OUP user menu

Small proteins of plant-pathogenic fungi secreted during host colonization

Martijn Rep
DOI: http://dx.doi.org/10.1016/j.femsle.2005.09.014 19-27 First published online: 1 December 2005


Small proteins secreted by plant pathogenic fungi in their hosts have been implicated in disease symptom development as well as in R-gene mediated disease resistance. Characteristically, this class of proteins shows very limited phylogenetic distribution, possibly due to accelerated evolution stimulated by plant–pathogen arms races. Partly due to lack of clues from primary sequences, insight into the biochemical functions or molecular targets of these proteins has been slow to emerge. However, for some proteins important progress has recently been made in this direction. Expression of the genes for small secreted proteins is in many cases specifically induced after infection, which should help to advance our still very limited understanding of how plant pathogens recognize and respond to the host environment.

  • Fungal pathogenicity
  • Avirulence
  • Cysteine-rich

1 Introduction

Although many fungi are notorious pathogens of plants, most fungi do not cause plant diseases and several species are even beneficial for plants as endophytes or root colonizers. This raises the question of what exactly distinguishes pathogenic fungi from their non-pathogenic relatives. In other words, what are the determinants of fungal pathogenicity towards plants? A priori, one would expect these to be involved in dealing with the active defence mechanisms that differentiate a living plant from a dead one. Active defence mechanisms may be countered by fungi in several different ways, including suppression of particular signal transduction or gene expression processes in plant cells, protection against antifungal compounds or enzymes, or, in the case of necrotrophic pathogens, induction of host cell death. From several studies, it has become apparent that, next to secondary metabolites, small secreted proteins can play important and even decisive roles in these processes. As a rule, the genes encoding these proteins are highly species-specific and may even differentiate pathogenic subspecies from closely related saprophytic subspecies. Small, secreted proteins are also interesting for another reason: they are by far the most common fungal avirulence factors. These are factors that can be recognized by the innate immune system of the host, leading to disease resistance.

Enough justification, then, to review the current state of knowledge on the role of small secreted proteins in virulence and avirulence in plant–fungus interactions. “Small” is a relative term, so a limit must be set. Here, I only consider proteins of less than 200 amino acid residues, which excludes hydrolases that are involved in cell wall breakdown and/or nutrient acquisition. I also limit this review to proteins that are likely or proven to be secreted in the host. Of those, I do not discuss proteins that have a structural function in the cell wall and belong to families that are widespread among fungi, like hydrophobins, metallothioneins and members of the ceratoplatanin family, even though some of these are important for pathogenic development [14]. I also do not review research on oomycetes, which morphologically resemble fungi and are also known to secrete proteins that play critical roles in plant–pathogen interactions [5].

2 Methods of identification

Various approaches have led to the identification of small, secreted proteins or their genes from plant pathogenic fungi.1 The most straightforward of these is isolation of the protein from the extracellular fluids of infected plant tissue, followed by protein sequencing by Edman degradation or tandem mass spectrometry. In this way, Avr4, Avr4E, Avr9 and five Ecp's (“Extracellular proteins”) from Cladosporium fulvum were identified, as well as Six1 (“Secreted in Xylem 1”) from Fusarium oxysporum, and two peptides from Uromyces vignae (see Table 1 for references). In some of these cases, necrosis-inducing activity in leaves was used to purify a protein from apoplastic extracts [69]. In a related approach, Nip1, ToxA and ToxB were isolated from axenic cultures based on their necrosis- or (in the case of ToxB) chlorosis-inducing activity [1013]. ToxA was directly shown to be produced inside the host [14], while both Nip1 and ToxB are strongly implied to be produced in the host by the change in virulence of fungi transformed with the respective genes [11] (L. Ciuffetti, personal communication).

