Saponins and plant defence — a soap story

Saponins and plant defence — a soap story

reviews Saponins and plant defence asoapstory oo m Saponins are plant glycosides that derive their name from their soap-like properties. They occur...

2MB Sizes 0 Downloads 2 Views


Saponins and plant defence asoapstory oo


Saponins are plant glycosides that derive their name from their soap-like properties. They occur in a great many plant species, and have been implicated as pre-formed determinants of resistance to fungal attack. A number of fungi that succeed in breaching these antimicrobial plant defences produce saponin-detoxifying enzymes. The importance of one of these (avenacinase, produced by Gaeumannomyces graminis) in determining host range has been demonstrated. Recently, avenacinase has been shown to be closely related to another saponin-detoxifying enzyme, tomatinase, which is produced by the tomato pathogen Septoria lycopersici, suggesting that common mechanisms for saponin detoxification may be widespread in phytopathogenic fungi. S

aponins consist of triterpenoid, steroid or steroidal glycoalkaloid molecules bearing one or more sugar chains ~-3. They are characterized by their surfactant properties, and give a stable foam when shaken with water. The detergent action of saponins was recognized hundreds of years ago, when extracts of plants such as Saponaria officinalis were used to make soap (Fig. 1). The chemical structures of key saponins which will feature in this review are illustrated in Fig. 2. Saponins such as avenacin A-1 from oat roots (a triterpenoid) and e-tomatine from tomato (a steroidal glycoalkaloid) each have a single sugar chain attached to carbon C-3 via oxygen, and so are referred to as monodesmosidic saponins (from the Greek, desmos, meaning chain) (Fig. 2a). Bisdesmosidic saponins have two sugar chains, usually one at the C-3 carbon, and one at C-28 (triterpenoid saponins) or at C-26 (steroidal saponins). Bisdesmosidic saponins lack many of the properties and activities of monodesmosidic saponins, but can be converted into the biologically active monodesmosidic form by removal of the sugar at the C-26 or C-28 position. For some bisdesmosidic saponins, specific plant glycosyl hydrolases that carry out this conversion have been identified 4-6. The best characterized of these enzymes is the oat enzyme avenacosidase ~, which activates avenacosides A and B, the bisdesmosidic steroidal saponins found in oat leaves (Fig. 2b). In healthy oat leaf tissue, avenacosidase is confined to the plastids 6, while avenacosides A and B are located in the vacuole 7. Wounding of plant tissue resulting from mechanical damage or pathogen attack causes a breakdown of compartmentalization, allowing the enzyme to come into contact with its saponin substrates. The C-26 glucose molecule is then removed by hydrolysis to yield the monodesmosidic 26-desgluco avenacosides A and B, which are fungitoxic. Although bisdesmosidic saponins such as the avenacosides lack antimicrobial activity, they are still regarded as pre-formed inhibitors because their conversion to the monodesmosidic form does not require induction of gene expression in response to pathogen attack.

Toxicity of saponins to fungi The primary mode of action of saponins towards fungi is similar to that of polyene antibiotics, and involves the formation of complexes with membrane sterols~.3. This results in pore formation and loss of membrane integrity. The precise way in which saponins become incorporated into membranes is unclear, although various models have been pro4

January 1996, Vol. 1, No. 1

posed 1. The nature of the aglycone and the oligosaccharide moiety of the saponin molecule are both likely to contribute to the membraneolytic properties of individual saponins 3. In the case of the steroidal glycoalkaloids, antifungal activity has been shown to be pH-dependentS, ~. The susceptibility of membranes to saponin action may be influenced by sterol composition, since detailed sterol-binding studies involving the steroidal glycoalkaloid e-tomatine indicate a requirement for membrane sterols with free 3#-hydroxyl groups 1°. The possibility that saponins may also have more subtle effects on fungi in addition to their membraneolytic action cannot be excluded, although little is known about this area. An alternative mechanism of action of saponins has been proposed in which the aglycone, rather than the s~ponin itself, is believed to be the membraneolytic agent 4. Here the sugars are regarded as a solubilizing group aiding delivery of the saponin to the membrane, where membrane-bound glycosyl hydrolases are then presumed to activate the saponin by converting it to the aglycone. This view has received little support because it is incompatible with evidence showing that the free aglycone of ~-tomatine does not bind cholesterol 11, and that saponins affect the permeability of artificial membranes (which lack hydrolytic enzymes)8. It is not clear how plants protect themselves from the lytic effects of their own saponins. The mechanism of action of saponins is relatively nonspecific, and may be expected to affect all eukaryotic organisms which have sterols in their membranes. The avenacosides are stored as bisdesmosidic molecules and so do not pose a threat to plant cell membrane integrity. The (~-tomatine and other monodesmosidic saponins, however, are stored as biologically active molecules. The subcellular location of these other saponins is uncertain. They may be sequestered in the vacuole or in other organelles, the membranes of which may avoid lysis by having reduced overall sterol levels or by containing a high proportion of sterols substituted at the 3~-hydroxyl position lo.

