90) Andolfi Et Al., Review Perspective

7/31/2019 90) Andolfi Et Al., Review Perspective

1/32

Perspectives in Medicinal Chemistry 2008:2 81112 81

PERSPECTIVE

Correspondence: Antonio Evidente, Dipartimento di Scienze del Suolo, della Pianta, dellAmbiente e delleProduzioni Animali, Universit di Napoli Federico II, Via Universit 100, 80055 Portici, Italy.Tel: +39 081 2539178; Fax: +39 081 2539186; E-mail: [email protected]

Copyright in this article, its metadata, and any supplementary data is held by its author or authors. It is published under the

Creative Commons Attribution By licence. For further information go to: http://creativecommons.org/licenses/by/3.0/.

Bioactive and Structural Metabolites ofPseudomonasand Burkholderia Species Causal Agents of CultivatedMushrooms Diseases1

Anna Andol fi1, Alessio Cimmino1, Pietro Lo Cantore2, Nicola Sante Iacobellis2and Antonio Evidente1

1Dipartimento di Scienze del Suolo, della Pianta, dellAmbiente e delle Produzioni Animali, Universitdi Napoli Federico II, Via Universit 100, 80055 Portici, Italy. 2Dipartimento di Biologia, Difesa e Bio-tecnologie Agro-Forestali, Universit degli Studi della Basilicata, Viale dellAteneo Lucano 10, 85100Potenza, Italy.

Abstract:Pseudomonas tolaasii,P. reactans andBurkholderia gladioli pv. agaricicola, are responsible of diseases on somespecies of cultivated mushrooms. The main bioactive metabolites produced by both Pseudomonas strains are the lipodepsip-eptides (LDPs) tolaasin I and II and the so called White Line Inducing Principle (WLIP), respectively, LDPs which have beenextensively studied for their role in the disease process and for their biological properties. In particular, their antimicrobialactivity and the alteration of biological and model membranes (red blood cell and liposomes) was established. In the case oftolaasin I interaction with membranes was also related to the tridimensional structure in solution as determined by NMR com-

bined with molecular dynamic calculation techniques. Recently, five news minor tolaasins, tolaasins A-E, were isolated fromthe culture filtrates ofP. tolaasii and their chemical structure was determined by extensive use of NMR and MS spectroscopy.

Furthermore, their antimicrobial activity was evaluated on target micro-organisms (fungiincluding the cultivated mushroomsAgaricus bisporus,Lentinus edodes, andPleurotus spp.chromista, yeast and bacteria). The Gram positive bacteria resultedthe most sensible and a significant structure-activity relationships was apparent. The isolation and structure determination of

bioactive metabolites produced byB. gladioli pv. agaricicola are still in progress but preliminary results indicate their peptidenature. Furthermore, the exopolysaccharide (EPS) from the culturefiltrates ofB. gladioli pv. agaricicola, as well as the O-chainand lipid A, from the lipo-polysaccharide (LPS) of the three bacteria, were isolated and the structures determined.

Keywords: cultivated mushrooms and bacterial diseases,Pseudomonas tolaasii,P. reactans andBurkholderia gladioli pv.agaricicola, lipodepsipetides, mycopathogenic bacteria, antimicrobial activity, permeabilising effects on membranes, exo-

polysaccharides and lipo-polysaccharides

Biography:Antonio Evidente is Professor of Organic Chemistry. He received his degree in Chemistryin 1975 at University of Naples Federico II. He worked as a chemist of bioactive natural compounds atthe University of Naples Federico II and from 19871989 at the University of Basilicata, Potenza. Dr.Evidentes research has been mainly devoted to the biosynthesis, biochemistry, chemistry, spectroscopyand synthesis of bioactive metabolites (alkaloids, phytotoxins, plant growth regulators, antibiotics,mycotoxins, fungicides, phytoalexins, herbicides, proteins and polysaccharides) produced by phyto-

pathogenic fungi and bacteria and plants. In particular, the stereostructural determination and synthesisof bioactive metabolites and studies on structure-activity relationships and on their action mode were

performed by using advanced spectroscopic and chemical methods. He is a leader of national andinternational research projects on the isolation and structure determination of bioactive microbial and

plant metabolites including lipo- and exopolysaccharides. Some of these have been used to correlate the tridimensionalstructure in solution with their role in the plant pathogenesis and as tool for the correct classification of the producer micro-organisms. He is author of more than 230 publications on international journals and books.

Biography: Anna Andolfi is researcher of Organic Chemistry in Agriculture Faculty of University ofNaples Federico II. She was graduated in Pharmaceutics Chemistry and Technology at University of NaplesFederico II, with the maximum score, in 1993. She attained the Dottorato degree in Agricultural Chem-istry in 1997. Anna Andolfi has over 15 years experienced in organic chemistry, mainly in chemistry and

spectroscopy of bioactive metabolites (phytotoxins, plant growth regulators, antibiotics, mycotoxins,fungicides, phytoalexins, elicitors, herbicides, proteins and polysaccharides) produced by phytopathogenicfungi and bacteria and plants.

1This paper is dedicated to the memory of Gianfranco Menestrina
http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/


2/32

82

Andolfi et al

Perspectives in Medicinal Chemistry 2008:2

Anna Andolfi participated to national in national andinternational research projects.She is co-author of 50 publications on international journals.

Biography:Alessio Cimmino graduatedwith the maximum score in FoodScience and Technology at Universityof Naples Federico II. After thegraduation he studied in detail the

isolation and chemical characterizationof natural compounds by usingchromatographic and spectroscopictechniques. At the moment is frequenting the third year ofDottorato studies in Agrobiology and Agrochemistry at theDepartment of Soil, Plant, Environment Science and AnimalProduction, University of Naples Federico II, SupervisorProfessor Antonio Evidente. Currently he is performing hisresearch work in Research Center Borstel, Leibniz Center forMedicine and Biosciences, Department of StructuralBiochemistry, Supervisor Professor Otto Holst. He is co-author of 11 publications on internationals journals.

Biography: Nicola Sante Iacobellis is

Professor of Plant Pathology, Universitdegli Studi della Basilicata (2000-to date).He was Associate Professor of PlantPathology at the above University(19922000) and Research scientist of

National Council of Research (CNR) ofItaly, Istituto Tossine e Micotossine daParassiti Vegetali, CNR (19831992)

being the Scientific responsible of thesection Phytotoxins (19881992). Hespent several periods at foreign laboratory. He was theresponsible of several research projects on Molecular aspectsof Plant-Microbe interactions, epidemiology and advanceddiagnosis of plant pathogenic bacteria, new alternative and eco-

compatible control strategies of bacterial and fungal plantdiseases. He was the chairmen and member of committees forthe organization of several national and international Congresses.He was the senior editor of the bookPseudomonas syringae andrelated pathogens and author of about 280 publications whichappeared on journals, proceedings of meetings and books.

Biography:Pietro Lo Cantore graduatedin Agriculture Science at Universitdegli Studi della Basilicata in 1997.On the 2001 he attained the Dottoratodegree in Applied Biology with athes i s on : P hys io log i ca l andm o l e c u l a r c h a r a c t e r i z a t i o n

ofPseudomonas tolaasii andP.reactans:role of tolaasin and WLIP in pathogenesis.During 2001 Pietro Lo Cantorefrequented the: Reparto di Biomolecole e MembraneBiologiche c/o Centro di Fisica Stati Aggregati di Povo(Trento) and course on Biology of the relationships source-sink organized by the Italian Society of Plant Physiology.He was the scientific responsible of the Project Biologicalcharacterization of the tolaasin and WLIP lipodepsipeptidesof Pseudomonas tolaasii andP.reactans for Young Research-ers financed by the Universit degli Studi della Basilicata.

From 2002 to date he is research fellow at the above universityand is mainly involved an research programmes an molecularaspects of plant-microbe interaction and on new alternativeand eco-compatible control strategies of bacterial and fungal

plant diseases. He is member of the Italian Society of PlantPathology and is author of 87 publications appeared oninternational and national journals and books.

IntroductionBrown blotch ofAgaricus bisporus (Lange) Imbach(Tolaas, 1915) and the yellowing ofPleurotus ostreatus(Jacq.: Fr.) Kum (Ferri, 1985, pp 398; Iacobellis andLo Cantore, 1997; Iacobellis and Lo Cantore, 1998a)are important diseases of the above cultivated mush-rooms known to be caused byPseudomonas tolaasii(Paine, 1919). However, recent studies have shownthat the above diseases are complex ones. In particu-lar, besidesP. tolaasii alsoP. reactans, a saprotrophassociated with cultivated mushrooms not yet clas-sified and known in bibliography for its use in thewhite line assay for the specific identification ofP.tolaasii (Wong and Preece, 1979), is responsible,though in different manner, of the symptom develop-ment on the hosts. Furthermore,P. reactans, has beenshown to be the casual agent of the yellowing ofP.eryngii (D.C.: Fr.) Qul a mushroom species culti-vated in southern Italy (Iacobellis and Lo Cantore,1997; Iacobellis and Lo Cantore, 1998b; Lo Cantore,2001; Lo Cantore and Iacobellis, 2002; Iacobellis andLo Cantore, 2003). The pathogenicity ofP. reactansmay be considered a novelty though indications in

this regard have been reported in independent studies(Kim et al. 1995; Wells et al. 1996).Burkholderia gladiolipv. agaricicola is the causal

agent of the soft rot disease of the edible mushroomA. bitorquison which it cause the rapid developmentof deep oozing lesions on the pileus which rendersthe mushroom unmarketable (Atkey et al. 1991,p 431; Lincoln et al. 1991). The disease appeargeographically limited since it has been reported firstin England (Lincoln et al. 1991) and then in NewZealand (Gill and Cole, 1992) on A. bitorquis.Preliminary analysis of the partially purified

antimicrobial metabolites produced by ofB. gladiolipv.agaricicola indicate their peptide nature. After that,the disease was reported on several common Japanesecultivated mushroom (Gill and Tsuneda, 1997).

Isolation of Tolaasin I from CultureFiltrate ofPseudomonas tolaasiiVirulent strains of phytopathogenic as well asmycopathogenic Pseudomonas spp. produce


3/32

83

Bioactive and structural metabolites ofPseudomonas and Burkholderia species


in vitro toxic and antimicrobial lipodepsipeptides(LDPs) containing unusual aminoacids also witha D-stereochemistry and are classified in twogroups according to their primary structures. Thefirst group includes nonapeptides such assyringomycins (Fig. 1), syringotoxins, syringostatinsand pseudomycins. The second group comprises

molecules containing 18 to 25 amino acid residues,most of which having a D-stereochemistry, such assyringopeptins (Fig. 2), tolaasins, fuscopeptins andcorpeptins, (Bender et al. 1999). In the latter groupthe C-terminal region forms a lactone ring of 5(tolaasins, fuscopeptins and corpeptins) to 8(syringopeptins) amino acids.