View this table:
Table 1

Small proteins of pathogenic fungi secreted in host plants

ProteinSpeciesLengtha# CysbBiological activityInvolvement in virulence/pathogenicityMatching R-geneRegulation of gene expressionProtein localizationReferencesc
Avr2Cladosporium fulvum588Protease inhibitor, induces HR d in the presence of the tomato Rcr3 proteaseNot required for pathogenicity on tomatoeCf-2 (tomato)Leaf apoplast [19,54]
Avr4Cladosporium fulvum868Induces HRNot required for pathogenicity on tomatoHcr9-4D (Cf-4) (tomato)Induced in plantaLeaf apoplast; Fungal cell wall chitin [7,22,25,27,73]
Avr4ECladosporium fulvum1016Induces HRNot required for pathogenicityHcr9-4E (tomato)Leaf apoplast [9]
Avr9Cladosporium fulvum286induces HRNot required for virulence or pathogenicityCf-9 (tomato)Induced in planta; induced by nitrogen starvationLeaf apoplast [8,24,33,52,63,74]
Ecp1Cladosporium fulvum656Induces HRRequired for full virulence on tomatoCf-ECP1 (tomato)Induced in plantaLeaf apoplast [29,31,75]
Ecp2Cladosporium fulvum1424Induces HRRequired for full virulence on tomatoCf-ECP2 (tomato)Induced in plantaLeaf apoplast [29,31,75,76]
Ecp3Cladosporium fulvum(19 kDa)?Induces necrosis in some tomato accessions, possibly R-gene dependent(Cf-ECP3) (tomato)Leaf apoplast [42]
Ecp4Cladosporium fulvum1016Induces HRCf-ECP4 (tomato)Leaf apoplast [42]
Ecp5Cladosporium fulvum986Induces necrosis in some tomato accessions, possibly R-gene dependent(Cf-ECP5) (tomato)Leaf apoplast [42]
CgDN3Colletotrichum gloeosporioides560Suppresses cell deathRequired for pathogenicity on StylosanthesStrong expression in primary infection vesicles; induced by nitrogen starvation [17]
Six1Fusarium oxysporum100 or 189f6 or 8Required for full virulence on tomatoI-3 (tomato)Induced in plantaXylem sap [32,35]
Avr-PitaMagnaporthe grisea179g8Putative metalloproteaseNot required for pathogenicity on ricePi-ta (rice)Induced in plantaProbably translocated into host cells after secretion [30,38]
Avr1-CO39pMagnaporthe grisea673Not required for pathogenicity on riceConfers avirulence towards rice cultivar CO39Probably translocated into host cells after secretion [36] S. Leong (pers. commun.)
Pwl1Magnaporthe grisea1242Confer avirulence towards weeping lovegrass [37,39]
AvrL567AMelampsora lini1271A and B forms induce HRL5, L6, L7 (flax)Induced in plantaProbably translocated into host cells after secretion [62]
ToxAPyrenophora tritici-repentis1182Host-selective toxinRequired for pathogenicity on wheatExpressed in axenic cultureTranslocated into host cells after secretion [10,14,26,40]
ToxBPyrenophora tritici-repentis644Host-selective toxinRequired for pathogenicity on wheatExpressed in axenic culture [12,41,71]
Nip1Rynchosporium secalis6010Induces necrosis in several plant speciesNot required for pathogenicity on barleyRrs1 (barley)Expressed in axenic cultureLeaf apoplast (probably) [11,69]
‘peptide 1’, ‘peptide 2’Uromyces vignae500Induce HR in cowpea, resistant cultivar onlyExpressed in axenic cultureLeaf apoplast [6]
Mig1Ustilago maydis1615Not required for virulence or pathogenicity on maizeInduced in planta (until teliospore formation) [16]
Mig2-1Ustilago maydis140g8Not required for virulence or pathogenicity on maizeInduced in planta [16,34]
  • aOf mature protein; in bold: based on (partial) amino acid sequences; otherwise sizes are predicted from coding sequences.

  • bNumber of cysteine residues in (predicted) mature protein. ? indicates that the number of cysteines is unknown.

  • cNot exhaustive for some proteins on which a large number of papers have appeared.

  • dHR: hypersensitive response (R-gene dependent cell death).

  • e“Not required for pathogenicity” indicates that natural, pathogenic isolates exist that lack the (intact) gene, but no gene knock-out mutant has been studied to determine whether the gene contributes to virulence.

