test potential inhibitors, the enzyme was

purified, characterized and shown to be a novel

enzyme, consistent with the unusual transformation it

catalyzes (Fig 1) Purified BO was used to screen a

library of 78 compounds containing crucifer

phytoal-exins and analogues for potential inhibitory activity

Surprisingly, we determined that the crucifer

phytoal-exins camalexin and cyclobrassinin inhibited BO

activ-ity substantially but BO activactiv-ity was not affected by

most of the synthetic compounds This discovery

suggests that, if camalexin was co-produced with brass-inin [5], it might protect Brassica sp from attacks by

L maculans because camalexin is not metabolized

by this pathogen and is a strong mycelial growth inhibitor

Results

Purification of BO activity Fungal cultures initiated from spores were grown under standard conditions and crude cell-free homo-genates were prepared from mycelia, as reported in the Experimental procedures The enzyme was purified by monitoring BO activity using brassinin as substrate Table 1 indicates the degree of purification and yield obtained for each step This purification protocol involved four steps: first employing DEAE-Sephacel, followed by chromatofocusing with PBE resin, then Superdex 200 and, finally, Q-Sepharose chromato-graphy Fractions with BO activity obtained in the last chromatography column were pooled, concentrated and used for biochemical analysis The purity of the protein isolated after Q-Sepharose chromatography was examined by SDS⁄ PAGE, which, upon staining with Coomassie brilliant blue R-250, revealed only one band having the apparent molecular mass of 57 kDa (Fig 2) In addition, Superdex 200 chromatography of the purified protein suggested that it was a native monomer because it was eluted at a position corre-sponding to a molecular mass similar to that deter-mined by SDS⁄ PAGE

Fig 1 Detoxification of the phytoalexin brassinin by the ‘blackleg’ fungi L maculans (L m.) and L biglobosa (L b.).

Table 1 Enzyme yields and purification factors for BO Recoveries are expressed as a percentage of initial activity and purification factors are calculated on the basis of specific activities (lmolÆmin)1= U).

Purification step

activity (mUÆmg)1)

Recovery (%)

Purification factor (fold) Protein (mg) Activity (mU)

a Mycelia from 1 L cultures yielded approximately 120 mg of protein.

Cellular localization of BO

The cellular localization of BO isolated from L

macu-lans was established after fractionation of the crude

protein extract into soluble, membrane and cell wall

fractions, and these fractions were used for enzymatic

assays As shown in Table 2, the BO specific activity

was found to be the highest in the cell wall fraction,

suggesting that BO was secreted (i.e a cell wall

pro-tein) However, the total BO activity was found to be

the highest in the soluble fraction, which could imply

that this protein was present in the cytoplasm as well

Because these inconclusive results were likely due to

contamination of the soluble protein fraction with cell

wall proteins, an additional fractionation was carried

out using concanavalin A chromatography [15,16], a

lectin used for purification of glycoproteins [17,18]

The majority of secreted proteins are glycosylated and

thus bind lectins specifically, namely those containing mannose or glucose (e.g concanavalin A) [19–21] Hence, the protein extracts of both soluble and cell wall fractions were subjected to concanavalin A Sepha-rose chromatography A single peak of activity was obtained after eluting each column with methyl-a-d-glucopyranoside (see supplementary Fig S1) Similar results were obtained using protein extracts from the first purification step using DEAE-Sephacel chroma-tography The maximum enzyme recovery was obtained using a relatively high concentration of methyl-a-d-glucopyranoside (1.0 m) These results suggest that BO is glycosylated and likely localized in the cell wall

Analysis of deglycosylated BO The cellular localization assays and the ability of BO

to bind concanavalin A suggested that BO was an N-glycosylated protein To determine whether BO is indeed a glycoprotein, purified BO was subjected to treatment with N-glycosidase (PNGase) F, an enzyme that cleaves N-linked oligosaccharides from proteins SDS⁄ PAGE analysis showed a shift in the migration

of BO (46 kDa) in the sample treated with PNGase versus the untreated sample (57 kDa) (Fig 2A), dem-onstrating that BO is an N-glycosylated protein (approximately 20%) To further characterize the nat-ure of the N-glycosylation of BO, samples of purified