Fungal strategies for dealing with saponins How do fungi that infect saponin-containing plants protect themselves? Pathogens such as Cladosporium fulvum may achieve this simply by circumventing the host plant saponin altogether. This biotrophic fungus restricts its growth to the intercellular spaces of tomato leaves, thus avoiding the release of ~-tomatine. Other tomato-infecting fungi may lower the pH at the infection site to levels at which ~-tomatine is ineffective as a membraneolytic agent 4. @1996, Elsevier Science Ltd

reviews In general, however, there appear to be two main mechanisms of resistance of fungi to saponins ~2. The first of these involves intrinsic resistance by virtue of membrane composition. Oomycetes such as Pythium and Phytophthora lack membrane sterols and are highly resistant to saponins in vitro ~3.These fungi can incorporate exogenous sterols into their membranes, and show an increase in sensitivity to saponins when sterols are added to the growth medium14. Conversely, fungi that normally contain sterols in their membranes become more resistant to saponins when they are grown in the presence of inhibitors of sterol biosynthesis 15. The importance of membrane composition in determining the ability of fungi to infect saponin-containing plants has been illustrated by experiments involving a tomato-attacking isolate of Fusarium solani. The wild-type isolate was pathogenic only to ripe tomato fruits (which contain relatively low levels of c~-tomatine) and was unable to infect green fruits (which are rich in ~-tomatine). However, mutants of this isolate defective in sterol biosynthesis showed increased resistance to a-tomatine, and were able to infect green fruit ~5. The second major mechanism of resistance of fungi to saponins involves the production of saponin detoxifying enzymes. These enzymes will be the focus of the remainder of this review.

Saponin-detoxifying enzymes A number of phytopathogenic fungi produce specific glycosyl hydrolases that remove sugar molecules from the glycosy] chains at the C-3 carbon position, to give products which are less toxic to fungal growth4,~2,~7. Some of these enzymes are constitutively expressed, while others are induced by their saponin substrates. Despite the widespread distribution of saponins in the plant kingdom, detailed studies of saponin detoxification by fungi have been restricted primarily to pathogens of oat and tomato. This is probably because the structures and antifungal properties of oat and tomato saponins are well documented, and the saponin profiles of these plants relatively simple (in contrast to alfalfa, for example, where over 20 different saponins have been identified). The importance of saponin-detoxifying enzymes in determining fungal pathogenicity to plants has recently been demonstrated for oat-attacking isolates of the soil fungus Gaeumannomyces graminis, which detoxify the oat root saponin avenacin A-1 (Ref. 18).

Detoxification of oat root saponins by Gaeumannomyces

graminis Oaeumannomyces graminis vat. tritici is a major pathogen of wheat and barley, and causes the disease known as 'take-all'. Oats are resistant to infection by G. grarninis var. tritici, but another variety of G. graminis exists (var. avenae) which can infect oats as well as wheat 12. Oat roots contain a group of structurally related triterpenoid saponins known as avenacins, of which avenacin A-1 has been implicated as the major antifungal component. Isolates of G. graminis var. avenae are relatively resistant to avenacin A-1 in vitro, while G. graminis vat. tritici isolares are more sensitive. Consequently, avenacin A-1 has been implicated in determining the resistance of oats to attack by G. graminis var. tritici 19. The localization of this saponin in the epidermal cells of roots is consistent with the observation that G. graminis var. tritici isolates rarely penetrate beyond the outer cell layers of oat roots2° (Fig. 3).

Fig. 1. Soapwort (Saponaria officinalis). In the past, soapwort was commonlycultivated near wool mills so that the soapy extracts from its leaves and roots could be used for washing wool.Reproduced, with permission, from Ref. 31.