Tolaasin I (8, Fig. 3), a lipodepsipeptide pro-duced in culture by virulent strains ofP. tolaasii(Rainey et al. 1991), bears a -OH octanoic acidblocking group at the N-terminus, a sequence ofseven successive D-amino acids in the N-terminal

region of the peptide (Pro 2 -Val 8), with a Ser-Leu-Val repeat, and then alternate L- and D-aminoacids (Nutkins et al. 1991) (Fig. 3). It also con-tains aZ-dehydroalanineBut) amino acid at posi-tions 1 and 13, and a D-homoserine (Hse 16) anda D-2,4-diaminobutyric acid (D-Dab 17). Finally,

a lactone ring is formed between the hydroxyl ofD-Thr 14 and the C-terminal L-Lys 18.

Biological studies demonstrated that tolaasin Iexhibits antimicrobial activity, strong resistance toenzymatic degradation and inactivation as wellantigenicity (Rainey et al. 1991). This suggests fortolaasin I a potential role as a therapeutic peptide.

Tolaasin I is considered the main virulence fac-tor ofP. tolaasii and responsible for the symptomsdevelopment on the mushrooms (Rainey et al.1991; 1992; Soler-Rivas et al. 1999a) and this isapparently due to the ability of this molecule todisrupt cell membrane function by the transmem-brane pore formation (Brodey et al. 1991; Raineyet al. 1991; Hutchinson and Johnstone, 1993).Tolaasin I was isolated from the cell-free bacterialculture filtrate of the type strain NCPPB2192 ofP. tolaasii following a modification of the methoddescribed by Peng (1996). The further purification

of the tolaasin I preparation by HPLC showed thatP. tolaasii, besides the already known tolaasin Iand II, which differ for the presence of Hse andGly, respectively, in position 16, (Nutkins et al.1991), produced in vitro other chemical correlatedmetabolites confirming previous findings (Shirataet al. 1995). HPLC profile of a crude tolaasinpreparation (Fig. 4) carried out on anAnalyticalBrownlee Aquapore RP-300 (C8) column usingwater (0.1% TFA, v/v) and CH3CN as solventsystem with a flow rates of 1.0 ml/min-1 [a Brown-lee Aquapore RP-300 (C8) column using the same

solvent system with flow rate 4.7 ml/min was usedfor semipreparative purpose] showed the presenceoffive minor peaks, namedtolaasins A, B, C, D,and E (Bassarello et al. 2004).

The chemical nature of tolaasin I was ascertainedby comparing its NMR and MS data with thosereported in the literature (Nutkins et al. 1991).

Isolation of White Line InducingPrinciple (WLIP) from CultureFiltrate ofPseudomonas reactansVirulent strains ofP. reactans produce in culturean extracellular substance called the White LineInducing Principle (WLIP, 9, Fig. 5), known forits ability to interact with tolaasin I and form awhite precipitates in the white line assay usefulfor the identification ofP. tolaasii (Wong and Pre-ece, 1979). The ability of WLIP to precipitatetolaasin I, the main virulence factor ofP. tolaasii,has been suggested to be useful for the control of

CH2

O

CO CH-NH-HyAsp-DhThr

CHOH

CH2Cl

R-CO-NH-CH-CO-Ser-Dab-Dab-Arg-Phe

1 SR-E: R = -CH2-CH-(OH)-(CH

2)

8CH

3

2 SR-A1: R = -CH

2-CH-(OH)-(CH

2)

6CH

3

3 SR-G: R = -CH2-CH-(OH)-(CH

2)

10CH

3

Dab = diaminobutyric acid

HyAsp = 3-hydroxyaspartic acid

DhThr = 2,3-dehydrothreonine

Figure 1. Structure of syringomicins A1, E and G (1, 2 and 3,respectively).


4/32

84

Andolfi et al


Figure 2. Structure of syringopeptins 22A, 22B, 25A and 25B (4, 5, 6 and 7, respectively).

N

HN

O

NHH

OH

O H HN

O

NH

HN

NH

HN

NH

HN

NH

O

O

O

O

O

H

H

OH

H

H

H

H

CONH2

O

O

HN N

HH O

OHN

O

NH

O

H

HN

NH

O

O

H

OH

H

NH2O

O

H

NH2

H

H

O

NHO

4

HO

H

8

Tolaasin I: -hydroxyoctanoyL--But1-DPro-DSer-DLeu-DVal5-DSer-DLeu-DVal-LVal-DGln10-LLeu-DVal--But-DalloThr-LIle15-LHse-DDab-LLys18

But=z-dehydroalanine

3

Figure 3. Tolaasin I (8) primary structure.

R-CO-NH-C-CO-Pro-Val-Ala-Ala-Val-Leu-Ala-Ala-Dhb-Val-Dhb-Ala-Val-Ala-Ala-Dhb-NH-CH-CO-Ser-Ala-Val-Ala

O

CO CH- NH -Dab-DabH2C

OH

CH

CH3

6 SP 25A R = -CH2-CH-(OH)-(CH2)6-CH3

7 SP 25B R = -CH2-CH-(OH)-(CH2)8-CH3

R-CO-NH-C-CO-Pro-Val-Val-Ala-Ala-Val-Val-Dhb-Ala-Val-Ala-Ala-Dhb-NH-CH-CO-Ser-Ala-Dhb-Ala

CH 3

O

CO CH-NH- - -Dab Dab

H2C

OH

CH

CH3

4 SP 22A R = -CH2-CH-(OH)-(CH2)6-CH3

5 SP 22B R = -CH2-CH-(OH)-(CH2)8-CH3

Dhb=dehydroaminobutyricacid

CHCH 3


5/32

85



Figure 4. HPLC profile of crude tolaasins preparation.

HN

HOH

N

NH

O

OH H

NNH

O

OHN

NH

O

O

HNH

O

OH

O

H

HO H

H

H

OHCOOH

OH

H

NH

H

O

WLIP: -hydroxydecanoyl-L-Leu-DGlu-Dallo-Thr3-DVal-DLeu5-DSer-LLeu-DSer-LIle10

9

Figure 5. WLIP Primary structure (9).


6/32

86

Andolfi et al


brown blotches caused byP. tolaasii (Soler-Rivaset al. 1999b). WLIP, whose biological features aresubstantially unknown, is a lipodepsipeptide witha molecular weight of 1125 Da composed of aN-terminal -hydroxydecanoic acid and a peptitemoiety of nine aminoacids, six of which ofD-form. The molecule presents a lactone ring

between D-allo-threonine (D-Thr3) and N-terminalL-isoleucine (Ileu9) (Mortischire-Smith et al.1991b). WLIP is structurally similar to viscosin(Laycock et al. 1991), except for the chirality ofleucine in position 5 that it is D- in WLIP and L- inviscosin. This substance is a potent biosurfactantthat appears important in the biology of the phy-topathogenic bacteriumP.fluorescens (Neu et al.1990; Laycock et al. 1991).

The isolation and purification of crude WLIPfrom the culture filtrates of strain NCPPB1311 ofP. reactans was performed according to the proce-

dure reported by Mortishire-Smith et al. (1991b).Crude WLIP was obtained by fractionated precipi-tation in acid conditions whereas crystalline WLIPwas prepared by the diffusion of water vapour intoa solution of crude WLIP in methanol.

The identity of WLIP was ascertained bycomparing NMR and MS data with those reportedin literature (Mortishire-Smith et al. 1991b). Thedetermination of the absolute structure of thelipodepsipetide by X-ray analysis allowed to ruledout its supposed identity with viscosin from whichdiffer in the form of the crystalline cell.

Strain NCPPB1311 ofP. reactans produced inculture high levels of WLIP with a yield of 169 mg/lwhich was more than ten fold higher of tolaasin I(13 mg/l) produced by strain NCPPB2192 ofP.tolaasii (Lo Cantore et al. 2006).

Biological Charaterizationof Tolaasin I and of WLIP

Antimicrobial activity of bacterialcultures and pure toxinsThe antimicrobial assays of HCLP grade tolaasin Iand crystalline WLIP confirmed the antimicrobialactivity of tolaasin I (Rainey et al. 1991; 1992)and showed that also WLIP is a lipodepsipeptideinhibiting the growth of bacteria and fungi(Tables 1 and 2) (Lo Cantore et al. 2006). In general

tolaasin I showed an antimicrobial activity higherthan WLIP. The Gram positive bacteriumBacillusmegaterium resulted the most sensible micro-organism to both tolaasin I and WLIP with aMinimal Inhibitory Quantity (M.I.Q) of 0.04 gand 0.32 g, respectively (Table 1). Also the otherGram positive bacteria used in this study wereinhibited by both LPDs with M.I.Q significantlydifferent onRhodococcus fascians and Curtobac-triumflaccufaciens (Table 1). On the contrary theGram negative bacteria, includingEscherichia coli,were inhibited by tolaasin I but not by WLIP

Table 1. Antimicrobial activity of tolaasin I and WLIP towards bacteria.

Bacteria No strains Minimal Inhibitory Quantity (g)

Tolaasin I WLIP

Gram positiveBacillus megaterium 1 0.04 0.32Rhodococcus fascians 2 0.16 1.28Clavibacter michiganensis subsp. michiganensis 2 0.32 0.64C. michiganensis subsp. sepedonicus 1 0.32 0.64Curtobacterium flaccunfaciens 1 0.16 0.64

Gram negativeEscherichia coli 1 0.64 -b

Pseudomonas tolaasii 1 -a P. reactans 1 1.28 P. syringae pv.phaseolicola 1 1.28 P. corrugata 1 0.64

Agrobacterium tumefaciens 1 0.64 Erwinia herbicola 1 0.64

Xanthomonas campestris pv.phaseoli 1 1.28

ano inhibitory activity were observed with a quantity of tolaasin I equal at 1.28 g.bno inhibitory activity were observed with a quantity of WLIP equal at 10.24 g.


7/32

87



(Table 1). It is of interest the fact that either culturesof strains NCPPB2192 or NCPPB1311 ofP. tolaasii

and P. reactans, respectively, inhibited thegrowth ofE. coli K12 but only when the assayswere performed on a minimal medium agar (MMA)(Lavermicocca et al. 1997). In fact, the aboveactivity was not observed when the assays wereperformed on KB plates. The fact that the presenceof the E. coli growth inhibiting metabolites wasobserved only on a minimal medium and the activ-ity was reversed when peptone was added in theassay media suggest that eitherP. tolaasii andP. reactans produce besides tolaasins or WLIP,

other metabolites, whose chemical nature and rolein the biology of the producers should be investi-gated. It is not excluded that also the above bacte-ria produce metabolites similar to those producedbyP. syringae strains involved in the inhibition ofamino acids biosynthetic pathways (Bender et al.1999; Iacobellis and Latorraca, 1999; Arrebolaet al. 2003; Cazorla et al. 2003).