BO were treated with endo-b-N-acetylglucosaminidase,

an enzyme that cleaves all high-mannose oligosaccharides from proteins Purified BO treated with endo-b-N-acetylglucosaminidase (Fig 2B) also showed a shift in the migration of BO (47 kDa) com-pared with the untreated sample (57 kDa) (Fig 2B)

Identification of BO tryptic peptides by LC-ESI-MS⁄ MS

Glycoproteins can escape analysis at any level of a peptide mass mapping procedure, in particular, during tryptic digestion, due to potential steric disturbance through interaction of the protein with proteolytic sites

of trypsin [22] For this reason, to determine the pep-tide sequence, analysis was performed with purified BO after treatment with PNGase F The deglycosylated

BO band in Fig 3 was digested with trypsin and then analyzed by LC-MS⁄ MS using mascot algorithms In total, 20 peptides were deduced from the LC-MS⁄ MS spectral data (Table 3) The sequence homology of the identified peptides was analyzed using the NCBI blast algorithm Peptides did not match significantly with proteins available in the NCBI blast database

Fig 2 SDS ⁄ PAGE of protein fractions from purification of BO.

Lane M, marker proteins (molecular masses are indicated); lane 1,

crude homogenate (40 lg); lane 2, DEAE-Sephacel pooled fractions

(10 lg); lane 3, Polybuffer exchanger 94 chromatography (10 lg);

lane 4, Superdex 200-pooled fractions (1.5 lg); lane 5, purified BO

after Q-Sepharose chromatography (1 lg).

Table 2 Fractionation of proteins from L maculans for cellular

localization of BO.

Protein

fraction

Volume

(mL)

Protein (mg)

Specific activity (nmolÆmin)1Æmg)1)

Total activity (nmolÆmin)1)

However, analysis of peptides using NCBI blast

data-base pertaining to fungi revealed that the majority of

peptides in Table 3 had some homology to a putative

short-chain dehydrogenase from Aspergillus terreus

NIH2624 (accession no XP_001210968) and putative

NADP-dependent flavin oxidoreductase from

Asper-gillus nidulans FGSC A4 (accession no XP_663310)

(results not shown)

Characterization of BO

BO required the presence of an electron acceptor for activity The purified enzyme was examined in the presence of various electron acceptors at concentra-tions of 0.10 and 0.50 mm As shown in Table 4, BO could accept a wide range of cofactors, including phen-azine methosulfate (PMS), 1,4-benzoquinone, 1,2-naphthoquinone, 2,6-dichloroindophenol;, coenzyme 2,3-dimethoxy-5-methyl-1,4-benzoquinone (Q0) and FMN The highest BO activity was obtained with PMS The quinones 1,4-antraquinone and coenzyme 2,3-meth-oxy-5-methyl-6-geranyl-1,4-benzoquinone (Q10) were not accepted, whereas the flavin derivative FMN acted

as an electron acceptor A number of other electron acceptors, such as FAD, duraquinone, NADP, cyto-chrome c and CuCl2, had low or no detectable effect

on BO activity The absorbance spectrum of purified

BO (0.1 mgÆmL)1) revealed a peak at 280 nm, typical

of proteins containing aromatic amino acids, but no chromophores characteristic of flavin or quinone dependant oxidoreductases were detected (no absorp-tion observed in the range 300–600 nm; results not shown)

The kinetic parameters for BO activity were deter-mined using brassinin as substrate in the presence of PMS as an electron acceptor Substrate saturation curves were fitted to the Michaelis–Menten equation to obtain the kinetic parameters The apparent Km and

kcatwere 0.15 mm and 0.95 s)1, respectively The cata-lytic efficiency (kcat⁄ Km) was determined to be of

6333 s)1Æm)1 The apparent Kmfor PMS was 0.30 lm The influence of pH on the activity of the BO was investigated in the range pH 3–11 The pH optima were determined to be in the range 8.0–10.0 (results not shown) The temperature dependence of BO activ-ity was tested in the range 8–75C, and the apparent optimum temperature was 45C (results not shown)