There is very little variation to be found in the levels of the avenaeins in the genus Arena. However, one diploid oat species (Avena longiglumis) has been shown to lack avenacin A-l, and significantly this oat species is susceptible to infection by G. graminis var. tritici 2°. The resistance of isolates of G. graminis vat. avenae to avenacin A-1 is associated with their ability to produce the saponin-detoxifying enzyme, avenacinase 17,~1,22. Avenacinase is a ~-glucosyl hydrolase which can remove both the ~,1-2- and ~,l-4-1inked terminal D-glucose molecules from avenacin A-1 (Fig. 2a) to give products that are less toxic to fungal growth. The gene encoding avenacinase (AVN1) has been isolated, and the cloned DNA used to generate avenacinase-minus mutants by homologous recombination 18. These mutants showed the anticipated increase in sensitivity to avenacin A-1 in vitro, and were no longer able to infect oats (Fig. 4). They were, however, still fully pathogenic to wheat, which is not known to contain saponins. These experiments indicate that avenacinase is a determinant of host range, but that it is not required for pathogenicity to alternative hosts which do not contain avenacin A-1. Unexpectedly, DNA sequences that cross-hybridize to the AVN! gene are also found in many isolates ofvar, tritici and in other related Gaeumannomyces and Phialophora January 1996, Vol. 1, No. 1


reviews species 12. The fungi that contain these cross-hybridizing DNA sequences all (a) have a very weak ability to deglucosyl%' O ate avenacin A-l, which is detectable only after prolonged incubation with A v~ O HNMe the substrate. This activity resides in 'avenacinase-like proteins' or ALPs, which ~-D-glu(1--~2) ~ "~ "-,~:v "*OH have very similar physicochemical ~/{X-L-ara(1-e)O ~ Avenacin A-1 properties to avenacinase and which are specifically recognized by polyclonal ~-g-glu(1-->4) anti-avenacinase antisera. Purification of the ALP from G. graminis var. tritici indicated a specific activity for avenacin A-1 which was 25 times lower than that of avenacinase from var. avenae. Transformation of the A V N I gene into 13-o-glu(1 var. tritici will test whether the inabil--->2'lS-g-glu(1)-->4)--13-g-gal(-->)O 1 atine ity to effectively detoxify avenacin A-1 / is the only factor preventing var. tritici 13-D-xyl(1-->3) from infecting oats. The function of the (b) ALPs is not known. While they do not appear to be able to deglucosylate ; ~ . . . ~ ' ~ O(~--1)13-D-glu* avenacin A-1 effectively, they may be important for detoxification of other saponins encountered during infection of alternative hosts such as wild grasses, ~-D-glu(1--->3)--6-D-glu(1--~2) the saponin contents of which are ~ A i e ~-o-glu nacoside B (1~)o unknown. / Alternative mechanisms of avenacin C~-L-rha(1-->4) resistance do exist in G. graminis. A group of isolates from Australia were Fig. 2. Structures of oat and tomato saponins. (a) The monodesmosidic saponins originally classified as var. tritici on the avenacin A-1 (a triterpenoid oat root saponin) and a-tomatine (the tomato steroidal basis of ascospore length, but are able glycoa]kaloid); (b) the bisdesmosidic foliar oat saponin avenacoside B. These three to infect oats and are resistant to saponins all have a sugar chain attached to carbon C-3 via oxygen. Avenacoside B avenacin A-1. These isolates do not, has an additional sugar moiety ([3-D-glucose) attached via oxygen to C-26. however, produce avenacinase in vitro, Avenacoside A differs from avenacoside B only in that it lacks the terminal ~(1-3)and they lack a DNA homologue for the linked D-glucose molecule. The 26-desgluco avenacosides are formed by removal of AVN1 gene 23. In this case, saponin the glucose indicated by an asterisk. resistance is likely to be due to some form of intrinsic resistance at the membrane level. Ribosomal DNA sequence analysis suggests that these isolates may be more closely related to var. avenae than to var. tritici, despite the differences in ascospore morphology.

Fungal detoxification of c~-tomatine

Fig. 3. Location of the autofluorescent saponin avenacin A-1 in oat root epidermal cells~°. The figure shows a crosssection of a young root of Arena strigosa. The diameter of the root is 0.39ram; the excitation and emission wavelengths are 365 and 397 nm, respectively.



