Mass spectrometric studies on the intrinsic stability of destruxin E fromMetarhizium anisopliae

RAPID COMMUNICATIONS IN MASS SPECTROMETRY

Rapid Commun. Mass Spectrom. 2004; 18: 2577–2586

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/rcm.1659

Mass spectrometric studies on the intrinsic stability

of destruxin E from Metarhizium anisopliae

Ed Dudley1*, Chengshu Wang2, Anke Skrobek2, Russell P. Newton1,2 and Tariq M. Butt2

1Biomolecular Analysis Mass Spectrometry Laboratory, University of Wales Swansea, Swansea SA2 8PP, UK2School of Biological Sciences, University of Wales Swansea, Swansea SA2 8PP, UK

Received 3 June 2004; Revised 1 September 2004; Accepted 1 September 2004

Destruxins are of current interest as bioactive agents. They are cyclic hexadepsipeptides produced

by fungi, the most common destruxins, A, B and E, differing in the structure of a side chain. Before

they can be widely used, the potential risk of destruxins and their metabolites entering the human

food chain must to be assessed; thus, knowledge of the structures of their degradation products is

essential. Here we report a study aimed at identifying, by tandem mass spectrometry and accurate

mass analysis, the products resulting from thermally and temporally induced degradation of des-

truxin E. The degradation products fell into two groups: those with relatively simple modifications

of the side chain and those involving much more complex rearrangements. The structures of most

of the degradation products were deduced from the MS data, with the major product being destrux-

in E diol: significantly, this compound had previously been reported to have only been produced as

a metabolic product of enzyme action rather than as a simple degradation product as demonstrated

here. Copyright # 2004 John Wiley & Sons, Ltd.

The entomogenous fungus Metarhizium anisopliae shows

considerable promise as a biological control agent (BCA)

for use in integrated pest management programmes,1 where

it may be applied as an environmental insecticide agent.2,3

Several strains have been commercialised yet still much infor-

mation is required by registration authorities regarding the

risks it may pose to human health and the environment. Of

particular concern are the bioactive metabolites secreted by

Metarhizium. One major group of metabolites produced by

this fungus consists of cyclic hexadepsipeptides called des-

truxins.4,5 These compounds are also known to be produced

by the insect pathogen Aschersonia sp.6 and the plant patho-

gen Alternaria brassicae.7 The most commonly produced

destruxins are A, B and E with the latter usually predominat-

ing.4,8 The chemical structures of cyclic hexadepsipeptide

destruxins A, B and E have the same backbone, with the gen-

eral formula: –D-HA1-L-Pro2-L-Ile3-L-MeVal4-L-MeAla5-b-

Ala6-, where HA is a D-a-hydroxy acid residue. The difference

is found in the R residue (see Fig. 4) bound to HA which is –

CH2–CH CH2, –CH2–CH(CH3)2, and for

dtx A, B and E, respectively. Some of the properties of these

compounds have been reviewed by Vey et al.5 and Pedras

et al.9 Little is known about the stability and fate of these com-

pounds in the environment yet such information could help

in risk assessment and possibly accelerate the registration

process. Mass spectrometry has previously been applied to

the analysis of destruxin metabolism,10 identification from

fungus extracts11 and their gas-phase fragmentation,12 and

was therefore considered to be a useful technique for the study

of destruxin E and related compounds. This paper provides the

first detailed mass spectrometric analysis of the intrinsic stabi-

lity of purified destruxins with attention focused on destruxin

E, the most bioactive of the metabolites,5,13 and its degradation

products. This study forms part of a major programme to deter-

mine if destruxins enter the food chain and pose a risk to

humans and the environment.

EXPERIMENTAL

ChemicalsAll chemicals used were purchased from Fisher Scientific

(UK) with the exception of leucine enkephalin purchased

from Sigma Aldrich (UK). Water for the analytical analysis

was prepared in-house using an Elix and Milli-Q system

(Millipore, UK).