Identification of inhibitors of BO Several analogues of brassinin and phytoalexins (78 compounds; see supplementary Table S1) were synthe-sized, purified and characterized spectroscopically, as reported previously [13,14] The activity of BO was examined in the presence of these compounds at 0.10 mm (supplementary Table S1); the compounds showing inhibition were also tested at 0.30 mm (Table 5) Camalexin, cyclobrassinin, thiabendazole and isobrassinin inhibited BO activity, whereas none

of the remaining compounds had an effect Further-more, none of the compounds shown in supplementary Table S1 were substrates of BO Considering the

A

B

Fig 3 SDS ⁄ PAGE of deglycosylated BO Purified BO was

incu-bated with and without (A) PNGase F and (B)

endo-b-N-acetyl-glucosaminidase as described in the Experimental procedures.

Deglycosylated samples were separated by SDS ⁄ PAGE and

migra-tion of deglycosylated BO was estimated by comparison with

molec-ular markers (A) Overnight incubation of BO in nondenaturing

conditions with PNGase F results in a reduction of molecular mass

of BO (46 kDa) compared with nontreated BO (57 kDa) Treatment

of BO with PNGase F in denaturing conditions for 3 h also results in

a reduction of molecular mass of BO (46 kDa) compared with

non-treated BO (57 kDa) (B) Endo-b-N-acetylglucosaminidase treatment

of BO in denaturing conditions for 3 h results in a reduction of

mole-cular mass of BO (47 kDa) compared with nontreated BO (57 kDa).

substantially higher inhibitory effect of both camalexin

and cyclobrassinin, it was of great importance to

deter-mine the type of inhibition that each compound

displayed The kinetics of inhibition of BO is shown in the form of Lineweaver–Burk double reciprocal plots (1⁄ S versus 1 ⁄ V) using 0.10 and 0.30 mm concentra-tions of camalexin and cyclobrassinin (Fig 4) The results showed that the intersection points of all curves were on the 1⁄ V axis (i.e both camalexin and cyclo-brassinin competitively inhibited BO activity)

Kinetic mechanism of BO The bisubstrate reaction mechanism of BO involves the oxidation of brassinin by an electron acceptor such

as PMS Steady-state kinetic studies were performed to

Table 3 Masses and scores of tryptic peptides a obtained from deglycosylated BO after treatment by PNGase Observed, mass ⁄ charge of observed peptide; M r (expt), observed mass of peptide; M r (calc), calculated mass of matched peptide; Delta, difference (error) between the experimental and calculated masses; Score, ions score A score of 49 or greater indicates that the probability of an incorrect match is < 5%.

Table 4 Effect of electron acceptors on BO activity BO activities

measured under standard assay conditions described in the

Experi-mental procedures; results are expressed as the means ± SD of

three independent experiments; relative activity is expressed as

percentage of the reaction rate obtained with PMS ND, not

detected.

Cofactor (electron acceptor)

Relative activity (%) 0.10 m M 0.50 m M

a A rate of 100% corresponds to 840 mU mg)1 protein b From

Pedras et al [31].

Table 5 Effect of the phytoalexins camalexin and cyclobrassinin, the brassinin regioisomer isobrassinin and fungicide thiabendazole

on BO activity (a complete list with 78 tested compounds is pro-vided in the supplementary Table S1) BO activity was measured under standard conditions described in the Experimental proce-dures; inhibition is expressed as percentage of control activity; results are expressed as the means ± SD of at least four indepen-dent experiments.

Compound

Inhibition (%)

investigate the kinetic mechanism of BO Varying the

concentration of brassinin (0.05–0.30 mm) and keeping

the concentration of PMS constant (0.10, 0.20 and

0.60 lm) gave an intersecting pattern to the left of the

1⁄ S axis (Fig 5A) A second set of experiments was

performed varying the concentration of PMS (0.05–

0.30 lm) and keeping the concentration of brassinin

constant (0.05, 0.10 and 0.15 mm) (Fig 5B) The

inter-section point was on the 1⁄ V axis Both sets of data

were indicative of a sequential mechanism but did not

distinguish between an ordered or random sequential

mechanism These two types of kinetic mechanisms

could be distinguished using camalexin as the dead-end

inhibitor of BO Thus, kinetic data obtained from

experiments performed with various PMS

concen-trations (0.05–0.40 lm) and constant concenconcen-trations of

camalexin (0.10 and 0.30 mm) gave the characteristic

plot of uncompetitive inhibition (Fig 5C) By contrast, data obtained by varying the concentration of brassinin (0.05–0.30 mm) and keeping the concentra-tion of camalexin constant showed that camalexin was