Saponin-detoxifying enzymes are also produced by a number of fungi which infect tomato 12. These enzymes act on the tomato steroidal glycoa]kaloid ~-tomatine and are collectively referred to as tomatinases, although they do not all work in the same way (Fig. 5). ~-tomatine has a tetrasaccharide moiety (~-lycotetraose) consisting of two molecules of glucose and one each of galactose and xylose, attached to the C-3 carbon. The tomatinases of Septoria Iycopersici and Verticillium albo-atrum both remove a single terminal glucose molecule from the lycotetraose group to give ~2-tomatine, while those of Botrytis cinerea and Fusarium oxysporum f. sp. lycopersici release the intact lycotetraose group to give the aglycone, tomatidine. Alternaria solani, another tomato-infecting fungus, also converts ~-tomatine to the aglycone but does so by release of the four monosaccharides. These deglycosylation events all destroy the ability of the saponin to complex with membrane sterols.

reviews avenacinase and S. lycopersici tomatinase belong to the latter 24 (Fig. 6). The members of family 3 most closely 8 ' T 8 related to avenacinase and tomatinase / are ~-glucosyl hydrolases from other = 6 fungi that have been studied because of 5 4 ; 5 their importance in cellobiose degradation (BGL1 from Trichoderma reesei, 4 :: ::: 4 BGL1 and BGL2 from Saccharoo_ i 3 1 mycopsis fibuligera, and BGLS from Candida pelliculosa ). Avenacinase and tomatinase are 11 1 more closely related to each other than C Ggt Gga Gga C Ggt Gga Gga either is to the other members of family Avn3, and may represent a subgroup of Fig. 4. Pathogenicity of wild-type and mutant fungi to (a) wheat and (b) oats 18. this family having the additional ability C = uninoculated control; Ggt, Gga and Gga Avn- -- plants inoculated with G. graminis to degrade saponins 24. More extensive var. tritici, var. avenae and an avenacinase-minus mutant ofvar, avenae, respectively. studies of substrate specificities of Pathogenicity was scored on an arbitrary scale of 0 to 8. these enzymes are required in order to establish whether this is*the case. Currently the activity (if any) of the The tomatinase enzymes of S. lycopersici and V. albo- other enzymes towards saponins has not been described atrum have a similar mechanism of action to that of aven- and, while avenacinase does have some activity towards acinase, since these enzymes all remove a terminal ~,l-2- cellobiose, it is considerably more active towards avenacin A-1. linked D-glucose molecule from their respective saponin Another less closely related member of family 3 has also been substrates by hydrolysis. Despite this similarity, avenaci- implicated in determining host range. The ~-glucosyl hydronase and these ~-tomatine degrading enzymes have been lase CBG1 from Agrobacterium tumefaciens hydrolyses coniconsidered separately in the literature, and until recently ferin to coniferyl alcohol, and has been associated with viruwere not believed to be related. The purification of tomati- lence to g~mnosperms 2s. nase from S. lycopersici has changed this view by revealing While the two saponin-detoxifying enzymes that have that tomatinase and avenacinase are closely related pro- been characterized from fungi belong to the family 3 group reins which share many properties, including immunologi- of ~-giucosyl hydrolases, the oat enzyme avenacosidase cal cross-reactivity24. This relatedness was exp]oited to (which activates the foliar oat saponins avenacosides A and isolate the tomatinase gene from S. lycopersici by using an B by removal of the C-26 glucose) belongs to family 1. Interestingly, family 1 also contains two other plant avenacinase cDNA clone as a heterologous probe. While tomatinase and avenacinase are 68% similar at the enzymes which activate pre-formed inhibitors of antifungal amino acid level, there are clear differences in their sub- substances. These are myrosinase and linamarase, strate specificities which reflect the host preferences of the which hydrolyse glucosinolates and cyanogenic glucosides, fungi from which the enzymes originated. Avenacinase has respectively29, 3o. a relative rate of hydrolysis towards a-tomatine of 2% of that towards avenacin A-l, while the S. lycopersici tomatinase enzyme also has only weak activity towards the nonSeptoria lycopersici host saponin (less than 0.01% of its activity towards ~-tomatine)24. The specificity of the S. lycopersici tomatinase for Verticillium albo-atrum ~-tomatine is influenced by the nature of the aglycone as well as the sugar chain, since this enzyme only has a weak ability to deglucosylate another saponin, F-gitonin, which has the same sugar chain as e-tomatine, but which is a 13-D-glu(l~2) steroidal saponin rather than a steroidal glycoalkaloid25. To date, of the various fungal tomatinase enzymes which have o,ycooo been described that of S.lycopersici is the only one which had been fully purified and for which the corresponding gene has been cloned. It should now be possible to generate tomatinase-minus mutants of S. lycopersici by targeted gene disruption in order to test the role of this enzyme in I Alternaria solani ~ l pathogenicity.