Cultures and metabolitesMetarhizium anisopliae strains V275 (isolated fromCydia pome-

nella, Austria) and V245 (isolated from hay field soil, Finland)

were grown in Czapek Dox liquid media and crude extracts

of the metabolites prepared as described by Wang et al.13

Destruxin E (dtx E) was purified using a Dionex HPLC

system, equipped with a C18 reverse-phase column

(AcclaimTM, silica, particle size: 5mm, pore diameter: 120 A,

length: 4.6� 250 mm) and a UVD 340U diode-array detector

at a flow rate of 1 mL/min.13 Purified samples from different

batch cultures were stored at �208C until required.

Storage under different temperature conditionsTo evaluate the temperature stability of the purified dtx E

as well as the comparative stability in crude extracts, the

Copyright # 2004 John Wiley & Sons, Ltd.

*Correspondence to: E. Dudley, BAMS Laboratory, University ofWales Swansea, Swansea, SA2 8PP, UK.E-mail: [email protected]/grant sponsor: European Commission; contract/grant number: QLK1-2001-01391.

purified and crude extracts of metabolites were prepared in

brown high-performance liquid chromatography (HPLC)

vials (Fisher) in triplicate and stored at �20, 4, 25, and 378C.

Material was also stored in an unheated greenhouse which

fluctuated in temperature (range 10–508C between March

and September). Purified and crude metabolites were

retrieved from the above treatments at 1 month intervals, dis-

solved in 0.5 mL methanol, and analysed by HPLC under the

conditions described above. Peak areas of the destruxin com-

pounds and their degradation products (where applicable)

were measured. UV peaks arising from suspected degrada-

tion products were numbered in order to identify them in

further experiments. The breakdown ratio (BR) of dtx E at

monthly time points was calculated as:

BRð%Þ ¼ SA� FA

SA� 100;

where SA is the start amount (mg) and FA is the final amount

(mg) of the destruxin.

Analytical analysisAny significant degradation products detected from the

HPLC-UV data were analysed by mass spectrometry in

order to study the structures of these compounds after the

fractionation of the different peaks. For the fractionation of

the degradation samples, the samples were run on a Vision

workstation with an inbuilt AFC2000 fraction collector

(Applied Biosystems, UK). An Acclaim column and gradient

were used as described previously in this paper with occa-

sional lengthening in time of the gradient sections in order

to improve the resolution between some of the closely eluting

peaks. Fractions were collected during the HPLC separation

and important fractions identified from their UV response

and correlation with the previously observed chromato-

graphic peaks numbered as describer earlier. The fractions

were concentrated to approximately 100mL using a Speed-

vac concentrator 5301 (Eppendorf) and these concentrated

fractions were applied to the following electrospray mass

spectrometry analyses.

The samples were first loop injected (1 mL) into a flow of

50:50 methanol/water at a flow rate of 4mL/min from an

Ultimate pump using a FAMOS autosampler (Dionex, UK)