  • fThe 100 amino acid form of Six1 (“p12”) corresponds to the N-terminal part of the 189 amino acid form (“p22”).

  • gAssuming processing by a Kex2-like protease; the mature form of Avr-Pita was estimated to be 176 amino acids by Jia et al. [30].

  • hThe allelic PWL3 and PWL4 genes encode proteins similar to Pwl1 and Pwl2 but are apparently not expressed in planta.

  • i MIG2-3 is a pseudogene; MIG2-2 and MIG2-6 encode much larger proteins.

An entirely different approach is the enrichment of fungal DNA sequences that are specifically expressed in the host through cDNA subtraction methods. MIG1 and MIG2 from Ustilago maydis were isolated in this way, using differential display of infected versus non-infected maize leaves [15,16]. CgDN3 from Colletotrichum gloeosporioides was found through differential screening of a cDNA library from nitrogen-starved mycelium, and was then shown to be expressed in planta and functionally involved in interaction with the host [17]. An entirely different way in which genes for small secreted proteins have been found is genetic mapping based on an avirulence phenotype. With this approach, AvrL567 from Melampsora lini and Avr-Pita, Avr1-CO39, PWL1 and PWL2 from Magnaporthe grisea have been identified (see Table 1 for references). Secretion of the encoded proteins in planta was not shown directly in these cases but inferred from the presence of an N-terminal signal peptide, the avirulence phenotype associated with the gene, and gene expression in planta. Finally, an elegant screen based on induction of necrosis upon Agrobacterium tumefaciens/ PVX-mediated cDNA expression in leaves yielded coding sequences for previously identified secreted proteins [18] as well as Avr2 [19] from C. fulvum.

Exploration of genome sequences of plant-pathogenic fungi is now yielding large numbers of hypothetical small secreted proteins [20]. Extracellular proteomes of infected plants are currently being analysed as well, and together these approaches will soon lead to a sharp increase in the number of proteins that are known to be secreted by fungi during colonization of their host. The key question is of course: what roles do these proteins have in the interaction between fungi and plants?

3 Structural characteristics

With the current amount of protein sequences in public databases, amino acid sequence comparisons frequently yield the first clues on possible functions of novel proteins. However, database searches with sequences of small secreted proteins from fungi commonly do not yield homologues nor recognizable protein domains (the exception in Table 1 is Avr-Pita, which contains metalloprotease motifs). The only recognizable features are often the presence of a signal peptide for secretion and, in many but not all cases, an even number of cysteine residues (Table 1). These cysteines are implied, and in some cases proven [2124], to form cysteine-bridges. Such connections are likely to stabilize protein tertiary structure, as was supported by the protease sensitivity of Avr4 cysteine-replacement mutants [22,25], and by the requirement of the two cysteines of ToxA for its biological activity [26]. In the absence of obvious sequence similarity to proteins with known tertiary structure, structural information has to be obtained experimentally. Thus, far, this has been done with NMR for Avr4 and Avr9 from C. fulvum and Nip1 from Rynchosporium secalis. Nip1 forms a novel β-sheet fold with five disulfide bonds [21], while Avr9 forms a cystine-knot similar to carboxypeptidase inhibitors [23]. Avr4 was classified as an invertebrate-type chitin binding domain based on disulfide bonding pattern and limited sequence similarity and was indeed found to bind chitin [22,27]. Overall, however, primary and tertiary structures have given few clues on possible protein functions.

Some proteins listed in Table 1 are further processed after removal of the N-terminal signal peptide, but the significance of this is not known in any of these cases. Between one and 12 amino acids are apparently removed from the N-termini of Avr9 [28], Avr4 [7], Ecp1 and Ecp2 [29], while from Avr4 20 C-terminal amino acids encoded by the gene are removed as well [25]. Larger prodomains are present in ToxA (38 amino acids [26]) and Six1 (74 amino acids, H.C. van der Does, M. Meijer and M. Rep, unpublished results). The prodomain of Six1 ends with the sequence LRKR and is therefore probably removed by a Kex2-like protease in the secretory route. The same may be true for Avr-Pita [30] and the Mig2 proteins [15] (see Table 1). The prodomain of ToxA may be important for folding but its removal is not necessary for toxic activity [26]. Proteolytic processing after secretion does not necessarily have a ‘function’ but may simply be a consequence of the presence of proteases in the plant apoplast that remove N- or C-terminal ‘tails’ without disturbing the overall structure or function of the protein.