In general, filamentous fungi, includingA. bisporus, P. ostreatus, P. eryngii andLentinus

edodes, were inhibited by both LPDs but onaverage tolaasin I was about eight/ten fold more

active than WLIP (Table 2).Furthermore, tolaasin I inhibited the growth

of Candida albicans and Cryptococcusneoformans, yeast-like fungi responsible ofsystemic and cutaneous mycoses in animals andhumans but the sensibility was lower whencompared to the ones shown by phytopathogenicfungi. On the contrary, in the assay condition noactivity on the growth ofC. parapsilosis (Ashford)orMalassezia pachydermatis was observed (LoCantore et al. 2006).

The results of this study confi

rmed previousfindings on the toxicity of tolaasin I on fungi(Rainey et al. 1991; 1992) and clearly demonstratedthat also WLIP, as other LPDs of bacterial origin(Bender et al. 1999), is toxic on fungi though ingeneral its activity is lower than that showed bytolaasin I. The above findings strongly suggestthat also WLIP, as already reported for tolaasin I(Brodey et al. 1991; Rainey et al. 1991; 1992;Hutchison and Johnstone, 1993), is important in

Table 2. Antimicrobial activity of tolaasin I and WLIP towards phytopathogenic fungi, chromista, yeast and fila-mentous fungi responsible for mycoses of mammalian, and cultivated mushrooms.

Micro-organisms No strains Minimal Inhibitory Quantity (g)

Tolaasin I WLIP

Fusarium solani 3 0.32 5.12F. oxysporum 1 0.32 2.56F. graminearum 1 0.16 -a

Sclerotinia sclerotiorum 2 0.32 1.28S. minor 1 0.16 0.64Sclerotium rolfsii 1 0.32 -a

Botrytis cinerea 2 0.16 1.28Rhizoctonia solani 1 0.08 0.64Verticillium dahliae 1 0.16 1.28Trichoderma viride 1 0.16 1.28Phytopthora nicotianae 1 0.32 -a

P. citrophthora 1 0.32 -a

Armillaria mellea 2 0.32 2.56Heterobasidion annosum 2 0.16 1.28Rhodotorula pilimanae 1 0.64 -a

Candida albicans 1 5.12 -b

C. parapsilosis 1 -a Malassezia pachydermatis 1 -a Cryptococcus neoformans 1 2.56

Agaricus bisporus 2 0.08 1.28Pleurotus eryngii 3 0.32 2.56P. ostreatus 2 0.16 1.28Lentinus edodes 2 0.32 1.28

ano inhibitory activity were observed with a quantity of tolaasin I or WLIP equal to 10.24 g.bno inhibitory activity were observed with a quantity of WLIP equal at 25 g.


8/32

88

Andolfi et al


the microbial interaction with mushrooms. In thisrespect, when considering the observed lowerantifungal activity of WLIP it is also necessaryto remind that strain NCPPB1311 ofP. reactans,at least in vitro, produced a quantity of WLIPwhich is more than ten fold higher thequantity of tolaasin I produced by the type strain

NCPPB2192 ofP. tolaasii. The lower antifungalactivity of WLIP appear to be compensatedby the higher quantity produced in culture.Furthermore, the observation that morphologicalvariants ofP. reactans strains are avirulentand do not produce WLIP (Lo Cantore, 2001;Iacobellis and Lo Cantore, 2003) furthersupport the possibility that the latter compoundmay play an important role in the interaction ofP. reactans with the cultivated mushrooms(Lo Cantore et al. 2006).

Assay on tissue blocks and wholesporophores ofAgaricus bisporusAssays of the two LPDs on tissue blocks ofA. bisporus showed that the deposition on theirsurface of a drop of solution containing 0.08 gof tolaasin I caused brown sunken lesions. In thesame assay condition a similar effect was observedwhen drops of solution containing 1.28 g ofWLIP were used (Fig. 6). Similar results were

obtained when such LPDs have been deposited onthe surface of whole sporophores caps (Lo Cantoreet al. 2006).

Haemolytic activity and osmoticprotection

Agarose haemolysis assay showed that both WLIPand tolaasin I cause the red blood cells (RBCs)lysis (Table 3). however, the minimal haemolyticquantity of tolaasin I, in the assay conditions, wasmore than four fold higher than that of WLIP and,moreover than, as previously reported (Raineyet al. 1991), it was temperature dependent. How-ever, this effect was not observed in the case ofWLIP. The higher temperatures might facilitatethe formation of oligomers of tolaasin I leadingto the transmembrane pore formation (Lo Cantoreet al. 2006).

Assays in water solution confirmed the higherhaemolytic activity of WLIP. In fact, the minimalhaemolytic concentration was 12.6 and 2.8 M fortolaasin I and WLIP, respectively (Lo Cantore et al.2006). In similar assays, Rainey et al. (1991)reported that 0.25 M solution of tolaasin I deter-mined an haemolytic effect. However, the latterexperiments were performed on horse erythrocytesat 37 C and also with the purpose to measure thelysis rate.

In the solution assay, as shown in Figure 7,both tolaasin I and WLIP caused human RBCslysis. WLIP was more effective than tolaasin Iwith regard to both C50 (i.e. the minimalconcentration able to cause 50% haemolysis,which was close to 34.6 and 4.3 g/ml for tolaa-sin I and WLIP, respectively) and of the hae-molysis rate (Fig. 7, inset). The Hill coefficientassociated to the formation of the active unit,appeared to be higher for WLIP (8 2) than fortolaasin I (6 1).

The haemolytic activity of both LDPs can beprevented by adding osmoticants of adequate

dimensions to the external medium. The resultsreported in Figure 8 suggest for WLIP and tolaasinI, at a concentration corresponding to 1.5 C50, apore radius between 1.5 0.1 and 1.7 0.1, and2.1 0.1 nm, respectively.

The small molecular size of the above LDPssuggests that such large pores could be formedby an aggregate of monomers, which is alsoconsistent with the large Hill coefficientsobserved. The structure of the pore formed by

A B

Figure 6. Brown lesions on tissue blocks ofAgaricus bisporus (lowerthree blocks in each treatment), caused by deposition of 5 l solutionscontaining 5.12 g of WLIP (A) and 0.64 g of tolaasin I (B),respectively. On upper blocks, 5 l of sterile water was deposited.


9/32

89



WLIP appeared to be less stable with a morepronounced dependence on toxin concentration.In fact, at higher toxin dose (2 C50) a biggerosmoticant is necessary for fully protect of RBCsfrom lysis (Fig. 8, right bottom panel) (Caraiola

et al. 2006).Theoretical analysis of the protection

experiments with the Renkin equation permits anestimation of the functional pore radius which was1.1 0.1 and 2.0 0.4 nm, for WLIP and tolaasinI, respectively. These radius values are similar tothose found for other LDPs (Dalla Serra et al.

1999a). The previous smaller estimation of thefunctional pore radius of tolaasin I, reported byRainey et al. (1991) to be 0.9 1.0 nm, coulddepend on the lower tolaasin I concentration usedin the experiments by the authors.

The Solution Structure of Tolaasin I

Secondary structure of tolaasin INMR spectra of tolaasin I were obtained indifferent solvents (1H2O, C

2H3O1H, DMSO-d6,

and aqueous SDS). Assignment of proton spinsystems in all solvents used was obtained withthe sequential methodology (Wthrich, 1986).Tolaasin I secondary structure was delineatedfrom the qualitative pattern recognition approachof the sequential and medium-range NOEs asderived from 2D experiments. In 1H2O andC2H3O

1H, NOESY spectra essentially displayedsequential effects indicating that the conformationof the whole peptide was largely extended. InDMSO-d6, proton chemical shifts were found tocorrespond to those previously reported (Nutkinset al. 1991). NOESY spectra with 0.10 and 0.25second mixing times showed some short- andmedium-range NOEs in the region DSer3-DLeu7in correspondence of tight turns as well as helical

structure (Wagner et al. 1986). Most likely it wasdetected what is referred to as nascent helix(Dyson et al. 1988). The rest of the peptide provedto be essentially extended. The solution structureof tolaasin I was then studied in a 1H2O/

2H2O(90/10 v/v) solution in presence of SDS at pH 7.0,a value at which the activity on erythrocytes isgreatest (Raney et al. 1991). The amide region ofthe NOESY spectrum at 0.10 s mixing time isshown in Figure 9.

Table 3. Minimal haemolytic quantity (M.H.Q.) of tolaasin I or WLIP as determined in agarose plate assay atdifferent temperatures.

Lipodepsipeptide Average Diameter of the Haemolysis Area (mm)

quantity (g) Tolaasin I WLIP

25 C 37 C 25 C 37 C

10.24 8 16 24 24

5.12 7 11 22 222.56 7 7 18 181.28 0 0 11 110.64 0 0 9 90.32 0 0 0 0

10-1 100 101 102

0

20

40

60

80

100tolaasin WLIP

0

tolaasin

50

%h

aemolysis

[LDP] (g/ml)

0 2 4

110

WLIP

Figure 7. Dose dependence of the haemolytic activity of tolaasin I(circles) and WLIP (squares) after 4 hours at room temperature.Dotted lines are best fit experimental data points with Hill equation.Hill coefficients are 6.3 and 8.0 for tolaasin I and WLIP, respectively.Inset: Set of traces recorded on the microplate reader of kinetics ofhaemolysis of RBC exposed to 2 step sequential dilutions of tolaasinI (left) and WLIP (right), from top to bottom. Starting concentration(top panels) were 100 g/ml for tolaasin I and 6.25 g/ml for WLIP.The decrease of turbidity over time indicated the disappearance ofintact RBC. All panels have the same linear axis, ranging from 0 to4 hours and from 0 to 110 mOD for x- and y-axes, respectively.


10/32

90

Andolfi et al


Sequential strong, non ambiguous cross-peaksbetween all consecutive amino acids, with the onlyexception of the DGln10-LLeu11, LIle15-LHse16,LHse16-DDab17 cross-peaks due to peak overlapwere observed. Furthermore, several cross-peaks,namely DLeu4-DSer6, DVal8-DGln10, LLeu11-But13, DVal12-DalloThr14 and LHse16-LLys18,were clearly observed. The presence of strongsequential NOEs together with medium in theregion DPro2-LVal9 is a first indication of a helicalstructure. This NOE pattern, generally observedfor right-handed helical structures in peptides andproteins with L-aminoacids, suggests here the pres-ence of a left-handed helix.