connected to an LCQ DECAXP Plus ion trap electrospray

mass spectrometer (Thermo Finnigan, UK), operated in

positive ion mode with a spray voltage of 3.5 kV, a sheath

gas flow rate (arbitrary units) of 35, a capillary temperature

of 2008C and an electron multiplier voltage of �890 V. Mass

spectra were collected between m/z 50 and 2000. Major

[MþH]þ ions detected by this method for each of the fractions

were then subjected to repeated injections during which

the precursor [MþH]þ ion was fragmented using various

collision energies (usually enough to result in a precursor ion

of at most 10% relative abundance) and MS/MS data was

collected. Where the product ions were of sufficient abun-

dance, these also were fragmented to provide MS3 and MS4

spectra. The fractions were also infused at 1 mL/min onto

an Ultima Q-ToF (Waters, UK) using a nanolockspray ion

source and a reference sprayer which electrosprayed a

solution of 250 leucine enkephalin (fmol/mL in 50:50 0.1%

formic acid/acetonitrile). The fractions were diluted or

concentrated sufficiently to produce a reasonable signal

intensity and for this to be less than 200 counts per second,

thereby avoiding a dead time effect that would reduce mass

accuracy. Accurate mass electrospray conditions required

a spray voltage of 3 kV, a cone voltage of 100 V, a source

temperature of 808C, a cone gas flow of 50 L/h and a

nanoflow gas level of approximately 7 psi. Full scan mass

spectra were obtained between m/z 50 and 1600 with a

reference spray scan being performed once every 5 s, and

MS/MS scans were performed in the same manner with

varying collision energies being applied to the analyte spray

(enough to sufficiently fragment the precursor ion) and a

collision energy of 10 being applied to the reference spray.

The resulting mass spectra were adjusted using the reference

spectra and MassLynx 4.0 software in order to obtain the

resulting accurate mass spectra and empirical formulae of the

degradation products.

RESULTS

Thermal instability of destruxin EAfter storage for 6 months, destruxin E of Metarhizium aniso-

pliae exhibited both time- and temperature-dependent degra-

dation (Fig. 1). The destruxin was most stable when kept at

�208C. Significant degradation was seen in samples kept at

other temperatures with more than 80% of dtx E degraded

after 6 months at increased temperatures. Storage at 48Cshowed slightly better stability characteristics than at the

higher temperatures monitored but no significant difference

was indicated in the stability of dtx E at the various tempera-

tures examined above 48C. No difference was observed

between the dtx breakdown ratios in crude mixture and in

the purified status of dtx E (P¼ 0.92) (Table 1), indicating

that the degradation monitored in the purified compound

reflects that in the crude extract in terms of the level of degrada-

tion at least. Destruxin E degradation resulted in 15 extra peaks

with various signal intensities from different temperature

storage conditions (Figs. 2 and 3). Of these, peaks 1–4 were

present in dtx samples before storage (Figs. 2(a) and 3(a))

and detectable throughout the storage under different condi-

tions (Figs. 3(a)–3(f)). The remaining peaks arose in the sam-

ples after storage, indicating the degradation of dtx E. Of the

new peaks, only peak 5 and trace amounts of peaks 7 and 8

were detected from the storage under �208C (Fig. 3(b)),

while the others were detected to varying degrees in storage

conditions of 4, 25, 378C and greenhouse (Figs. 3(c)–3(f)).

Figure 1. Breakdown dynamics of the purified dtx E under

different conditions.

Copyright # 2004 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2004; 18: 2577–2586

2578 E. Dudley et al.

Most significantly, the intensity of peak 11 was noticeably

higher than others (Fig. 3). Peak 9 was not included, based

on the analysis described in the following section.

Structural elucidation of destruxin E degradationFor the 15 peaks mentioned above, mass spectrometric data

concerning their structures could be obtained from the

combined electrospray analyses for the majority of the peaks

(Table 2). In order to study the structures of the peaks, the

fragmentation of dtx E was studied, and the information con-

cerning its ring opening and the positions of fragmentation

during MS/MS analysis are shown in Fig. 4. The fragmenta-

tion of the peptide, once ring opening had occurred, followed

the usual pattern of fragmentation of peptides, resulting

mainly in B-series ions. Although m/z 271 was not detected

it was recorded as another possible characteristic ion and

its formation was considered when studying the degradation

products. This information was then used to identify charac-

teristic ions from the fragmentation spectra of dtx E deriva-

tives as shown in Fig. 4. The product ions obtained from

the degradation products were then compared with those

numbered in Fig. 4 in order to obtain structural information

for the unknown compounds. By comparing the m/z values

for these diagnostic ions in dtx E and any such ions seen dur-

ing fragmentation of the degradation products and noting

their difference in mass, it was possible to identify the nature

and position of several of the modifications (Table 3, Fig. 5).

From these data, [MþH]þ ions at m/z 626 (peak 2), 612

(peak 11), 610 (peak 10), 654 (peak 8) and 630 (peak 13) are

proposed as arising from R group modifications to dtx E.