I i:t T .....



Relatedness of saponin degrading enzymes to other glucosyl hydrolases Glycosyl hydro]ases have been grouped into approximately 40 families on the basis of sequence similarities26, 27. Within this classification the majority of ~-glucosyl hydrolases (EC fall into two of these families (family 1 and family 3). Gaeumannomyces graminis var. avenae

Fusarium oxysporum f. sp. lycopersici Botrytis cinerea

Fig. 5. Funga] pathogens of tomato employ a variety of mechanisms to detoxify ~-tomatine.

January 1996, Vol. 1, No. 1





S e l Tom Trr Bgll Sam B g l l Sam B g l 2 Cap G l u c b Agt Cbgl Asw G l u a 3 Consensus


G D E E E E A -| i L G F Q G F V

~ A I YT


/GiQ ~ ~ ~ ~

T~ g

A , I E EmVN

g i N S ~


A ~ Q HiA G ~ L F G S ~ T A A ~ H r ~



Yi I S ~ ~



~ G

E D S~"


Fig. 6. Relatedness of avenacinase and tomatinase to other family 3 ~-glucosylhydrolases24. The aspartic acid residue identified as the active site in GluA3 from Aspergillus wentii is indicated by an asterisk. Residues found in the majority of sequences are highlighted in black, and conservative substitutions are indicated by shaded boxes. Abbreviations: Gga Avn and Se] Tom (avenacinase from G. graminis and tomatinase from S. lycopersici, amino acids 268-300 and 269-301, respectively)(GenBank accession numbers U35463 and U35462); Trr Bgll (amino acids 256-288), ~ m Bgll (amino acids 284-316) and Bgl2 (amino acids 288-320), Cap Glucb (amino acids 288-320), Ag%Cbgl (amino acids 211-243) and Asw Glua3 (partial sequence derived from a peptide fragment) are ~-glucosylhydrolases from Trichoderma reesei, Saccharomycopsis fibuligera, Candida pelliculosa, Agrobacterium tumefaciens and Aspergillus wentii, respectively (EMBL accession numbers U09580, M22475, M22476, X02903, and GenBank accession number M59852).

Saponins and their detoxification - general significance for plant-pathogen interactions? The occurrence of DNA sequences related to avenacinase is not restricted to Gaeumannomyces and S. lycopersici. Southern blot analysis indicates that cross-hybridizing DNA sequences exist in a wide range of other fungi, although the significance of this observation is unclearu. While these sequences may encode family 3 ~-glucosyl hydrolases required simply for saprophytic growth and with no role in pathogenicity, it is also possible that some may have a role in detoxification of host plant saponins. This can now be readily tested by isolation of these DNA sequences using the cloned avenacinase or tomatinase genes, or PCR primers derived from their sequences. Targeted gene disruption will allow the generation of specific mutants which can then be tested for altered host range or pathogenicity. In this way it should be possible to determine whether the products of these DNA sequences have any role in infection of plants. This approach has the advantage that it can be applied to fungi for which the possible significance of saponin detoxification in the disease process has not yet been studied. Implications for disease control If saponin-'saponinase' combinations do have widespread significance for determining the host range of phytopathogenic fungi then this phenomenon may provide the basis for novel crop protection strategies which give control of a wide range of pathogens. Avenacinase and tomatinase are both extracellular enzymes and may present attractive accessible targets for inhibitor-based approaches. The manipulation of plant saponin biosynthetic pathways may also offer potential for disease control, but first it is important to establish the contribution of saponins to disease resistance. The isolation of plant mutants defective in saponin biosynthesis will allow this to be tested directly. Acknowledgements The Sainsbury Laboratory is supported by the Gatsby Charitable Foundation. I would like to thank Liz Atchison for providing the illustration of soapwort, Simon Southerton for providing Fig. 3 and my colleagues for their critical corn-


Janua~1996,Vol1,. No.1

ments on the manuscript. Because of a restriction on the number of references, in some cases reviews have been cited in place of the original articles.