Observation of several medium-range effectssuggests the presence of an ordered structure.However, it is important to underline that the stan-dard distances observed for helices, -sheets andturns in peptides and proteins containing only D- or

L-residues cannot be used here to infer simply thefolding of tolaasin in the region containing alter-nate D- and L- amino acids. Not withstanding, thepresence of key NOE effects between non-sequential residues of the same chirality hints at aturn like folding. In order to clarify the issue, theNOESY cross-peak intensities expected for a

regular left-handed helix covering the 214 region,including the DPro2-DVal8 and the alternatingL- and D-amino acids were calculated (Jourdanet al. 2003).

Three-dimensional structureof tolaasin IIn SDS micelles, tolaasin I forms an amphipathicleft-handed -helix in the region DPro2-DThr14.It comprises a sequence of seven D-amino acids(DPro2-DSer-DLeu-DVal5-DSer-DLeu-DVal8) and

a stretch ofL-D-L-D-D-amino acid residues (LVal9-DGln-LLeu-DVal-But-DalloThr14) (Jourdan et al.2003). A stereoview of the best fit superposition ofthe backbone atoms of the 20 rSA/rEM structuresis shown in Figure 10A.

This is the first recognized example of a natu-rally occurring molecule with a regular left-handed-helix including both D- and L-amino acids, andit could be relevant for the de novo design of pro-tein structure. As a comparison, the closely related

0.8 1.2 1.6 2.0

10-3

10-2

10-1

100

0 1 2 3 4 5 0 1 2 3 4 5

40

60

80

r1=1.5 0.1 nm

r2=1.7 0.1n

r = 2.1 0.1 nm

radius (nm)0.8 1.2 1.6 2.0

radius (nm)

relativepermeability

P900

P4000

P2000

P1500

turbidity(mOD)

time (h)

buf

P1500

P1000

P900P600

buf

time (h)

tolaasin WLIP

Figure 8. Haemolysis in the presence of osmotic protectants andRenkin fit. Upper panels: Representative traces of decrease in

turbidity of a RBC suspension in the presence of different externalPEG osmolites and constant LDP concentrations (i.e. 100 g/ml and6.25 g/ml for tolaasin I and WLIP, respectively). PEG size is indicatednext to each trace. Lower panels: Renkin representation of datacollected in experiments similar to those reported in the upper panels.Curves through points are best fits according to Renkin equation [seetext; Schultz and Solomon (1961); Ginsburg and Stein (1987)], whichgives a pore radius for tolaasin I of 2.1 0.1 nm. This value does notdependent on LDP concentration, at least at the concentrations usedin this study (i.e. 50100 g/ml): dotted is the best fit of data collectedat [tolaasin I] = 76 g/ml (triangles, 2 C50), solid line refers to bestfit of data collected at [tolaasin I] = 50 (squares, 1.5 C50) and100 g/ml (circles, 3 C50). On the other hand, WLIP shows differentvalues of pore radius, ranging from 1.5 to 1.7 nm depending on toxinconcentration. Dotted line is the best fit of data collected at[WLIP] = 6.25 g/ml (circles, 1.5 C50), solid line refers to best fit ofdata collected at [WLIP] = 9 g/ml (squares, 2 C50).

Figure 9. The amide region of NOESY spectrum of tolaasin I in1H2O/

2H2O (90/10 v/v) in the presence of SDS with a mixing time of0.10 s, 300 K, pH 7.0. Where space permits, the cross-peaks arelabeled with the sequence number of amino acid residues.


11/32

91



lipodepsipeptide syringopeptin 25-A also forms aleft-handed -helix, but only includes residues ofD-chirality (Ballio et al. 1995). A percentage of20% left-handed-helix has been calculated fromCD data, and it was hypothesized that the helixcould be located in the N-terminal sequence of 7D-amino acids (Mortishire-Smith et al. 1991a).However, the CD data do not allow to specificallyposition the helix in the sequence as 20% of thesequence represents only 3.6 residues, i.e. a singlehelix turn.

The C-terminal lactone loop adopts a boat-like conformation and is found shifted from the

helical axis, therefore tolaasin I assumes agolf-club conformation (Fig. 10B) (Jourdan et al.2003). This golf club-like overall conformationseems likely to anchor in plasma membrane bymeans of its amphipatic helical segment with thehydrophobic lactone loop lying close to the surface.Further studies remain highly desirable to under-

stand how such a structure can create ion channelsin plasma membranes. In particular, the formationof oligomeric structures, already hypothesized forsyringomycin and syringopeptins, (Feigin et al.1997; Hutchison and Johnstone, 1993; Dalla Serraet al. 1999a; 1999b) is now under investigation fortolaasin I.

It has been reported that the cationic heptapeptidering in polymixyns (Storm et al. 1977), ranalexin(Clark et al. 1994), and in type-I brevinins(Simmaco et al. 1994) is responsible for thedisruption of bacterial membrane permeability

and thus for antibiotic activity. The octapeptidelactone ring of syringopeptins shows twopositively charged residues as in the abovemolecules (Ballio et al. 1995). Also in thepentapeptidic C-terminal lactone ring in tolaasinI two positively charged groups (dDab17 andlLys18) are present. These findings imply that,independently from its size, the presence ofpositive charges into the loop together with theamphipathic helix, represent a fundamentalmembrane-permeabilizing motif (Oren and Shai,1998; Andreu and Rivas, 1998).

Structure-activity relationshipsof tolaasin IIt has been reported that tolaasin I is able to causehorse erythrocytes haemolysis as consequence oftransmembrane pores formation in erythrocyte(Raney et al. 1991). Because of its small size thiseffect may be the result of the monomer aggregationacross the membrane. Zn2+ inactivates ion channelformation probably via chelation of ionizable

groups in the membrane near the site of the poreformation (Brodey et al. 1991). Tolaasin Isecondary structure suggests that the amphipathicN-terminal -helix is deeply anchored into themembrane, leaving the hydrophilic lactone ringclose to the surface of membrane. A longeranchored tail might both result in a higher affinityto the membrane and a stronger disrupting effect.Moreover, the length of the N-terminal sequencemight play a central role in the formation of stable

Figure 10. Calculated structures of tolaasin I (A) MOLMOL stereoviewof the best-fit superposition of the 20 lowest energy rSA/rEMconfor-mations of tolaasin. Backbone atoms of residues 214 were usedfor the best fit. (B) Stickgolf-club like tolaasin I conformation. Dottedlines indicate hydrogen bondsalong the left-handed -helix. Thehelical axis inside the helix is also represented. The shaded areasimulates the ellipsoid that covers most of the backbone atoms ofthe lactone ring, and its largest axisis also shown.


12/32

92

Andolfi et al


oligomeric structures required for ion channelformation. This may explain the loss of activityin ion channels formation observed in the case oftolaasin 144, an altered tolaasin I derivativelacking DSer6-DLeu7-DVal8. Preliminary MDsimulations indicate that the N-terminal sequenceof tolaasin 144 is more flexible than tolaasin I and

it does not fold into a stable helix. These featuremay prevent the oligomeric aggregate formationand may explain the fact that tolaasin I lyseserythrocytes by two mechanisms (Hutchison andJohnstone, 1993). At low tolaasin I concentrationthe activity is the result of Zn2+-sensitive ionchannel formation whereas at high concentrationsit is the result of Zn2+-insensitive detergent actionof both molecules. It has been reported that the-helical structure, while important for cytotoxicitytoward mammalian cells, is not a prerequisite forantibacterial activity (Shai and Oren, 1996). Novel

antibiotic peptides useful as therapeutic drugsrequire strong antibiotic activity against bacterialand fungal cells without haemolytic effect (Raneyet al. 1991). Functional and structural studies withD-amino acids incorporated analogues of pardaxinreveal that the lack of a significant -helix doesnot prevent the lysis of bacterial membrane. Thepeptide is probably acting as a detergent in whathas been described as a carpet-like mechanism(Shai, 1995; Shai and Oren, 1996). Therefore, thedecrease of the helical content into lytyc peptidesby incorporating D-amino acids may be a useful

approach to discover new peptide antibiotics fortherapeutic use. In this context, tolaasin I andtolaasin 144 are very interesting molecules: thepresence of a stable-helix in the former warrantsits lytyc activity, while the decreased helicalcontent together with the detergent action exertedby tolaasin 144 strongly suggest its possible roleas antibacterial, but not cytotoxic, peptide.Furthermore, the above substances present L-amino acids into a D-amino acid sequence: thisinduces resistance to enzymatic degradation andinactivation (Wade et al. 1990), and increaseantigenicity (Benkirane et al. 1993). Theantimicrobial assays of tolaasin I, though limitedto a representative set of test micro-organisms,confirmed the high sensitivity of Gram-positivebacteria, as previously reported (Raney et al.1991), and showed a quite good activity towardthe yeast Rhodotorula pilimanae and the Gramnegative bacterium E. coli (Lo Cantore et al.2003a; 2006). The activity againstE. coli (strain

ITM100), in contrast with the results previouslyreported (Raney et al. 1991), indicates a variabilityof sensitivity among strains, and strongly suggeststhat before conclusions are drawn several strainsof the same micro-organisms must be checked. Itwould be of interest to test tolaasin I and tolaasin144 against Gram-positive and Gram-negative

bacteria of medical interest. Finally, the structuralinformation on tolaasin I can be extremely helpfulfor its possible use in the control of the brownblotch disease ofA. bisporus also in combinationwith other methods. Treatment of mushrooms capswith tolaasin I provokes pitting and browning(Raney et al. 1991) as well as activation of thedefense mechanism of the mushroom (Soler-Rivaset al. 1999a). Nutkins et al. (1991) reported thatpa rt ially pu ri fied toxin reproduces di seasesymptoms, as does the intact organism, althoughtolaasin I induces more prominent symptoms than

living bacteria (Moquet et al. 1996). Accordingly,tolaasin I appears to be an adequate target to stopthe sequence of events affectingA. bisporus. Thefact that WLIP appears to be able to inhibit thebrowning could confirm this hypothesis (Jourdanet al. 2003).

Isolation and Chemicaland Biological Chracterizationof News TolaasinsTolaasins A-E (1115, Fig. 11), were isolated

together with tolaasins I and II from the cell-freeculture filtrate of the type strain NCPPB2192 ofP. tolaasii essentially by semi-preparative HPLCas above cited (Fig. 4).