Fig 4 Lineweaver–Burk plots of BO activities in the presence of

the phytoalexins (A) camalexin and (B) cyclobrassinin Purified

enzyme obtained from Q-Sepharose chromatography was used for

BO activity measurements Enzyme activity was determined as

described in the Experimental procedures.

Fig 5 Distinguishing ping-pong versus sequential kinetic mecha-nisms for BO (A) Lineweaver–Burk plot for the oxidation of brassi-nin carried out in the presence of a fixed concentration of PMS and varied [brassinin] (B) Lineweaver–Burk plot for the oxidation

of brassinin carried out in the presence of a fixed concentration of brassinin and varied [PMS] (C) Distinguishing ordered sequential versus random sequential mechanisms for BO Lineweaver–Burk plot for the dead-end inhibition of BO by camalexin at the indicated concentrations of PMS in the presence of a fixed concentration of brassinin at 0.60 m M

a competitive inhibitor (Fig 4A) These results

demon-strate that BO catalysis occurs through an ordered

mechanism in which brassinin binds first to the enzyme

followed by PMS binding to the BO binary complex

Analysis of BO activity in plants inoculated with

L maculans

B napusplants susceptible to infection by L maculans

and Brassica juncea plants resistant to infection by

L maculans BJ 125 were inoculated, incubated and

analyzed for BO activity The results obtained

(Table 6) demonstrate that only infected leaves and

stems of the susceptible plants exhibited BO activity;

no BO activity was found in non-inoculated stems or

leaves or inoculated resistant plants (Table 6)

Further-more, analysis of phytoalexin production showed the

presence of methoxybrassinin and spirobrassinin in

infected leaves of B napus [5]

Mycelia extracts of cultures of L maculans BJ 125

showed BO activity when cultures were induced with

3-phenylindole but only traces in control cultures

These analyses confirm that BO activity in L maculans

is inducible

Discussion

The present study reports the purification and

charac-terization of BO, a phytoalexin detoxifying enzyme

produced by the plant pathogenic fungus L maculans

both in infected plants and in axenic fungal cultures This enzyme is a monomer with an apparent molecular mass of 57 kDa that catalyzes the transformation of the dithiocarbamate toxophore of brassinin into the corresponding nontoxic aldehyde (Fig 1) BO appears

to be the first enzyme that has been described to cata-lyze this unique functional group transformation A peak of BO activity obtained by chromatofocusing was observed at pH 7.1–7.2, suggesting this to be the pI of the enzyme

Elution of BO from a concanavalin A Sepharose column suggested it to be glycosylated [23] Concanav-alin A affinity chromatography has been used for puri-fication of secreted proteins N-glycosylated with sugars such as d-glucose and d-mannose [18,24,25] To dem-onstrate that BO was indeed a glycosylated protein, purified BO was deglycosylated using either PNGase F

or endo-b-N-acetylglucosaminidase and the molecular mass of the native and deglycosylated forms of enzyme were compared by SDS⁄ PAGE Treatment of BO with either N-glycosidase caused a decrease in the apparent molecular mass of BO of approximately 20% (Fig 3) PNGase F and endo-b-N-acetylglucosaminidase are enzymes used for the release of N-linked glycans from glycoproteins [26,27]

Taken together, the assays used for cellular locali-zation (Table 2) and the glycosylation analysis (Fig 3)

of BO suggest that this enzyme is localized in the cell wall This cellular localization of BO could allow a more efficient detoxification of brassinin In this con-text, it is pertinent to point out that the enzyme cata-lyzing the detoxification of the phytoalexin kievitone, kievitone hydratase (EC 4.2.1.95), is also a glyco-enzyme secreted by the bean fungal pathogen Fusarium solani f sp phaseoli [28]