References 1 Price, K.R., Johnson, I.T. and Fenwick, G.R. (1987) The chemistry and biologicalsignificance of saponins in food and feedingstuffs, CRC Crit. Rev. Food Sci. Nutrition 26, 27-133 2 Hostettmann, K., Hostettmann, M. and Marston, A. (1991) Saponins, Methods in Plant Biochemistry 7, 435-471 3 Fenwick, G.R. et al. (1992) Saponins, in Toxic Substances in Crop Plants (D'Mello,J.P.F., Duffus, C.M. and Duffus, J.H., eds), pp. 285-327, The Royal Society of Cambridge 4 SchSnbeck, F. and Schbsser, E. (1976) Preformed substances as potential protectants, in Physiological Plant Pathology (Heitefuss, R. and Williams, P.H., eds), pp. 653-678, Springer-Verlag 5 Lttning, H.U. and SchlSsser, E. (1975) Role of saponins in antifungal resistance V. Enzymatic activation of avenacosides, Z. Pflanzenkrankh. Pflanzenschutz 82,699-703 6 Nisius, H. (1988) The stromacentre inAvena plastids: an aggregation of ~-glucosidase responsible for the activation of oat-leaf saponins, Planta 173,474-481 7 Kesselmeier, J. and Urban, B. (1983) Subcellu]ar localization of saponins in green and etiolated leaves and green protoplasts of oat (Arena sativa), Protoplasma 114, 133-140 8 Roddick, J.G. and Drysdale, R.B. (1984) Destabilization of liposome membranes by the steroidal glycoalkaloid a-tomatine, Phytochemistry 23, 543-547 9 Fewell, A.M. and Roddick, J.G. (1993) Interactive antifungal activity of the glycoalkaloids ~-solanine and ~-chaconine,Phytochemistry 33, 323-328 10 Steel, C.C. and Drysdale, R.B. (1988) Electrolyte leakage from plant and fungal tissues and disruption of liposome membranes by ~-tomatine, Phytochemistry 27, 1025-1030 11 Arneson, P.A. and Durbin, R.D. (1968) Studies on the mode of action of tomatine as a fungitoxic agent, Plant Physiol. 43, 683-686 12 Osbourn, A.E. et aI. (1994) Detoxification of plant saponins by fungi, in Molecular Genetics of Plant-Microbe Interactions (Daniels, M.J., Downie, J.A. and Osbourn, A.E, eds), pp. 215-221, Kluwer 13 Arneson, P.A. and Durbin, R.D. (1967) Hydrolysis of tomatine by Septoria Iycopersici; a detoxification mechanism, Phytopathology 57, 1358-1360 14 SehlSsser, E. (1972) Sterol dependent membranelytic action of saponins, Phytopath. Z. 74, 91-94

reviews 15 Olsen, R.A. (1973) Triterpeneglycosides as inhibitors of fungal growth and metabolism. 6. The effect of aescin on fungi with reduced sterol contents, Physiol. Plant. 29, 145-149 16 D6fago, G. and Kern, H. (1983) Induction ofFusariam solani mutants insensitive to tomatine, their pathogenicity and aggressiveness to tomato fruits and pea plants, Physiol. Plant Pathol. 22, 29-37 17 Crombie, W.M.L. et al. (1986) Pathogenicity of the take-all fungus to oats: its relationship to the concentration and detoxification of the four avenacins, Phytochemistry 25, 2075-2083 18 Bowyer, P. et al. (1995) Host range of a plant pathogenic fungus is determined by a saponin detoxifying enzyme, Science 267, 371-374 19 Turner, E.M. (1953) The nature of resistance of oats to the take-all fungus, J. Exp. Bet. 4, 264-271 20 Osbourn, A.E. et al. (1994) An oat species lacking avenacin is susceptible to infection by Gaeumannomyces graminis var. tritici, Physiol. Mel. Plant Pathol. 45, 457467 21 Turner, E.M. (1961) An enzymic basis for pathogenic specificity in Ophiobelus graminis, J. Exp. Bet. 12, 169-175 22 Osbourn, A.E. et al. (1991) Partial characterization of avenacinase from Gaeumannomyces graminis var. avenae, Physiol. Mol. Plant Pathol. 38, 301-312 23 Bryan, G.T., Daniels, M.J. and Osbourn, A.E. (1995) Comparison of fungi within the Gaeumannomyces graminis-Phialophora complex by analysis of ribosomal DNA sequences, Appl. Env. Microbiol. 61, 681-689 24 Osbourn, A.E. et al. Fungal pathogens of oat roots and tomato leaves employ closely related enzymes to detoxify host plant saponins, Mol. Plant-Microbe Int. (in press)