A preliminary analysis of1H NMR and ESIMSdata of these metabolites indicated that the newlipodepsipetides are structurally related to tolaa-sins I(8)andII(10) as previously suggested byShirata et al. (1995). The authors isolated eighttolaasins, including tolaasins I and II fromP. tolaasii cultures filtrates and hypothesized, onlyon the basis of MS data, that their structural fea-tures were similar to 8 and 10. In our case seventolaasins, including tolaasins I and II, were iso-lated and the structures of two of them differ fromthose hypothesized from Shirata et al. (1995). Acomplete NMR-based structural elucidation oftolaasins A-E (1115) was not an easy task toaccomplish, due to the low amount of substancesavailable. For this reason, while a full character-ization of the 1H-NMR spin systems was achieved


13/32

93



through extensive use of 2D NMR techniques,13C NMR carbon chemical shifts could not beassigned. Attempts to aquire 13C NMR data bymeans of suitable 2D NMR experiments (HSQCand HMBC) at very long-acquisition time failed.The assignment of the proton spin systems of eachamino acid residue, including the identification ofthe N-terminal side chain, was mainly obtainedfrom the 2D TOCSY and the 2Q (Double quantumspectroscopy) experiments (Wthrich, 1986; p292). The 2Q spectrum was acquired in order toassess the sequence specific assignment of the spinsystems preliminarily characterized in the TOCSYspectrum, together with the pattern analysis ofsequential dipolar couplings displayed in the 2DNOESY spectrum (Wthrich, 1986; p-292). Fullproton assignments for tolaasins A-E (1115), in

DMSO-d6 at 300 K were obtained (Bassarelloet al. 2004).

Tolaasin A (11) showed a pseudomolecular ionpeak in the HRESIMS spectrum at m/z980.0912(M+2H)2+, consistent with the molecular formulaC91 H155N21O26 (exact mass calculated forC91H157N21O26/2 980.080435). HRESIMS data wereobtained also for the intense fragment ion (M-chain-But1+H) and interpreted as originating from theelimination of [HOOC-(CH2)3CONHC=CH(CH2)-C=O] with hydrogen transfer to the chargedspecies (see below). Extensive analysis of the1H NMR data of tolaasin A (11), including TOCSY,DQF COSY, and 2Q spectra, permitted confirmationof the same amino acid sequence present inthe parent tolaasin I (8). Moreover, the Thr14residue and the C-terminal Lys18 were found to

HNNH

HN

NH

HN

NH

HN

NH

HN

O

O

O

O

O

O

O

ONH

HN

O

ONH

O HN

NHO

OHN

H

O

NH

OO NH

2

O

O

NNH O

OOH

HO

OH

NH2

NH2

O

10 Tolaasin II

H

O

HNNH

HN

NH

HN

NH

HN

NH

HN

O

O

O

O

O

O

O

O

NH

HN

O

O

NH

OHN

NH

O

O

R2

HNO

NH

OO

NH2

O

H2N

O

NR

1HN

O

HO

OH

NH2

HNNH

HN

NH

HN

NH

HN

NH

HN

O

NH2

O

O

O

HO

OH

O

O

O

O

O

NH

HN

O

O

NH

OHN

NH

O

O

O

NNH

O

OOH

HN

O

OH

OH

NH

COOH

NH2

O

H2N

11 Tolaasin A: R1=HOOC(CH

2)

3CO; R

2=CH(CH

3)CH

2CH

3; R

3=CH

2CH

2OH

12Tolaasin B: R1=CH

3(CH

2)

4CH(OH)CH

2CO; R

2=CH(CH

3)

2; R

3=CH

2CH

2OH

13Tolaasin D: R1=CH

3(CH

2)

4CH(OH)CH

2CO; R

2=CH

2CH(CH

3)

2; R

3=CH

2CH

2OH

14Tolaaasin E (7): R1=CH

3(CH

2)

4CH(OH)CH

2CO; R

2=CH

2CH(CH

3)

2; R

3=H

R3

15Tolaasin C

Figure 11. Structures of tolaasins II and A-E (1015).


14/32

94

Andolfi et al


be involved in the lactone formation, as alreadyobserved for tolaasin I. On the other hand, a dif-ferent nature of the N-terminus chain was sug-gested by the analysis of the DQF-COSY spectrum.The N-terminus blocking group was unequivocallyidentified by the fragmentation peaks observed inthe ESIMS spectrum and appeared to have a

HOOC-(CH2)3-CO- structure. Finally, the ESIMS/MS (collision induced dissociation) spectrum,confirmed the same amino acid sequence for thetwo depsipeptides.

Tolaasin B (12) showed a pseudomolecular ionpeak at m/z1973 (M+H)+, 14 Daltons less thantolaasin I, suggesting that tolaasin B could differfrom the analog tolaasin I by the absence of amethylene group. NMR and the ESIMS data per-mitted to deduce for tolaasin B a molecular for-mula C93H161N21O25. Moreover, several peptidefragmentation peaks generated by a stepwise

spontaneous loss of amino acid residues from theN-terminus were observed. The TOCSY spectrum,almost superimposible to that of tolaasin I, permit-ted the identification of tolaasin B (12), whichshows the substitution of the isoleucine spin sys-tem present in tolaasin I (8) with a proton con-nectivity consistent with a valine residue. Twointense NOE correlations were observed betweenHse16 NH and Val H, and Val NH and Thr14H, confirming that the valine residue is at posi-tion 15. The presence of the intact lactone moietywas confirmed by CID MS/MS analysis. Taken

together the above data allowed to establish thestructure 12 for tolaasin B.

Tolaasin C (15) showed a pseudomolecular ionpeak at m/z 2005 (M+H)+ corresponding to amolecular weight of 2003 Da, 18 mass units greaterthan that of tolaasin I, suggesting that tolaasin Cmay be the acyclic version of tolaasin I derived bythe hydrolysis of its lactone ring (molecular for-mula C94H165N21O26). Comparison of 2D NMRspectra and the ESIMS fragmentation patterns oftolaasins C (15) and I (8) further supported thehypothesis that tolaasin C is a linear peptide gener-ated from the hydrolysis of the lactone ring. In fact,the proton chemical shifts of the dehydroamino-butyric acid residue at position-13 (But13) andthe threonine at position-14 suggested that thehydroxyl group of the side chain of this residue isnot involved in an ester functionality.

Tolaasin D (13) exihibited a pseudomolecularion peak at m/z1987 (M+H)+, suggesting the samemolecular formula C94H163N21O25 already observed

for tolaasin I. Thus tolaasin D (13) and I (8) areisomers. A careful comparison of the TOCSYspectra of tolaasin D and I indicated that tolaasinD contained a leucine residue in place of the iso-leucine present in the related lipodepsipeptide 8.The 2Q spectrum allowed to specifically assign theresonances of the leucine methylene protons. The

position 15 of the leucine was supported by twoNOE effects between Hse16 NH and Leu15 H,and between Leu15 NH and Thr14 H .

Tolaasin E (14) showed a pseudomolecular ionpeak at m/z 1943 (M+H)+ suggesting the samemolecular formula C92H159N21O24 of tolaasin II(10). Hence, tolaasin E (14) is isomeric with tolaa-sin II (10). The latter depsipeptide, contains aglycine residue in place of the homoserine residueof tolaasin I (8). In fact, in the TOCSY spectrumof tolaasin E it was possible to characterize thespin system of a glycine residue, confirming that

14 is indeed a tolaasin II analogue. The 2Q exper-iment indicated the presence of four rather thanthree leucine residues. A careful examination ofthe NOESY and TOCSY spectra confirmed thatthe first three leucine residues are at positions 4, 7and 11, as in tolaasin II (10). NOE correlationsallowed the placement of the fourth leucine atposition 15 in tolaasin II. A support to the struc-ture of tolaasin E was the presence of the -hydroxyoctanoil chain at the N-terminus asdeduced by ESIMS data through the spontaneousloss of 225 Da (the chain +But1) from the

pseudomolecular ion peak at m/z1943.The antimicrobial activity of HPLC grade tolaa-

sins A-E (1115)assayed in comparison withtolaasin I and II against the yeastR. pilimanae, thefungus Rizoctonia solani, the Gram-positivebacteriaB. megaterium andRodococcus fascians,respectively, and the Gram-negative bacteriaE. coli and Erwinia carotovora subsp. carotov-orais reported in Table 4.B. megaterium andR.fascians were the most sensible test micro-organ-isms. In fact, all the analogs, except tolaasin C,inhibited the growth of these bacteria though dif-ferences among their specific activities wereobserved. The most active analogs appear to betolaasin D followed by tolaasin I and II with aminimal inhibitory quantity of 0.16, 0.32 and 0.64g, respectively. Tolaasins A and B, and E wereless active, with a minimal inhibitory quantity of1.28 and 2.56 g, respectively. A similar sensitiv-ity to the above substances was shown by thefungusR. solani. None of the analogs inhibited the


15/32

95



growth of the Gram-negative bacteriaE. coli andE. carotovora subsp. carotovora at least at theconcentration used in the assays. Finally, thegrowth of the yeastR. pilimanae was inhibited onlyby tolaasins I, II, A and D though the sensitivity ofthe yeast was lower in comparison toB. megate-

rium (Table 4). The results of the antimicrobialactivity of the tolaasins suggest the importance inthis respect of the lactone and the N-terminus acylmoiety. Tolaasin A, which has the pentadioic acidinstead of the -hydroxyoctanoic acid, showed areduced activity in respect to tolaasin I while tolaa-sin C, the linear derivative of tolaasin I, due to thelactone opening, lacked the activity at least at theconcentration tested. The effect of the lactone ringopening on the lack of the antimicrobial activityhas been already reported for other bacteriallipodepsipeptides (Jourdan et al. 2003). Further-

more, the importance of the position 15 in thepeptide moiety of both tolaasins I and its deriva-tives in respect to the antimicrobial activity wasobserved. In fact, the replacement of isoleucine inposition 15 with a valine or leucine residue intolaasins B and D, respectively, determines thedecrease or the increase of the antimicrobial activ-ity when compared to tolaasin I. The presence ofleucine in position 15 in tolaasin E determined thereduction of the activity when compared to tolaa-sin II. Though the latter result is apparently incontrast with the effect of the same replacement inthe tolaasin D, it is important to consider thattolaasin II differ from tolaasin I for the replacementof the homoserine in position 16 with a glycineresidue (Bassarello et al. 2004).