The peptides deduced from the LC-ESI-MS⁄ MS spectral data of purified BO digested with trypsin (Table 3) did not show a significant match with other proteins available in the NCBI blast database Anal-yses of these peptides using the NCBI blast database pertaining to fungi showed that some peptides in Table 3 had homology with different putative oxido-reductases (results not shown) In addition, the majority of peptides in Table 3 showed some homo-logy to a putative short-chain dehydrogenase from

A terreus NIH2624 and putative NADP-dependent flavin oxidoreductase from A nidulans FGSC A4 These peptide sequences (Table 3) should be sufficient for identification of the complete sequence of the enzyme when the genome sequence of L maculans is available [sequencing of the genome of L maculans

is in progress (http://www.genoscope.cns.fr/externe/ English/Projets/#region)]

Table 6 BO activity in plants infected with L maculans isolate BJ

125 Tissues of B napus cv Westar (susceptible) and B juncea cv.

Cutlass (resistant) were homogenized in buffer and protein extracts

were assayed for BO activity, as described in the Experimental

procedures BO activity was determined in protein extracts of

mycelia of L maculans isolate BJ-125 (control cultures and

cul-tures incubated with 3-phenylindole, 0.05 m M ) The results are

expressed as the means ± SD of four independent experiments.

lmolÆmin)1= U; ND, not detected.

Tissues analyzed for BO activity

Specific activity (mUÆmg)1)

Inoculated leaves – B napus 1.10 ± 0.05

Inoculated stems – B napus 1.41 ± 0.05

Control leaves of whole plants – B napus ND

Inoculated leaves of whole plants – B napus 0.52 ± 0.07

Control mycelia – L maculans Traces a

Mycelia incubated with 3-phenylindole –

L maculans

2.31 ± 0.15

a £ 0.01 mUÆmg)1.

The wide range of cofactors that serve as electron

acceptors of BO (PMS, small quinones or FMN)

dem-onstrate that BO is not selective with respect to

elec-tron acceptors (Table 4) Interestingly, PMS was a

more efficient electron acceptor than some natural

cofactors (e.g FMN, FAD) Because BO has no

cova-lently attached cofactor, as indicated by UV-visible

spectroscopic analysis, it is possible that natural

elec-tron acceptors of BO could be components of the cell

wall of L maculans Some fungi can produce

extra-cellular quinone derivatives used in the biosynthesis of

melanin [29] and other metabolites For example, the

brown rot fungus Gloeophyllum trabeum secreted two

quinone derivatives used to reduce Fe3+ and produce

H2O2[30]

In view of the important role of brassinin in crucifer

phytoalexin biosynthesis and its effective detoxification

by L maculans, inhibitors of BO are being developed

[7,13] Toward this end, the effects of the phytoalexins

camalexin, 1-methylcamalexin, cyclobrassinin and

rutalexin, the commercial fungicide thiabendazole, and

several synthetic compounds (see supplementary

Table S1) on BO activity were evaluated

Unexpect-edly, the phytoalexins camalexin and cyclobrassinin

were the best inhibitors of BO activity, whereas none

of the designed compounds (supplementary Table S1)

showed inhibitory effects In addition, none of these

compounds (supplementary Table S1) were

trans-formed by BO An additional surprise was revealed by

kinetic analyses of the inhibition of BO activity

because both camalexin and cyclobrassinin were shown

to be competitive inhibitors (Fig 4) These molecules

are the first inhibitors reported for a phytoalexin

detoxifying enzyme In addition, because these

inhibi-tors are also phytoalexins, this discovery indicates that

the various constituents of a phytoalexin blend have

multiple physiological functions For example, in

addi-tion to antimicrobial activity, constituents of these

blends may inhibit specific enzymes produced by

fun-gal pathogens Furthermore, it is of interest to note

that L maculans is able to metabolize and detoxify

cyclobrassinin but unable to metabolize camalexin [31]