Table 3 shows the change in empirical formula between the

degradation products and dtx E and Fig. 5 shows the frag-

mentation spectra and interpretation for peak 11 (m/z 612).

Figure 6 shows the suspected R group modifications deduced

from the changes in empirical formulae for the other peaks

listed above. Interestingly, the structures of ions m/z 612

and 610 were the same as the previously reported dtx E diol

(Ed) and dtx F, respectively.9,14,15

It was suspected that the UV peak 4 (m/z 594) might

represent the co-purification of another destruxin, destruxin

B, as them/z value of this compound (non-accurate mass) and

fragmentation spectra would be expected to be essentially

identical to those of dtx E. However, with accurate mass data,

UV peak 4 is shown not to be due to destruxin B; the empirical

formulae of dtx B and E differ because of the different R

groups present, while them/z 594 ion from UV peak 4 and dtx

E have exactly the same empirical formulae (0 ppm difference

in the experimental and actual value for peak 4 and dtx E

empirical formula). Furthermore, the product ions matched

the empirical formulae expected for the dtx E fragments (with

the maximum difference from expected value being 5.6 ppm)

re-enforcing the conclusion that UV peak 4 was not destruxin

B, but a more chemically similar degradation product of

destruxin E. It was reasoned therefore that this ion was an

Table 1. Breakdown ratio (%) of Metarhizium anisopliae

destruxin E after storage under different conditions for

6 months

Temperatureof storage Purified dtx E

Dtx E fromV275 crude

extract

Dtx E fromV245 crude

extract

�208C 34.21� 2.36 35.25� 3.65 35.90� 2.6548C 60.53� 5.65 67.63� 6.35 66.67� 4.68258C 89.47� 7.12 87.05� 7.65 79.49� 3.72378C 81.58� 3.65 88.49� 4.64 74.36� 5.97GH 92.11� 6.23 89.21� 4.51 82.05� 6.31

*, GH, greenhouse; Values are mean� SD.

Figure 2. Selected HPLC chromatographic profiles of

destruxin E after storage for 6 months under different

conditions: (a) the chromatographic profile before storage;

(b, c) the chromatographic profiles after storage for 6 months

at conditions of 48C (b) and 258C (c).

Degradation products of destruxin E 2579


isomer of dtx E. Destruxin E has a number of chiral centres

within its structure that could give rise to the ion at peak 4.

Given the numerous degradation products arising from

structural changes in the R group, this latter part of the

molecule was considered to be a possible point for isomer-

isation. The proposed modification within the R group

requires the oxygen that forms the epoxy group being bound

to the first and second carbons of the side chain and the third

carbon being part of a CH3 group (Fig. 7(a)), although other

possible isomers could not be ruled out using the accurate

mass data obtained.

Peak 1 exhibited an [MþH]þ ion ofm/z 666 (Table 2) with an

empirical formula of C30H50N8O9 (0.5 ppm mass difference).

Fragmentation of this ion resulted in the ions at m/z 534, 437

and 324 and fragmentation of the m/z 534 ion produced the

m/z 437 and 324 ions (Fig. 7(b)). Very few product ions were

detected and therefore much of the structural characterisa-

tion was determined by the accurate mass analysis. The

formation of the structure shown would be rather complex

with the initial ring opening being followed by the replace-

ment of the oxygen of the hydroxy group of the hydroxyl acid

to which the R group is attached by a nitrogen group. Further,

Figure 3. Distribution and relative intensity of the peaks detected from the destruxin E samples after

storage for 6 months under different conditions: (a) the initial dtx E sample (before storage); (b) stored

at �208C; (c) stored at 48C; (d) stored at 258C; (e) stored at 378C; and (f) stored in a greenhouse. Y

axis, peak area of HPLC-UV signal.



loss of the C terminus of the ring-opened peptide up to the R

group would also be required. The R group of one such

molecule would then have to be modified and two of these

molecules (one modified, one unmodified) then have to join

to form the compound from which a protonated molecule at

m/z 666 would be produced (cf. Figs. 4 and 7(b) for structural

changes). The fragmentations shown in Fig. 6(b) were further

enforced by the accurate mass data with the formulae of the

fragments shown matching the observed data with differ-

ences of 0.5, 12.1 and 4.4 ppm for the ions at m/z 534, 437 and

324, respectively.