As previously reported tolaasin I forms anamphipatic left-handed -helix in the regionDProd2-DalloThr14 comprising the sequence ofseven D-amino acids and the adjacent L-D-L-D-D-region. Furthermore, the lactone macrocycle

adopts a boat-like conformation and is shiftedfrom the helical axis giving rise to a golf-cluboverall conformation. These structural featuresappear of importance on toxicity of tolaasin I(Jourdan et al. 2003). Therefore, structural changes,occurring in position 15 in the case of tolaasins B

and D, in position 16 in tolaasin II, and in bothposition 15 and 16 in tolaasin E, may modify thislactone conformation and consequently theantimicrobial activity (Bassarello et al. 2004).

Mode of Action of Tolaasin Iand WLIP on Natural and ModelMembranesThe activity WLIP and tolaasin I was compara-tively evaluated on lipid membranes. Both LDPsinduced the release of calcein from large unilamel-

lar vesicles (LUVs). Their activity was dependenton the toxin concentration and liposome composi-tion and in particular it increased with the sphin-gomyelin (SM) content of the membrane. Studiesof dynamic light scattering suggested a detergent-like activity for WLIP at high concentration (27M). This effect was not detected for tolaasin I atthe concentrations tested (28 M). Differenceswere also observed in LDPs secondary structure.In particular, the conformation of the smaller WLIPchanged slightly when it passed from the buffersolution to the lipid environment. On the contrary,a valuable increment in the helical content oftolaasin I, which was inserted in the membranecore and oriented parallel to the lipid acyl chains,was observed. The novelty of the pathogeneticcapability ofP. reactans and, in particular, the factthat avirulent variants of the pathogen have alsolost the ability to produce WLIP (Lo Cantore andIacobellis, 2002) suggest that this metabolite maybe important in the P. reactans-mushrooms

Table 4. Antimicrobial activity of tolaasins I, II, and A-E.

Micro-organism Tolaasins Minimal Inhibitory Quantity (g)

I II A B C D E

Rizoctonia solani1583 0.32 0.64 1.28 2.56 5.12 0.16 5.12Rhodotorula pilimanae ATCC26423 2.56 5.12 5.12 5.12 5.12 2.56 5.12Bacillus megaterium ITM100 0.32 0.64 1.28 2.56 5.12 0.16 2.56Rodococcus fascians NCPPB3067 0.32 0.64 1.28 1.28 5.12 0.16 2.56

Escherichia coliK12 ITM103 5.12 5.12 5.12 5.12 5.12 5.12 5.12Erwinia carotovora subsp. 5.12 5.12 5.12 5.12 5.12 5.12 5.12carotovora ICMP5702


16/32

96

Andolfi et al


interactions. Although the haemolytic and theantimicrobial activity of WLIP and its interactionswith synthetic membranes have been recentlydemonstrated (Lo Cantore et al. 2003a, p 25562;Lo Cantore et al. 2003b, p 26373; Lo Cantoreet al. 2006), the mechanism involved is not yetdefined. The effects of WLIP on model membranes

of different lipid composition, also in comparisonto tolaasin I, were investigated analysing the LDPsmode of action and the molecularity of the per-meabilising unit. In addition, the changes in thesecondary structure occurring when these mole-cules passed from the aqueous buffer to the mem-brane were studied. In particular, it was possibleto analyse the conformation of WLIP and tolaasinI during their interaction with the lipid membrane,i.e. their natural target, and, in the case of tolaasin I,to calculate the molecule orientation with respectto the lipid acyl chains (Coraiola et al. 2006).

Calcein release assayBoth WLIP and tolaasin I are able to damage bio-logical and artificial membranes through the forma-tion of transmembrane pores (Brodey et al. 1991;Cho and Kim, 2003; Lo Cantore et al. 2003b, p26373; Lo Cantore et al. 2006).

Tolaasin I was more active on SM- andphospholipids containing LUVs and less activewhen sterols were present (Table 5 and Fig. 12).This feature, observed for other peptin-like LDPs(Dalla Serra et al. 1999b; Menestrina, 2003,

p 18598), was correlated to their cell specificitybeing peptin-like LDPs more phytotoxic and lessantifungal than the smaller LDPs such as thenonapeptides syringomycines, pseudomycines andcormycin. The latter showed a clear preference forsterols (Bender et al. 1999; Dalla Serra et al. 1999b;Menestrina 2003; Julmanop et al. 1993; Taguchi

et al. 1994; Feigin et al. 1997; Scaloni et al. 2004).In the case of tolaasin I, the slight preference forergosterol (a fungal sterol) among sterols isconsistent with its antifungal activity.

On the contrary to the expectations, also WLIP,which is a nonapeptide, was more active on SM-containing LUVs. This could be explained by thefact that this LDP shows structure similaritywith the peptine-like LDPs with only a portion ofthe peptide moiety involved in the formation ofthe lactone ring. In comparison to tolaasin I, theinhibitory effect of sterols was less effective on the

WLIP activity.Nevertheless the two toxins showed similar

responses to changes of the lipid membrane com-position.

Firstly, the molecularity of the permeabilisingunit, estimated from the Parente-Rapaport modeland reported in Table 5, was larger with WLIP andin the same range of detergents (Dalla Serra et al.1999a). In the case of tolaasin I, Cho and Kim(2003) described the existence of subconductancestates, which could be related to the presence ofpores with slightly different molecularities .

Table 5. Permeabilizing activity of WLIP and tolaasin I on vesicles of different lipid composition and a statisticalanalysis of their pore-formation mechanism.

Lipid Compositiona

(mol%) WLIP Tolaasin I

PC SM sterol 1/Cb50 (M-1) K1

c (M-1) Md K2e 1/Cb50 (M

-1) K1c (M-1) Md K2

e

100 0 0 0.59 0.04 50 20 8 1 533 451 2.8 0.4 233 52 6 1 5.2 2.650 50 0 1.41 0.05 100 30 6 1 200 340 4.9 0.2 673 179 6 1 1.7 1.150 33 16.5* 0.70 0.05 100 30 12 1 217 247 2.8 0.2 200 50 6 1 12 1250 16.5 33* 0.59 0.01 50 20 8 1 533 451 1.2 0.2 100 30 7 1 19.7 17.4

50 0 50* 0.45 0.01 50 20 11 1 533 451 0.4 0.1 50 20 7 1 0.8 0.350 33 16.5 0.70 0.03 100 30 12 1 413 440 2.5 0.2 250 54 7 1 2.8 1.750 16.5 33 0.59 0.08 50 20 8 1 533 451 1.3 0.1 100 30 7 1 30 2050 0 50 0.54 0.01 50 20 10 1 533 451 1.0 0.2 50 20 5 1 30 20aLipid mixtures are reported on a molar base. PC: egg phosphatidylcholine; SM: sphyngomyelin; the sterol included in the vesicles wascholesterol* and ergosterol.

bThe permeabilising activity is expressed as the reciprocal of the concentration of toxin causing 50% of calcein release (C50). Data areaverage S.D. of at least two different experiments.

cThe partition coefficient of toxin monomers into the liposome.dThe number of monomers necessary for the formation of an active pore.eThe aggregation process of membrane-inserted monomers.c; d ; eData are average S.D. of at least three different experiments.


17/32

97



This instability of the pore structure well correlateswith the variability of the parameter M found byour analysis (Coraiola et al. 2006).

For both of LDPs the activity seemed mainlydue to the ability of monomers to bind to theliposome membrane (K1) rather than to the two-dimensional aggregation rate (K2) of the boundpeptide (Caraiola et al. 2006).

Dynamic light scatteringFrom the results reported in Figure 13 it wasevident that WLIP solution at concentration higherthan 27 M showed a detergent-like activity; infact, the intensity of the scattered light loweredprobably in consequence of vesicles micellization.On the contrary, both LDPs at low concentration(at a LDP to lipid molar ratio up to 0.3) caused anincrease of the liposome dimensions and a decreaseof the intensity of the diffused light. These effects

well correlated with vesicles aggregation, thatincreased the measured diameters, and with pre-cipitation of vesicle macro-aggregates whichcaused a decrease of signal intensity through adecrease of the concentration of scattering cen-tres.

Interestingly, the concentrations of WLIP and

tolaasin I used in these assays were of the same orderof magnitude required to induce mushroom tissuealteration in vivo (Lo Cantore et al. 2006; Brodeyet al. 1991; Hutchison and Johnstone, 1993). Alsothe different activity between the two LDPs wasobserved, being WLIP about ten times less activethan tolaasin I (Lo Cantore et al. 2003a; 2006).

Fourier-transformed infraredspectroscopy experimentsThe results reported in above paragraphs together

with those in Lo Cantore et al. (2006), supportedthe idea of the cell membrane as the main target

10-3 10-2 10-1 1000

20

40

60

80

100

10-2 10-1 100

R%

T/L

[Tolaasin I] (M)

Figure 12. Dose dependence activity of tolaasin I on liposomes withdifferent lipid composition: Calcein loaded liposomes were exposedto different peptide concentrations. The percentage of calcein releasewas determined after 45 min as a function of the toxin/lipid ratio (T/L)and expressed as % of the maximal value obtained with TritonX-100.Experiments were done at constant lipid concentration (about 9 M).The true toxin concentration used was reported in the upper scaleof the panel. Lines through the points are best fit according to thestatistical model described in the text. Best values offitting parametersare reported in Table 1. Vesicles were prepared with the followingcompositions expressed in molar ratio: (down triangle) PC:SM(50:50), (up triangle) PC:SM:Chol (50:33:16.5), (circle) PC:SM:Chol(50:16.5:33), (square) PC:Chol (50:50).

100 101

-90

-60

-30

10-1 100 101

0

5

10

15

20

25tol IWLIP

ILS%

T (M)

10-3 10-2 10-1T/L

d

iameter%

T (M)

10-3 10-2 10-1T/L

Figure 13. Effect of toxins on vesicle size as determined by dynamiclight scattering: The treatment of phosphatidylcholine vesicle withtolaasin I (open square) and WLIP (close circle) caused changes invesicle average size ( diameter) which was reported vs the toxinconcentration (T), normalized to the vesicle diameter in the absenceof toxin. Experiments were obtained at constant lipid concentration(80 M). The toxin/lipid ratio (T/L) was reported in the upper scale ofthe same panel. Inset. In this case we analyzed the variation of thescattered light intensity ( ILS), normalized to the intensity value inthe absence of toxin. Other parameters are as above.