Both camalexin and cyclobrassinin are biosynthesized

from l-tryptophan; however, although cyclobrassinin

is derived from brassinin and both co-occur in various

cultivated species [5], camalexin appears to be

pro-duced only in wild species (e.g Camelina sativa and

Arabidopsis thaliana) and is biosynthesized by a

diver-gent pathway [32] Furthermore, it should be noted

that camalexin (and the synthetic compound

3-pheny-lindole) could induce BO production substantially,

whereas the phytoalexin spirobrassinin (and

thiabenda-zole, a commercial fungicide) displayed no apparent

effect That the induction of BO was not related with the antifungal activity of these compounds was clari-fied by thiabendazole, which was a 50-fold more potent fungicide than camalexin but did not induce

BO [31] Due to the substantial inhibitory effect of camalexin on BO activity, a decrease of the rate of brassinin detoxification in cultures of L maculans co-incubated with brassinin and camalexin was expected However, our previous results did not show such a rate decrease [31] This apparent discrepancy between the results obtained with cell cultures [31] and the current results obtained with purified BO could be due to two opposite effects of camalexin: (a) induction

of BO and (b) inhibition of BO activity Therefore, the overall result was no detectable change in brassi-nin transformation rates in cultures of L maculans Nonetheless, because plants producing camalexin and brassinin were unknown until now, this apparent con-tradiction has not been investigated Without doubt, it would be most interesting to evaluate the disease resis-tance of such plants, which may be substantially higher because camalexin is not detoxified by crucifer patho-genic fungi such as blackleg or blackspot [7] and is a potent mycelial growth inhibitor of L maculans (com-plete inhibition at 0.5 mm) [31]

Recently, we proposed a mechanism for the trans-formation of brassinin to indole-3-carboxaldehyde [14], which invoked the formation of an imido dithiocar-bamate intermediate (I1) partly resembling a cyclo-brassinin structure, followed by formation of a fully conjugated intermediate (I2) partly resembling a cama-lexin structure (Fig 6) Because both cyclobrassinin and camalexin are competitive inhibitors of BO, these results lend support to the previously proposed reac-tion mechanism On the other hand, the absence of inhibition observed in the presence of N¢-methylbrassi-nin and 1-methylcamalexin suggests that these mole-cules do not fit in the active site of BO Furthermore, competitive inhibition is consistent with our steady-state kinetic studies indicating that BO followed an ordered kinetic mechanism (using PMS as electron acceptor and camalexin as dead-end inhibitor; Figs 4 and 5) This characteristic of BO is in contrast with flavoenzymes [33] and quinoenzymes [34,35] containing

a covalently bound cofactor, which are known to display a ternary complex or ping-pong kinetic mecha-nism Interestingly, plant cytokinin oxidases⁄ dehydrogenases (CKXs) catalyze the irreversible degra-dation of cytokinins (secondary amines) to aldehydes

in a single enzymatic step [36] This oxidative cleavage

of the side chain of cytokinins is somewhat related to the degradation of brassinin by BO In addition, some CKXs appear to be glycosylated and can transfer