25 Sandrock, R.W., DellaPenna, D. and VanEtten, H.D. Purification and characterization of ~2-tomatinase, an enzyme involved in the degradation of c~-tomatine and isolation of the gene encoding 132-tomatinase from Septoria lycopersici, Mol. Plant-Microbe Int. (in press) 26 Henrissat, B. (1991) A classification of giycosyl hydrolases based on amino acid sequence similarities, Biochem. J. 280, 309-316 27 Henrissat, B. and Bairoch, A. (1993) New families in the classification of glycosyl hydrolases based on amino acid sequence similarities, Biechem. J. 293, 781-788 28 Castle, L.A., Smith, K.D. and Morris, R.O. (1992) Cloning and sequencing of an Agrobacterlum tumefaciens ~-glucosidase gene involved in modifying a vir-inducing plant signal molecule, J. Bact. 174, 1478-1486 29 Lenman, M. et al. (1993) Characterization of a Brassica napus myrosinase pseudogene: myrosinases are members of the BGA family of J3-glycosidases, Plant Mol. Biol. 21, 463~I74 30 Oxtoby, E. et al. (1991) Nucleotide and derived amino acid sequence of the cyanogenic ~-glycoside (linamarase) from white clover (Trifolium repens L.), Plant Mel. Biol. 17, 209-219 31 Curtis, W. (1835) Flora Londinensis IV (Bohn, H., ed.), London

Anne Osbourn is at the Sainsbu~ Laboratory, The John Innes Centre, Colney Lane, Norwich; UK NR4 7UH,

Fungal hydrophobins: proteins that function at an interface Joseph G.H° Wessels Hydrophobins are small, moderately hydrophobic proteins secreted by fungi and containing eight cysteine residues in a conserved pattern. Most hydrophobins have been discovered as putative products of genes abundantly expressed in development, pathogenesis and symbiosis. Those that have been purified exhibit interfacial selfassembly into amphipathic protein films that can be extremely insoluble. These protein films arise at the surface of emergent structures, such as aerial hyphae, fruit bodies and air-borne spores and they line air channels in tissues, conferring hydrophobicity to these surfaces. Hydrophobins may also be responsible for adherence of hyphae to each other and to hydrophobic surfaces of other organisms, as in pathogenic interactions. he intimate relationships between plants and fungi is best appreciated if it is realized that these organisms belong to different kingdoms that have evolved separately but in complete interdependence. Although fungi may superficially resemble plants, they appear evolutionarily closer to animals than plants 1. That fungi depend on plant photosynthesis for nutrition is obvious, but that land plants could not have evolved without fungi is less generally appreciated. First of all, fungi are the principal degraders of dead plant remains, particularly of the lignocellulosic cell wall, and thus return to the atmosphere an estimated 25-50 x 109 tons of carbon annually fixed by plants. Of course, this is because they secrete large quantities of cellulases, ligninases and other enzymes, but equally important is their


(#1996, Elsevier Science Ltd

mode of growth by means of apically extending hyphae2. Under turgor, while secreting the depolymerizing enzymes, such hyphae can penetrate solid substrata such as wood and digest them from within a. Many fungi also penetrate living plants, but only in a few cases does this lead to death of the plant. In most cases a symbiotic relationship is set up. It is relatively recently that plant biologists have come to realize the enormous importance, at least in natural habitats, of the mutualistic symbioses between fungi and plant roots, known as mycorrhizas. These associations occur with nearly all plant species and without them there would be no rain forests nor heathlands, to mention just two well-known biotopes 4. While obtaining photosynthates from the plant, the fungus aids the plant with recruitment of water and minerals from poor soils and even recycles organic carbon January 1996, Vol. 1, No. 1