18/32

98

Andolfi et al


for both WLIP and tolaasin I. Therefore it seemedof interest to investigate the conformationalchanges occurring when these molecules interactwith the lipid environment. The secondary struc-ture of both LDPs was already reported but onlyin the soluble form (Mortishire-Smith et al. 1991b;Jourdan et al. 2003). In particular, the conformation

of tolaasin I was studied in SDS (Mortishire-Smithet al. 1991a, 1991b; Jourdan et al. 2003): in thismembrane-like environment it was demonstratedthe presence of a left-handed -helix, suggestinga refolding of the structure from the aqueous buf-fer where the conformation of the whole peptidewas largely extended (Jourdan et al. 2003). Con-sidering these results, FTIR spectroscopy was usedto determine the LPDs conformation and to followits changes when the molecules passed from hydro-philic to a membrane mimetic environment. Forthis reason, FTIR spectra were collected in buffer

solution, in an organic solvent (TFE, HFIP), inSDS or in the presence of lipid vescicles. Thestructure of peptides was analyzed in the lipid-mimetic conditions provided by TFE, HFIP andSDS, as well as in the model lipid membrane ofpalmitolyl-oleoyl-phosphatidylcholine (POPC)liposome. Changes in the secondary structure ofpeptides passing from the water buffer solution tothe more hydrophobic environment with a maxi-mum effect in the presence of the lipid membranewere expected (Table 6).

The rigidity of WLIP, cyclized via lactone

formation between the third and the C-terminalresidues (Mortishire-Smith et al. 1991b), couldexplain the slight variation observed in the amide Ispectrum even after the interaction with the membrane.Moreover, the presence in the same molecule ofunusual residues made the assignment of a secondarystructure to the curve fitting components more

difficult. Therefore the differential spectra obtainedafter H/D exchange or after the buffer substitutionwith a membrane-like environment were focused(Fig. 14). In fact, variations of spectra after H/Dexchange provided information about the localstructure of the molecule. The decrease of absorptionat 1670 cm1 (and the corresponding increment at

1634 cm1

), for example, indicated the presence offree protons (i.e. not involved in hydrogen bonds)which were able to exchange with deuterium, probablybecause of their good accessibility allowed by the localstructure of the molecule. This behaviour could beattributed to a turn configuration with fast exchangingprotons. Interestingly, this behaviour was dependenton the solvent used and decreased in a lipid-mimeticenvironment (effect measured by the small negativeand positive peaks at 1670 cm1 and 1634 cm1,respectively), suggesting a decrease of the protonaccessibility and, consequently, a change in the local

structure of the molecule (Fig. 15A). Therefore, itseemed that the lipid-mimetic conditions enhancedthe formation of hydrogen bonds inside the structurewhich become less susceptible to deuteration. Similarchanges were confirmed in differential spectra

Table 6. FTIR spectroscopy determination of the sec-ondary structure of tolaasin I in aqueous buffer, withand without lipids, and in different solvents.

% Secondary structurea

System h r

buffer 29 44 27TFE 24 45 30HFIP 24 48 28SDS 27 47 26POPC 21 63 15

a: total -structure (i.e. -sheet, -turn); h: helix; r: random. Errorsof FTIR determinations are 5%.

1700 1650 1600

-0.01

0.00

0.01

0.02

D2O - H

2O

L

H

T

B

S

Wavenumber (cm-1)

A(a.u.)

B T H S L0.0

0.1

0.2

0.3

Pea

karea

Figure 14. Differential spectra after H/D exchange of the WLIP amideprotons: Analysis of differences in the amide I band offilms of WLIPsamples deposited from a buffer solution (B, dashed line), TFE (T,dotted line), HFIP (H, solid line), SDS (S, thick dashed line) and afterbinding to the lipid membrane (L, thick dotted-dashed line). Differen-tial spectra were obtained by subtracting the hydrogenated spectraof WLIP in the different environments from the corresponding deuter-ated spectra. (Inset) Area of negative peaks between 1710 and 1660cm1. All the differences are between normalized spectra, i.e. withthe amide I peak area set at 1.


19/32

99



obtained by subtracting the deuterated spectrum ofthe soluble form in buffer from those obtained withother solvents (Fig. 15B). Variations in the absorptionat the above wavenumbers showed an increment at1670 cm1 and a corresponding decrease at 1634 cm1with the maximum in the lipid membrane. Thesechanges favoured a more stable local structure whichwas enhanced from the lipid-mimetic environmentsand that could be able to interact with the membraneand its hydrophobic moiety. In addition, the disorderingeffect on the vesicles membrane induced by the WLIPbinding supports the hypothesis of an insertion of themolecule into the membrane lipid core.

As expected an increment of the helical contentin the structure of tolaasin I was observed when itpassed from the water buffer solution to themembrane-like environment. This is consistent

with results obtained by CD and NMR spectroscopyin aqueous solution and in SDS, respectively(Mortishire-Smith et al. 1991a; Jourdan et al.2003). Nevertheless some considerations need tobe added. The peptide moiety of tolaasin I (18aminoacids) includes unusual residues withD-chirality, so the standard assignment of second-

ary structures to the curve fit components could benot always applied. Interesting recent studies wererecently reported on -helical membrane lyticpeptides, in which a few L-amino acids werereplaced with their D-enantiomers: incrementedamide I frequencies (16561670 cm1) comparedwith pure -helical structures (16451654 cm1)were observed and assigned to 310-helix ordynamic/distorted -helix (Papo and Shai, 2004).Previous studies had correlated this region with310-helix (Kennedy et al. 1991), -turn (Susi andByler, 1986) and interestingly to a left-handed

-helix, as well (Giancotti et al. 1973; Citra andAlxelsen, 1996). In this analysis two componentsaround the typical region of helix absorption, centredat 1654 cm1 and 1664 cm1, respectively (data notshown) were found (Fig. 16A and Table 7). Afterdeuteration the position of both peaks did notchange, suggesting a stable structure maintainedby hydrogen bonds, like a helix. Moreover, the sumof percentage absorption of these two componentskept at a constant value, independently of theenvironment (i.e. buffer, solvents and membranes),on the contrary the single absorptions changed: the

peak at 1654 cm1 increased passing from bufferto solvents at expenses of peaks at 1664 cm1. Afterinteraction with the membrane, the contribution ofthe former was maximum but analysis of thedichroic spectrum showed only one peak at higherwavenumber (1659 cm1) (Fig. 16B). For all thesereasons, the two contributions were joined in anunique peak and assignto it a helix structure whichfinally resulted the conformation adopted fromtolaasin I to bind the membrane and to insert in it.The significant decrease in the orientation orderparameter of the aliphatic chain region (Table 6)and the orientation of tolaasin I in the lipid bilayerprompt to interpret these results in favour of abarrel-stave mechanism of action versus a carpetlike hypothesis (Malev et al. 2002). This is indeedin agreement with Bassarello et al. (2004) whodescribed for tolaasin I the formation of a stableamphipathic helical structure and suggested itsinsertion into the lipid matrix for the formation ofan ionic channel.

0.00

0.01

0.02

0.03

1700 1650 1600

-0.01

0.00

0.01

Amide I Abuffer

POPC

Absorb

ance(a.u.)

Wavenumber (cm-1)

B

A(a.u.)

1

43

2

Figure 15. FTIR-ATR spectra of WLIP in buffer and in lipid mimeticenvironments: (A) Analysis of the amide I band of deuterated filmsof WLIP samples deposited from a buffer solution (buffer, thick dottedline) or after binding to the lipid membrane (POPC, solid line). Theoriginal spectrum (POPC) was deconvoluted and curve fitted toresolve the component frequencies. The corresponding Gaussianbands are reported as dotted lines, and their sum as a thick dashedline superimposed to the original spectrum. (B) Differential spectrawere obtained by subtracting the deuterated spectrum of the solubleform in buffer from that in TFE (1), in HFIP (3), in SDS (2) or in pres-ence of the lipid phase (4).


20/32

100

Andolfi et al


On the contrary, a carpet-like mechanism waspreviously proposed for magainin derived lipopep-tides which localized on the membrane surface anddid not significantly destabilise the acyl-chain order.In fact, Avrahami and Shai (2002) observed nega-tive order parmeters for both natural magainin andlipo-magainin (magainin analogues conjugated

with lipophilic acids), typical of helices orientednearly parallel to the membrane surface. In addition,after incorporation of the peptides into the lipidmembrane, the lipid order was only poorly dis-turbed and to the same extent, suggesting that bothpeptides did not insert deeply into the lipid core.

A different mechanism of action could beproposed for WLIP, in agreement wi th theevidences here reported for membrane disruptionin a detergent-like manner. In fact, it does not havea sufficient length to transverse the entire membrane,similarly to syringomicin for which Malev et al.

(2002) hypothesized a toroidal pore.In conclusion, the results described shed light

on the permabilising effects induced by WLIP andtolaasin I on liposome and on the secondary struc-ture of these molecules during their interactionwith the lipid membrane, which is their biologicaltarget (Caraiola et al. 2006).

Exo- and Lipo-Polysaccharidefrom Pseudomonas tolaasii,P. reactans and Burkholderia

gladiolipv. agaricicolaIn the last decades the chemical composition andstructures of exopolysaccharides (EPSs), complexmolecules of the bacterial capsules, of several plant

0.00

0.01

0.02

0.03

0.04

1700 1650 1600

0.00

0.01

r

hAbs

orbance

Amide I'

90

0

POPC + tolaasin I A

090

Dichroism B

Wavenumber (cm-1)

Figure 16. Analysis of FTIR-ATR spectra of tolaasin I in POPC layerswith polarizer: (A) Spectra were taken with either parallel (0) or per-pendicular (90) polarization. The amide I region of tolaasin I boundto vesicles was reported after subtraction of the l ipid contribution. Thebest fit curve with Gaussian components (dotted lines) was superim-posed as a thick dashed line to the 90 polarized trace (solid line). Theabsorption bands in the parallel and perpendicular configuration wereused to calculate the orientation of the corresponding structural elementas reported in Table 7. Bands are: h (helix), (-structure), r (random).(B) Dichroic spectrum obtained by subtracting the 90 polarized spec-trum (after multiplication by Riso i.e. 1.54) from that at 0.

Table 7. Assignment and dichroic ratio of some IR bands observed in lipid vesicles with and without toxin.

Wavenumber (cm1) Vibrationa Direction ()b Dichroic Ratio Angle ()c

Lipid alone2920 as CH2 stretching 90 1.32 0.01 33 12850 s CH2 stretching 90 1.34 0.01 34 1LUVs +WLIP2920 as CH2 stretching 90 1.48 0.01 41 12850 s CH2 stretching 90 1.52 0.01 42 1LUVs +tolaasin I2920 as CH2 stretching 90 1.61 0.01 45 12850 s CH2 stretching 90 1.63 0.01 45 11660 Amide I helix 30 2.51 0.10 47 2aas: asymmetric; s: symmetric.bDirection of the variation of the dipole moment associated to the vibration with respect to the direction of the main molecular axis(aliphatic chain or-helix axis).cAverage angle between the direction of the molecular axis and the perpendicular to the crystal plane (i.e. membrane plane).