electrons to artificial electron acceptors such as PMS

and coenzyme Q0 [37–40], similar to BO Yet, unlike

BO, CKXs have FAD covalently bound and the

cata-lytic cycle occurs through a ternary complex

mecha-nism [33] That is, comparison of the characteristics

and function of BO with ‘somewhat similar’ enzymes

emphasizes its uniqueness and explains its lack of

sequence homology to proteins available in current

databanks

Analysis of BO activity in plant tissues (stem and

leaf) susceptible and resistant to L maculans, cvs

Westar (B napus) and Cutlass (B juncea),

respec-tively, revealed that BO is produced only in

suscepti-ble plants (Tasuscepti-ble 6) That is, BO is an enzyme

produced in vivo in susceptible tissues but not in

resis-tant ones, during infection by L maculans

Further-more, production of BO in vitro fungal cultures

requires induction with specific compounds (e.g

3-phenylindole) (Table 6) Taken together, these

results demonstrate that BO is not an inconsequential

enzyme produced just when the pathogen has all

growth requirements satisfied By contrast, BO is

per-haps one of the best arms used by the pathogen

L maculans to overcome the inducible antifungal

plant defenses (phytoalexins) In this context, it is

pertinent to recall the precursor function of brassinin

vis-a`-vis phytoalexins and thus the negative impact on

the plant if it is depleted of it

Detoxification of phytoalexins from the family

Legu-minosae has shown the significance of phytoalexin

detoxification in the interaction of plants with fungi

[7] Pioneering work on the detoxification of the

phyto-alexin pisatin by pisatin demethylase, produced by the

plant pathogenic fungus Nectria haematococca,

demon-strated that this enzyme functioned as a virulence trait

[41] Such a precedent and our overall results indicate

that BO could be a virulence trait of L maculans as

well, a product of pathogen evolution over numerous life cycles of interaction with brassica plants

The apparent role of BO in the pathogenicity of

L maculansmay be confirmed once the gene(s) for this enzyme has been cloned Notwithstanding future dis-coveries, a first generation of BO inhibitors able to protect plants from fungal attacks by L maculans can now be modeled on the structural elements of camalexin, a ‘natural inhibitor’ In addition, purified

BO will facilitate in vitro evaluation and optimization

of such inhibitors, which could be developed into selective crucifer protectants after toxicity screens

Experimental procedures

General experimental procedures Chemicals and deglycosylating enzymes were purchased from Sigma-Aldrich (Oakville, Canada) and chromatogra-phy media and buffers from GE Healthcare (Quebec, Can-ada) HPLC analysis was carried out with a system equipped with a quaternary pump, an automatic injector, a photodiode array detector (wavelength range 190–600 nm),

a degasser and Hypersil octadecylsilane column (5 micron particle size silica, 200· 4.6 mm), and an in-line filter The retention times (tR) are reported using a linear gradient elu-tion with CH3CN-H2O, 25 : 75 to CH3CN, 100%, for

35 min at a flow rate of 1.0 mLÆmin)1 All operations regarding protein extraction, purification and assays were carried out at 4C, except where noted otherwise Solvents used in syntheses were treated as previously reported [13]

Fungal cultures Fungal spores of L maculans virulent isolate BJ 125 were obtained from the IBCN collection, Agriculture and Agri-Food Canada Research Station (Saskatoon, Canada)

Brassinin

Fig 6 Proposed mechanism of transformation of brassinin to indole-3-carboxaldehyde catalyzed by BO [14]: note the similarity of the chem-ical structures of the phytoalexins cyclobrassinin and camalexin and those of intermediates I 1 and I 2 , respectively.

Liquid cultures were initiated as described previously

[13,42] and induced with 3-phenylindole (0.05 mm) after

48 h The cultures were incubated for an additional 24 h

and then gravity filtered to separate mycelia from culture

broth

Preparation of protein extracts

Frozen mycelia (22 g) obtained from cultures of L

macu-lans(or plant tissues) were suspended in ice-cold extraction

buffer (20 mL) and ground (mortar) for 10 min The

extraction buffer consisted of 25 mm diethanolamine

(DEA) (pH 8.3), 5% (v⁄ v) glycerol, 1 mm d,l-dithiothreitol

and 1 : 200 (v⁄ v) protease inhibitor cocktail (P-8215;

Sigma-Aldrich) The suspension was centrifuged for 60 min

at 58 000 g The resulting supernatant (20 mL) was used

for chromatographic analyses

Chromatographic purification of the enzyme

exhibiting BO activity

In step 1, the soluble protein extract from mycelia (20 mL)

was equilibrated by dialyzing against 20 mm Tris–HCl

buf-fer (pH 8.0) containing 2% glycerol (v⁄ v) and loaded on a

DEAE-Sephacel (Amersham Biosciences, Uppsala, Sweden)