4 Virulence, avirulence and cell death

The most unambiguous way to establish whether a particular protein plays a role in virulence or avirulence is to create and analyse a gene knock-out mutant. In this way, a role in virulence has been proven for CgDN3 [17], ECP1, ECP2 [31] and SIX1 [32] and dismissed for AVR9 [33], MIG1 [16] and the MIG2 genes [15,34]. Interestingly, study of a CgDN3 knock-out mutant revealed that loss of pathogenicity appeared to be due to a cell death response of the host that was not seen when the wild type fungus was used for infection, suggesting that CgDN3 plays a role in suppression of host defence responses [17]. Also with a gene knock-out strategy, a requirement for avirulence on certain host cultivars was established for AVR9 [33] and SIX1 [35]. More commonly, a requirement for avirulence has been demonstrated by transformation of a candidate avirulence gene to a naturally virulent strain, as was done with genes from C. fulvum [7,9,19], R. secalis [11], and M. grisea [3639]. Conversely, a function in virulence was demonstrated for the host-selective toxin ToxA from Pyrenophora tritici-repentis by transformation of the gene to a naturally occurring non-virulent strain [40].

Instead of the creation of stable transformants, secreted proteins have been tested directly for suppression or induction of plant defence responses through injection into leaves of a host plant. In this way, P. tritici-repentis ToxA and ToxB were shown to be sufficient for the induction of cell death or chlorosis specifically in susceptible host plants [10,12,40,41]. In contrast, several avirulence proteins of C. fulvum were demonstrated to induce cell death only when injected in plants carrying the corresponding disease resistance gene [79,42]. A similar effect could be obtained by transient expression of the corresponding genes in leaf tissue using Agrobacterium-mediated transformation [9,43] or in whole plants using the PVX expression system [9,18,19,25,44]. A more complicated picture emerged from related experiments with Nip1 from R. secalis. This protein is the avirulence factor that is required for Rrs1-mediated resistance of barley [11]. However, infiltration of this protein into leaves can lead to necrosis in several plant species [45].

These examples illustrate that induction of cell death can be either related to resistance, in which case it is called the hypersensitive response (HR), or to susceptibility. In both cases programmed cell death appears to be involved, which is mediated by specific receptors and signal transduction pathways [46,47]. The difference in outcome may in part be attributed to the growth requirements of the fungus: biotrophic fungi require living tissue to proliferate and are therefore inhibited by a hypersensitive response, while necrotrophic fungi feed on dead tissue and take advantage of host cell death. Clear examples of proteins that promote disease by causing cell damage (death or chlorosis) are the above-mentioned host-specific toxins of P. tritici-repentis. From other fungi, proteins have been identified as well that cause cell death when injected into plant leaves [48,49]. However, in these cases it is still unclear if the observed cell death is relevant for the fungus–plant interaction, i.e., whether the proteins are actually produced in planta and whether their production is related to susceptibility or resistance of the host.

5 Targets

Ultimately, phenotypes of mutants and cell death responses of plants are not sufficient to understand the function of secreted proteins in molecular terms. A crucial step towards such an understanding is the identification of the molecular targets of a protein after its secretion. Much is still to be learned in this respect. There are some indications that secreted proteins can act through influencing the plasmamembrane H+-ATPase. Nip1 stimulates barley plasmamembrane H+-ATPase, which may be the cause of increased opening of stomata during infection [50]. Furthermore, ToxA-induced electrolyte leakage from wheat cells could be inhibited by vanadate, an inhibitor of H+-ATPases [51]. However, this does not mean that H+-ATPases are direct molecular targets since these effects could be indirect and the underlying protein–protein interactions have not yet been uncovered.