21/32

101



pathogenic bacteria have been determined. SomeEPSs have a very complex structure, being hetero-geneous branched heteropolymers. The differencesin the chemical composition and structures of theEPSs purified from different plant pathogens sug-gest that these molecules may play specific anddifferent roles in plant-pathogen interaction

(Evidente and Motta, 2002, p 581629).The involvement of extracellular polysac-charidic or glycopolysaccharidic substances inbacterial and fungal plant diseases has often beenreported (Hodgson et al. 1949; Harborne, 1983, p74382; Denny, 1995; van Alfen, 1989) but theirrole in the disease process is still to be clarified(Denny, 1995; van Alfen, 1989).

EPSs are considered either molecules able toavoid or delay the activation of plant defence, act-ing as signal in plant pathogen cross-talk as wellas they may have a significant role in the virulence

of the pathogen of interest. In several phytopatho-genic bacterial species belonging to the genusAgrobacterium, Clavibacter,Erwinia,Pseudomo-nas andXanthomonas, the production of polysac-charidic or glycopolysaccharidic substances wasproved to be associated with water soaking and/orwilting symptoms and, moreover, they played animportant role in plant colonisation by thebacterium (Denny, 1995). Furthermore, EPSs fromX. campestrispv. vesicatoria were proved to inducephytotoxic effects (chlorosis, necrosis, electrolyteleakage) on the homologous host (Walkes and

Garro, 1996).These macromolecules appear to interfer with

water movement in plant due to mechanical plug-ging of the vessels which leads to wilt symptomsdevelopment. The phenomenon appears to berelated to molecule size and viscosity rather thanto their structure (Harborne, 1983, p 74382),though some results on their apparent host specific-ity (Rudolph et al. 1989, p 177218) and on viscos-ity interference (McWain and Gregory, 1972)would suggest a possible different behaviour in atleast in some cases.

Lipopolysaccharides (LPSs) are complex mol-ecules of the outer membrane of Gram negativebacteria containing a polysaccharide (hydrophilic)tail and a lipid (hydrophobic) head, which isanchored in the outer membrane. The moleculeconsists of three segments: lipid A, core polysac-charide, and the antigenic O-chain. The core poly-saccharide is further subdivided into an innercore and an outer core. The three segments differ

in their composition, biosynthesis and biologicalfunction. Although this model of LPS was origi-nally proposed for Salmonella spp., the basicfeatures appear valid for most, if not all, Gram-negative bacteria (Chatterjee and Vidaver, 1986, p94114).

The isolation and structure determination of

LPSs from phytopathogenic bacteria, and in par-ticular that of the corresponding O-chain, assumedgreat importance for the possible role of this mol-ecule in the host-pathogen interaction. LPSs appearto play an important role in the interaction of cellsof pathogenic bacteria and plant and animal cellhosts (Medzhiton and Janway, 1997; 1998.; Dowet al. 2000). LPSs activate host defence systems ineither vertebrate and invertebrate inducing theproduction of antimicrobial peptides or, in mam-malian, that of immunoregulatory, inflammatoryand cytotoxic molecules. The use of mutants of

plant pathogenic bacteria defective or lacking LPSslead to assess their role in the virulence expressionand in the recognition process which takes placein the first phases of the interaction of the pathogenand plant cells.

The O-chain and lipid A, from the LPS ofP. tolaas ii , P. reactans and B. gladioli pv.agaricicola as well as the EPS of the latter wasisolated and the structures determined.

Lipopolysaccharide ofPseudomonas

tolaasiiDried cells of type strain NCPPB2192 ofP. tolaasiiwere extracted according to the water/phenolmethod (Westphal and Jann, 1965) and the LPSfraction was found exclusively in the phenol phase.The LPS material was subjected to vertical gelelectrophoresis and it appeared as a simple ladderlike pattern, typical of a regular repetitive lipopoly-saccharide species of smooth LPS form. A mild acidhydrolysis allowed the removal of the lipid A moi-ety by precipitation leaving the polysaccharide insolution which was then purified by gel filtration.The compositional analysis of pure O-polysaccha-ride, via acetylated O-methyl glycoside derivatives,showed the presence of 2-acetamido-2-deoxy-gulu-ronic acid and 2-acetamido-2,6-di-deoxy-glucose(QuipN). Methylation analysis showed the presenceof 3-substituted quinovosamine. The 1H NMR spec-trum of the polysaccharide showed in the anomericregion a variety of signals not all attributable toanomeric protons and not integrating for the same


22/32

102

Andolfi et al


area (Fig. 17). Therefore, a detailed 2D NMRanalysis (DQF-COSY, TOCSY, NOESY, gHSQC,gHMBC) allowed the complete assignment of allresonances. Interestingly, the HSQC spectrumshowed the signals of some spin systems (A and A)and (B and B) and these were indicative of theiranomeric positions.

A gHSQC spectrum registered without decou-pling allowed the measurement of the 1JC,H ano-meric coupling, thus demonstrating that spinsystem A and A have-configuration whereas thespin system B and B have -configuration. Resi-due A had typical proton and carbon signals of auronic acid substituted at C-4 whereas residue Ahad proton and carbon resonances of a 4-substituted3-O-acetyl uronic acid. The value of the couplingconstants and the 13C chemical shifts were distinc-tive of a gulo-configurated hexopyranose (DeCastro et al. 2003). The other spin systems (B and

B), showing the same resonances and both char-acterised by the anomeric signals, were identified3-substituted -QuipNAc. The acetyl substitutionat C-2 of all residues and at C-3 of residue A wasproven by a HMBC spectrum. The sequence of themonosaccharides in the repeating unit was inferredusing interresidual NOE data measured by 2D-

NOESY and long range scalar connectivitiesmeasured by HMBC spectrum (Fig. 18).

Selective and milder de-O-acetylation condi-tions were carried out (ammonium hydroxide at4 C for 4 days) yielded a polysaccharide with adisaccharide repeating unit as demonstrated by 2DNMR analysis. This experiment demonstrated

2-acetamido-2-deoxy-gulopyranuronamide and2-acetamido-2,6-di-deoxy-glucopyranose asconstituents of the disaccharide unit. The absoluteconfiguration of these residues was determined byusing both NMR and the Exciton Coupling Methodapproaches (Harada and Nakanishi, 1983) carriedout on the correspondingp-bromobenzoyl deriva-tives of O-methyl glycosides N-acetylated obtainedby a strong methanolysis of O-chain followed byacetylation.

The absolute configuration of both gulopyra-nose units derivative was established on the basis

of13C chemical shifts of the de-acetylated deriva-tive. The analysis has been carried out consideringthe two possible relative configurations for thecompound in the disaccharide repeating unit, thatis:4)--L-Gulp-NAcAN-(13)--D-QuipNAc-(1 and 4)--D-GulpNAcAN-(13)--D-QuipNAc- (1. The 13C chemical shifts of

Figure 17.1H NMR spectrum of the O-chain of the LPS from Pseudomonas tolaasii.


23/32

103



gulopyranose residue univocally indicate

L-configuration in agreement with previous data(Lipkind et al. 1988; Shashkov et al. 1988).

Hence, the O-chain of the LPS from the mush-rooms associated bacteriumP. tolaasii consists ina tetrasaccharide repeating unit built up of twounits 2-acetamido-2,6-di-deoxy-glucopyranoseand two units of 2-acetamido-2-deoxy-gulopy-ranuronamide, one of which is acetylated at C-3position, as shown below:4)--L-GulpNAc3AcAN-(13)--D-QuipNAc-(14)--L-GulpNAcAN-(13)--D-QuipNAc-(1 (Molinaro et al. 2003).

Lipopolysaccharide ofPseudomonasreactansSDS PAGE electrophoresis of water and phenolextracts arising from phenol-water extraction ofstrain NCPPB1311 ofP. reactans cells showed thatonly the phenol phase had the typical banding pat-tern of smooth LPS. Methylation analysis of thelatter fraction revealed the presence of terminal-Glc and 4-substituted-Glc units and, furthermore,its 1H NMR spectrum showed an anomeric signalat 5.41 ppm, as broad singlet, in addition to carbi-nolic signals in the range 3.484.21 ppm. All thesedata suggested an amylose-like structure charac-terised by an average-molecular weight of 300 KDaas obtained by gel-permeation cromatography.

The presence of 2-keto-3-deoxy-D-mannoocta-noic acid (Kdo), dodecanoic acid (12:0),3-hydroxydecanoic acid [10:0 (3-OH)],3-hydroxydodecanoic acid, [12:0 (3-OH)] in the

phenol phase confirmed the LPS nature of this

fraction. The mild-acid hydrolysis of the fractionremoved, as precipitate, the lipid A moiety leavingin the supernatant, the saccharide part which, puri-fied by gel-permeation chromatography, gave riseto two fractions; the one in the void volume rep-resented the polysaccharidic part.

The GC-MS analysis of the high-molecular-weight polymer showed the presence of GlcpNAc,QuipNAc4NAc, and alanine (Ala) and the D con-figuration of both sugar residues and the L con-figuration of Ala. The methylation data indicatedthe presence of 3-substituted glucosamine arising

from Glcp2Am and the 3-substituted bacil-losamine. The 1H and 13C NMR spectra (Figs. 19and 20) showed three anomeric signals of the sameintegral intensity, suggesting a trisaccharide repeat-ing unit. Sugar moieties are named with A-C lettersin order of decreasing chemical shift (Table 8).

The anomeric configurations, on the basis oftheir chemical shift values positively confirmed bytheir3JH-H and

1JC-H values measured from a cou-pled HSQC experiment, were established to be , and forA, B and C residues, respectively. Thecomplete assignment of all proton and carbonsignals was achieved by 2D homonuclear andheteronuclear experiments (DQF-COSY, TOCSY,ROESY, HSQC, HMBC). The 13C NMR spectrumwas also informative in showing several signals ofcarbons bearing nitrogen including two signalsassigned to a C=N and a CH3 group, respectively,of an acetamidino group (Am), the presenceof which was established by chemical andspectroscopic work carried out on the O-chain. The

Figure 18. Relevant HMBC correlation of the anomeric region of the O-chain from Pseudomonas tolaasii.


24/32

104

Andolfi et al


Residue H-1/C-1 H-2/C-2 H-3/C-3 H-4/C-4 H-5/C-5 H-6/C-6

QuipNAc4NAc 5.199 4.039 3.933 3.75

Documents

90) Andolfi Et Al., Review Perspective