anion-exchange column (1.6· 12 cm) Proteins were eluted

with the same buffer, first alone and then with a 0.0–0.40 m

NaCl gradient Fractions (5 mL) were collected and 100 lL

assayed for BO activity Peak fractions (8–13) showing BO

activity were pooled and used in the second step of

purifica-tion In step 2, fractions showing BO activity from step 1

(30 mL) were concentrated to 6 mL, equilibrated in 25 mm

ethanolamine buffer (pH 9.4) and applied to a column

(0.9· 20 cm) of Polybuffer exchanger PBE 94 resin (GE

Healthcare) equilibrated in the same buffer Elution was

performed with Polybuffer 96, ten-fold diluted with distilled

water and adjusted to pH 6.0 Fractions of 3 mL were

col-lected and 50 lL of each fraction were assayed for BO

activity A peak of BO activity was observed at pH 7.1–7.2

In step 3, pooled fractions 38–40 showing BO activity after

step 3 were concentrated to 500 lL and fractionated by fast

protein liquid chromatography (GE Healthcare) on a

Superdex 200 HR10⁄ 30 column, pre-calibrated with the

fol-lowing markers of known molecular mass: bleu dextran

(2000 kDa), BSA (67 kDa), ovalbumin (43 kDa),

chymo-trypsin (25 kDa) and ribonuclease (13.7 kDa) Equilibration

and elution were performed at 8C with 25 mm Tris-HCl

(pH 8.0), 1% glycerol and 0.15 m NaCl Fractions of

0.5 mL were collected at a flow rate of 0.4 mLÆmin)1, and

10 lL of each fraction were assayed for BO activity In

step 4, the protein extract of 1.5 mL obtained from step 3

was equilibrated by dialyzing against 20 mm DEA buffer

(pH 8.3) and 1% glycerol The protein extract was loaded

on a Q-Sepharose (GE Healthcare) cation-exchange column

(1.0· 5 cm) The proteins were eluted with the same buffer,

first alone and then with a 0.0–0.3 m NaCl discontinuous gradient using 2.5 mL of NaCl solution, increasing by 0.025 m Fractions (1 mL) were collected and 50 lL assayed for BO activity Peak fractions 14–15 were pooled and concentrated to 500 lL, and then used for biochemical analysis

Analysis of deglycosylated BO Purified BO was treated with PNGase F (G5166) or endo-b-N-acetylglucosaminidase (A-0810) following the manu-facturer’s protocols Reactions were incubated at 37C overnight with 1 lL (7.7 units) of PNGase F in nondena-turing and 3 h in denanondena-turing (0.2% SDS, 50 mm b-mercap-toethanol and 1% of Triton X-100) conditions in the appropriate buffer (30 lL of total reaction volume) Endo-b-N-acetylglucosaminidase (1 lL: 5 mU) was incubated with purified BO at 37C for 3 h in denaturing (0.2% SDS, 50 mm b-mercaptoethanol) conditions with the appro-priate buffer (30 lL of total reaction volume) After incubation, 3 lL of SDS⁄ PAGE buffer was added to each reaction and samples were analyzed by SDS⁄ PAGE

SDS⁄ PAGE Protein-denaturing SDS⁄ PAGE was carried out using 10% polyacrylamide gels Standard markers (molecular mass range 25–200 kDa; Bio-Rad, Hercules, CA, USA) were used to determine the approximate molecular masses of purified proteins in gels stained with Coomassie brilliant blue R-250

Identification of tryptic peptides of BO by LC-ESI-MS⁄ MS

Analyses were carried out by the Plant Biotechnology Insti-tute, National Research Council of Canada (Saskatoon, Canada) Protein gel slice was manually excised from Coo-massie stained gels and placed in a 96-well microtitre plate The protein was then automatically destained, reduced with dithiothreitol, alkylated with iodoacetamide and digested with porcine trypsin [43] (sequencing grade; Promega, Mad-ison, WI, USA) and the resulting peptides transferred to a 96-well PCR plate· 3; all steps were performed on a Mass-PREP protein digest station (Waters⁄ Micromass, Manches-ter, UK) The digest was evaporated to dryness, then dissolved in 20 lL of 1% aqueous TFA, of which 5 lL was injected onto a NanoAcquity UPLC (Waters, Milford,

MA, USA) interfaced to a Q-Tof Ultima Global hybrid tandem mass spectrometer fitted with a Z-spray nanoelec-trospray ion source (Waters⁄ Micromass) Solvent A con-sisted of 0.1% formic acid in water, whereas solvent B consisted of 0.1% formic acid in acetonitrile The peptide digest sample was loaded onto a C18 trapping column