Fragmentation of the m/z 562 ion (peak 5) produced the

ions at m/z 437 and 324 as observed for peak 1 (Fig. 7(b)).

The accurate mass data from the m/z 562 ion suggested an

empirical formula of C28H44N5O7 (0.3 ppm) and the empirical

formulae for the product ions at m/z 324 and 437 matched

those of the product ions of peak 1 (ppm differences of 3.2 and

0.8, respectively). The data support the hypothesis that the

Figure 4. Ring opening and fragmentation of destruxin E under MS/MS conditions.



m/z 562 ion arose from the m/z 666 ion by loss of the N-

terminus section, as indicated in Fig. 7(b).

Peak 6 represents a compound that gives rise to a pro-

tonated molecule and its sodium adduct at m/z 684 and 706,

respectively. The accurate mass data for both ions indicate

that they differed from those expected for dtx E by the

addition of CH4N3O2 (4 ppm for m/z 684 and 2.7 ppm for

m/z 706). Fragmentation of them/z 684 ion produced very low

intensity MS/MS data; however, significant product ions

were detected at m/z 632, 571, 543, 526, 441, 413, 344, and 199

(Fig. 7(c)).The possibility that the m/z 684 ion was similar to

the ion at m/z 666 with the additional presence of an extra

water moiety was considered as they both exhibited a

product ion at m/z 324; however, the difference in fragmenta-

tion patterns and in the accurate mass data led to the rejection

of this option. The product ions of m/z 571 and 543 were

sufficient to give reliable accurate mass data and were shown

to be identical to the fragments 2 and 3 of dtx E (Fig. 4)

but with the addition of the extra atoms as described above

(7.7 and 8.5 ppm, respectively). Other product ions provided

unreliable accurate mass data and so the nominal masses

were considered along with the known fragmentation of

destruxin-like compounds. The mass differences between the

m/z 684 and the dtx E ions could generate an idea of the

structural differences between them but it was impossible to

identify exactly where the modification occurred on the R

group. The proposed structure represented a ‘best fit’ of the

data obtained (Fig. 7(c)). The asterisk in Fig. 7(c) denotes a

transfer of a proton from the ion formed in the mass spec-

trometer to the neutral C terminus to form the ion at m/z 441.

Peak 3 contained three significant ions: m/z 553 (approxi-

mately 30% relative abundance), m/z 1105 (100% relative

abundance) and m/z 1127 (10% abundance). Due to its high

m/z value, the m/z 1105 ion gave no accurate mass-defined

empirical formula and the fragmentation of this ion did not

yield reproducible MS/MS spectra. The m/z 553 ion was too

low in abundance to produce any useful fragmentation

spectra but an empirical formula of C28H49N4O7 could be

obtained (5.6 ppm). This empirical formula represented a loss

of NCO from, and an addition of a hydrogen atom to, dtx E.

The m/z 1105 ion corresponded to the protonated dimer

[2MþH]þ, a deduction supported by the presence of an ion of

m/z 1127 in which the proton is replaced by a sodium ion on

the dimer. The loss of the NCO from the dtx E would involve

ring opening and it is believed that this would occur at a

point from which NCO could be easily lost (Fig. 7(d)). It is

noteworthy that a second site also has the possibility of losing

HCN – indicated by the asterisk in Fig. 7(d), giving identical

accurate mass data. However, this possibility was considered

less likely as this site is adjacent to a proline amino acid

residue and these sections of peptide structures are not prone

to fragmentation or degradation due to the stability imparted

by the cyclic side chain structure of the proline amino acid.19

The ion atm/z 512 (peak 15) possessed an empirical formula

consistent with the structure of the protonated molecule of

dtx E minus a C4H4NO moiety. The MS/MS analysis of the

Table 3. Peaks with identities arising from R group modification of destruxin E

Ion (UV peak No.) Accurate mass formula Addition to E MS/MS ions detected (identity of the ion)