For Avr9 of C. fulvum, which is secreted into the apoplast, a high-affinity binding site (HABS) was found on the plasmamembranes of leaf cells of tomato and other solanaceous plant species using radiolabelled Avr9 [52]. The Cf-9 resistance gene of tomato confers Avr9-dependent resistance to C. fulvum, suggesting that the Cf-9 protein, a membrane protein with a large extracellular domain, could be a receptor for Avr9. However, direct binding of Avr9 to Cf-9 could not be demonstrated using various methods, so the presence of Avr9 is probably sensed only after it binds to its ‘virulence’ target – possibly the HABS [53]. A striking example of such indirect recognition was recently reported for Avr2, another avirulence factor of C. fulvum. Induction of cell death in tomato by Avr2 depends on the presence of the resistance gene Cf-2, which like Cf-9 encodes an extracellular receptor-like protein. It also depends on the presence of Rcr3, an extracellular tomato protease. Although not apparent from its primary sequence, Avr2 turned out to be an inhibitor of Rcr3, and it is likely that an Avr2-induced change in the conformation of Rcr3 triggers Cf-2 to activate an HR [54].

Unlike the resistance genes against Cladosporium, resistance genes against several other fungal pathogens turned out to encode intracellular proteins [5560]. This raised the question of whether recognition of the corresponding avirulence factors depends on translocation of the factor into the cytoplasm of plant cells, either by direct injection as is well established for bacteria [61], or through uptake by plant cells from the apoplast. In a still exceptional case, direct binding of Avr-Pita from the rice blast fungus M. grisea to its corresponding intracellular resistance protein was demonstrated [30]. Avr-Pita has a signal peptide for secretion, so presumably it is secreted by the pathogen and subsequently taken up by host cells. An intracellular target for proteins secreted by plant pathogenic fungi may not be uncommon. AvrL567 from flax rust is also likely to be transferred into host cells since artificial production of the protein inside plant cells that express a matching resistance gene results in HR [62]. The same has been observed for Avr1-CO39p from M. grisea (S. Leong, pers. commun.). It is still unclear how these proteins end up inside plant cells. A recent breakthrough in this respect was the demonstration that a GFP-tagged version of ToxA is internalized only by wheat cells from a ToxA-sensitive cultivar, and possibly interacts with a chloroplast protein (Manning and Ciuffetti, manuscript submitted for publication).

On the other hand, proteins secreted by fungi in host plants do not necessarily bind to a host protein – they may also confer protection against host defences in other ways. Such a function has been demonstrated for Avr4 from C. fulvum, which binds to chitin in the fungal cell wall [22] and, at least in vitro, protects fungi against plant chitinases that are commonly produced constitutively or in response to infections (H. van den Burg, personal communication).

6 The genes respond to the host environment

In most cases investigated, expression of the genes for small, in planta secreted proteins is low or undetectable under axenic conditions [15,16,38,39] (H.C. van der Does and M. Rep, unpublished observations). This has opened up the possibility to use these genes to identify the conditions or compounds that are sensed by the pathogen and cause it to reprogram its gene expression for survival and growth in the host. One clue that has been followed was that starvation for nitrogen in axenic culture induced expression of the AVR9 gene [63], which required the global nitrogen response factor NRF1 [64] as well as binding sequences for GATA-type regulators (like NRF1) in the promoter [65]. Also, expression of the gene for CgDN3 of C. gloeosporioides [11] is induced by nitrogen starvation. This has led to the hypothesis that fungal pathogens experience nitrogen starvation inside the plant, and that this may be a way to sense the host environment. However, tomato leaf apoplastic nitrogen content actually increases after infection by C. fulvum [66], and the enzyme that allows utilization of γ-aminobutyric acid, a major nitrogen source in the apoplast, is expressed during infection but not during nitrogen starvation [67]. Indeed, many genes for in planta secreted proteins investigated are not induced by nitrogen starvation or other types of stress [15,16,63] (H.C. van der Does and M. Rep, unpublished observations). Clearly, we know too little about the cues from the host that trigger specific gene expression patterns in plant pathogenic fungi. Promoters of the genes for small, in planta secreted proteins may help to find those cues and unravel the signal transduction pathways involved. One approach to this end are promoter deletion studies to identify elements that regulate in planta induction of expression, as has been initiated for MIG1 [16], MIG2-1 [15] and MIG2-5 [34]. At least one type of regulation that applies to the MIG1 and MIG2 genes may be histone modification, since deletion of a histone deacetylase gene derepresses expression of MIG1 [68] and MIG2 [34].