626 (2) C30H52N5O9 (2.8 ppm) CH4O 513(2), 485(3), 400(4), 303(5), 157(6), 114(8)654 (8) C31H52N5O10 (1.2 ppm) C2H4O2 541(2), 513(3), 428(4), 331(5), 157(6), 114(8)612 (11) C29H50N5O9 (1.9 ppm) H2O 499(2), 471(3), 386(4), 289(5), 157(6), 114(8)610 (10) C30H52N5O8 (1.3 ppm) CH4 497(2), 469(3), 384(4), 287(5), 157(6), 114(8)630 (13) C29H49N5O8Cl (6.3 ppm [MþH]þ, 1 ppm [MþNa]þ) HCl 489(3), 404(4), 307(5), 157(6), 114(8)

Table 2. Identifications and verifications of destruxin E and its degradation products

UV peak m/z detected Mass accuracy (ppm) LCQ MS2/3Q-ToF AccMass MS

Q-ToF Acc MassMS/MS Modifications

E 594 0.5 H H H —1 666 0.5 H H H RO2 626 2.8 H H H RGM3 553, 1127 5.6 NC H NC RO4 594 0 H H H RGM5 562 0.3 H H H RO6 684 4 H H H RO7 ND NA ND ND ND NA8 654 1.2 H H H RGM9 ACN NA ND ND ND NA10 610 6.3 (1.0 Na adduct) H H H RGM11 612 1.9 H H H RGM12 ND NA ND ND ND NA13 630 1.3 H H H RGM14 603 NA ND H H NA15 512 37.1 ND H H RO

H, detected; NA, not available; NC, not clear; ND, not detected; ACN, possible gradient acetonitrile; RO, ring opening; RGM, R group modifica-tions (Fig. 4).



compound gave product ions at m/z 324, 211 and 183 and the

structure of the proposed compound and its fragmentations

are outlined in Fig. 7(e). It is proposed that the compound

arose from the loss of CH3NHCHCH3CO from the N ter-

minus of the ring-opened dtx E (Fig. 4) with the donation of

protons from this lost group back to the degradation product.

For the described fragmentations, the accuracy of the m/z

values in relation to the expected ions was 37.1 ppm diffe-

rence between the observed m/z and that expected for the ion

atm/z 512 with differences of 2.6, 10 and 11.3 ppm for the ions

at m/z 324, 211 and 183, respectively. The mass accuracy for

the [MþH]þ ion at m/z 512 is low (possible due to an unstable

signal); however, the MS/MS data gave better mass accuracy

and helped support this identification.

The structures of peaks 7, 9, 12 and 14 were not deduced.

Peak 7 was too low in intensity to be detected with any MS

Figure 5. The proposed ring opening and fragmentation of peak 11 (m/z 612) under MS/MS conditions.



data (Fig. 2(c)). Peak 9 was thought to arise from the acetoni-

trile gradient as the peak occurred around the time when the

HPLC gradient reached 100% organic solvent and mobile

phase related peaks are commonly detected in UV detection

at these time points. Peak 12 co-eluted with peak 11 which

was far in excess and masked the identity of the former (as

shown in Figs. 2(b) and 2(c)). Despite altering the gradient

used for the HPLC run, no improved separation was obtained

between these two peaks. Peak 14 was again of low intensity

but gave a possible [MþH]þ ion atm/z 603; however, its signal

intensity was too low to provide any further useful mass

spectrometric data for structural elucidation.

DISCUSSION

Purified destruxin E has been shown to degrade when stored

at all the temperatures examined, with the degradation

becoming more pronounced at temperatures at or above

258C. The degree of degradation in the purified sample has

also been shown to be mimicked in crude extracts; however,

to date no data concerning the comparative degradation

products from crude extracts has been obtained. The data

allowed the study of the degradation products formed from

purified destruxin E and suggested that two classes of degra-

dation product are formed (Table 2). The first group of degra-

dation products arises purely as a result of the R group of

the destruxin E being modified in some way; these modi-

fications are summarised in Table 3. The second class of

degradation products is formed by the ring opening of the

destruxin E followed by loss of parts of the structure and/

or modifications of the structure. This second class of product

proved more complex and therefore more difficult to obtain

useful structural information for. Highly intriguingly, among

15 detected peaks, mass spectrometric analysis of peaks 11

and 10 identified them as destruxin E diol (dtx Ed) and

destruxin F, respectively.14,15 Dtx Ed has been reported to

be easily formed from dtx E.8,9,16 The less toxic dtx Ed was

previously described as dtx E17,18 and later confirmed as a

new toxin.14,15 Dtx Ed has always been thought to arise

from the enzymatic hydrolysis of dtx E inside the mycelium

with its levels increasing with a comparative decrease in

destruxin E levels;8 however, its observation in this study

suggests that the conversion of dtx Ed from dtx E is a non-

enzymatic process since it could occur under storage condi-

tions and exhibits the change in comparative amounts as

previously reported.8 Dtx F was reported previously as a ‘nat-

ural’ product from the culture filtrate of M. anisopliae;14 the

data in our study suggests that it may instead be a degrada-

tion product formed from destruxin E rather than a newly

synthesised toxin. Peaks 1–4 were pre-existing in the dtx E

sample before the start of storage experiments. Mass analysis

indicated these ions were each related to the structure of dtx

E, eliminating the possibility of contamination during HPLC

purification; however, the dtx E derivates could have arisen

during the sample preparation storage under �208C for

1 month before the start of the experiments. In this respect,

peak 5, with an [MþH]þ ion at m/z 562, could be speculated

to be a derivative arising from peak 1 (m/z 666) rather than

arising directly from dtx E. This argument is based on the

structural similarities between the two compounds as peak

1 appears to lose part of its structure to form peak 5. Also,

as shown in Fig. 3, peak 1 decreases in UV response over

the different temperatures whereas peak 5 shows a compara-

tive increase in signal, adding weight to the idea that peak 1

degrades to form peak 5. Of further interest is peak 13

(m/z 630) which corresponds to destruxin E chlorohydrin

(again previously believed to be a new toxin released by

Metarrhizium anisopliae11 and shown here to arise from sam-

ple degradation). This degradation product contains a chlor-

ine atom which was not added to the sample at any point

during storage. A possible explanation for the formation

of this compound arises from the fact that destruxin E was

extracted in dichloromethane and this might cause contami-

nation of the storage solution with chlorine. Whilst this may

explain the formation of the chlorohydrin product in this pre-

sent study it is unknown whether the degradation reaction of

destruxin E with chloride ions is responsible for the presence

of this compound in fungal extracts. The other degradation

products detected involve ring opening with some loss of

the structure (peak 3 and 15) or possibly extensive modifica-

tion of the structure (peak 6).

The R group modified peaks give reliable structural

information and accurate mass and fragmentation data

allow the investigation of the structure of some of the more

unusual degradation products. This is demonstrated by the

fact that only three of these products have a mass lower than

their parent compound, dtx E, suggesting that complex

rearrangements occurred during degradation and these are

comparatively more complicated to elucidate. We feel that

the structures presented here for such compounds accurately

reflect the empirical formulae suggested by the accurate mass

analysis, and represent the most probable structures based

both on these data and on the fragmentation data obtained,

despite the complications arising from such unusual trans-

formations. In this study, thermal instability and degradation

of dtx E were studied under laboratory conditions; however,

environmentally, the detoxification/modification of de-

struxins by other biotic and abiotic factors is still poorly

Figure 6. R group modifications to destruxin E during

degradation.



understood. Further studies are required to elucidate the

persistence of destruxins in soil, their entry into the food

chain, their modifications by other organisms, plus particu-

larly the toxicity study of the degraded products identified in

this study.