7 Do genes for small, secreted proteins undergo accelerated evolution?

None of the proteins listed in Table 1 are known to have close homologues beyond species or genus boundary. This makes the question of the natural history of the respective genes intriguing and at the same time hard to reconstruct. Even within a species the gene can be absent in part of the population (see, for instance [35,36,69]). Fast evolvement of coding regions due to molecular arms races between pathogens and hosts might explain the absence of close homologues in other fungal species. Indeed, the allelic PWL3 and PWL4 genes in M. grisea as well as the AvrL567 genes in Melampsori linii appear to have experienced accelerated evolution [37,62]. Also, the strong bias in favour of nonsynonymous substitutions between the SIX1 gene and the homologous pseudogene SIX1-H in F. oxysporum could be an indication of accelerated evolution [70] (M. Rep, unpublished observations). Accelerated evolution may have been enhanced by high frequency of gene duplications due to proximity to transposable and/or repetitive elements, a situation which appears to be common for genes for small, in planta secreted proteins [15,35,37,62,70,71]. Genomic locations enriched in repetitive elements could also explain cases of spontaneous gene loss in laboratory strains [35,39,70]. A telomeric location may cause gene instability as well, as was shown for AVR-Pita [38]. But if selection against avirulence and unstable genomic locations provide ample incentive for gene loss, what about gene gain? Is vertical inheritance and selective retention of fast evolving genes sufficient to explain the absence of homologues in related fungi? In some cases, horizontal gene transfer between fungi could contribute to a ‘patchy’ phylogenetic distribution of genes involved in host-specific pathogenicity, as has been suggested for the PEP (Pea pathogenicity) genes of Nectria haematococca that are found in only two out of 16 species of the Nectria/Fusarium solani species complex [72]. However, many small secreted proteins appear not to be required for pathogenicity (Table 1), which would exclude selection on pathogenicity as an explanation for their selective retention, irrespective of whether they were obtained vertically or horizontally. On the other hand, a role of these proteins in virulence on unknown hosts or on fitness under untested environmental conditions can never be ruled out completely.

With the current pace of gene and protein identification from many different plant pathogenic fungi, and the unique story that each single protein has in store regarding its structure, evolution, biological function and molecular targets, certainly much exciting work lies ahead with this remarkable class of fungal proteins.


Many thanks to Charlotte van der Does, Harrold van den Burg and Frank Takken for their useful suggestions for improvement of the manuscript. Thanks also to Sally Leong and Lynda Ciuffetti for sharing unpublished results. The research of M. Rep has been made possible by a fellowship of the Royal Netherlands Academy of Arts and Sciences.


  • 1 I here consider a protein as ‘identifieD' when its coding sequence has been obtained but also when only a part of the protein has been sequenced (of the proteins listed in Table 1, the latter is the case for Ecp3 and the two peptides from U. vignae).


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
  44. [44].
  45. [45].
  46. [46].
  47. [47].
  48. [48].
  49. [49].
  50. [50].
  51. [51].
  52. [52].
  53. [53].
  54. [54].
  55. [55].
  56. [56].
  57. [57].
  58. [58].
  59. [59].
  60. [60].
  61. [61].
  62. [62].
  63. [63].
  64. [64].
  65. [65].
  66. [66].
  67. [67].
  68. [68].
  69. [69].
  70. [70].
  71. [71].
  72. [72].
  73. [73].
  74. [74].
  75. [75].
  76. [76].
View Abstract