AcknowledgementsThis work was supported by the Quality of Life and Manage-

ment of Living Resources Programme of the European

Commission, Key Action 1 on Food, Nutrition and Health,

QLK1-2001-01391 (RAFBCA).

Figure 7. Other degradation products: (a) isomerisation of the R group in peak 4; (b) identity and fragmentation of peaks 1 and

5; (c) identity and fragmentation of peak 6; (d) identity and fragmentation of peak 3; and (e) identity and fragmentation of peak 15.



REFERENCES

1. Butt TM, Jackson CW, Magan N (eds). Fungi as BiocontrolAgents: Progress, Problems and Potential. CAB International:Wallingford, UK, 2001.

2. Ignoffo CM. In CRC Handbook of Natural Pesticides, vol. V.Microbial insecticides, Part A, Entomogenous protozoa andfungi. CRC Press Inc.: Boca Raton, 1988; 243.

3. Vestergaard S, Gillespie AT, Butt TM, Schreiter G, EilenbergJ. Biocontrol Sci. Technol. 1995; 5: 185.

4. Amiri-Besheli B, Khambay B, Cameron S, Deadman ML,Butt TM. Mycol. Res. 2000; 104: 447.

5. Vey A, Hoagland R, Butt TM. In Fungi as Biocontrol Agents:Progress, Problems and Potential, Butt TM, Jackson CW,Magan N (eds). CAB International: Wallingford, UK,2001; 311–345.

6. Krasnoff SB, Gibson DM, Belofsky GN, Gloer KB, Gloer JB.J. Nat Prod. 1996; 59: 485.

7. Pedras MSC, Zaharia IL, Gai Y, Zhou Y, Ward DE. Proc. NatlAcad. Sci. USA 2001; 98: 747.

8. Loutelier C, Cherton J-C, Lange C, Traris M, Vey A.J. Chromatogr. A 1996; 738: 181.

9. Pedras MSC, Zaharia LI, Ward DE. Phytochemistry 2002; 59:579.

10. Hubert M, Cherton J-C, Vey A, Lange C. Rapid Commun.Mass Spectrom. 1999; 13: 179.

11. Potterat O, Wagner K, Haag H. J. Chromatogr. A 2000; 872:85.

12. Loutelier C, Hubet M, Marcual A, Lange C, Mercadier J-L,Cavelier F, Verducci J, Jacquier R. Rapid. Commun. MassSpectrom. 1995; 9: 1512.

13. Wang C-S, Skrobek A, Butt TM. FEMS Microbiol. Lett. 2003;226: 373.

14. Wahlman M, Davidson BS. J. Nat. Prod. (Lloydia) 1993; 56: 643.15. Jegorov A, Sedmera P, Havlicek V, Matha V. Phytochemistry

1998; 49: 1815.16. Wang C-S, Skrobek A, Butt TM. J. Invertebr. Pathol. 2004; 85:

168.17. Cherton J-C, Lange C, Mulheim C, Pais M, Cassier P, Vey A.

J. Chromatogr. B 1991; 566: 511.18. Lange C, Loutelier C, Cherton JC, Cassier P, Vey A, Pais M.

Rapid Commun. Mass Spectrom. 1992; 6: 28.19. Liebler DC. Introduction to Proteomics: Tools for the New

Biology. Humana Press, Inc.: Towowa, NJ, 2003; 89–99.



https://www.researchgate.net/publication/288951951_Fungi_as_Biocontrol_Agents_Progress_Problems_and_Potential?el=1_x_8&enrichId=rgreq-d05f30d8-8dce-40c1-a3f8-5a228e855412&enrichSource=Y292ZXJQYWdlOzgyNDc4MjU7QVM6MTA0MjgyNzY2NDQ2NTk4QDE0MDE4NzQzNTc4NzI=



Mass spectrometric studies on the intrinsic stability of destruxin E fromMetarhizium anisopliae

Documents