Autophagy vitalizes the pathogenicity of pathogenic fungi

Abstract

Plant pathogenic fungi utilize a series of complex infection structures, in particular the appressorium, to gain entry to and colonize plant tissue. As a consequence of the accumulation of huge quantities of glycerol in the cell the appressorium generates immense intracellular turgor pressure allowing the penetration peg of the appressorium to penetrate the leaf cuticle. Autophagic processes are ubiquitous in eukaryotic cells and facilitate the bulk degradation of macromolecules and organelles. The study of autophagic processes has been extended from the model yeast Saccharomyces cerevisiae to pathogenic fungi such as the rice blast fungus Magnaporthe oryzae. Significantly, null mutants for the expression of M. oryzae autophagy gene homologs lose their pathogenicity for infection of host plants. Clarification of the functions and network of interactions between the proteins expressed by M. oryzae autophagy genes will lead to a better understanding of the role of autophagy in fungal pathogenesis and help in the development of new strategies for disease control.

Keywords: :

• appressorium
• autophagy
• autophagy-related (ATG) genes
• infection structure
• pathogenicity

Introduction

The life span of cellular macromolecules is much shorter than that of the cell itself. Ideally cells should have mechanisms to “recycle” any reusable materials in order to reduce their dependence on the extracellular environment. The vacuole in fungi is a major site where recycling is executed. The materials catabolized in this acidic compartment may be of endogenous or exogenous origin having been delivered by autophagic processes or endocytosis, respectively. Thus, autophagic processes result in the transport of cellular contents to vacuoles for degradation, thereby contributing to the homeostatic balance between synthesis and degradation within cells. It is the action of the resident complement of acid hydrolases within vacuoles that degrades the material delivered into the vacuolar lumen. The molecular “building blocks” so recovered are reused in biosynthetic pathways, providing a mechanism by which cells can respond to nutrient stress effectively by reallocating resources for use in the most essential cellular processes.^1,2

The yeast Saccharomyces cerevisiae, itself a member of the fungi, is arguably the best-studied model system by which to investigate the molecular mechanisms of autophagy.^2-4 With respect to mechanism, there are several distinct autophagic processes which occur in yeast cells: macroautophagy, microautophagy, cytoplasm-to-vacuole targeting (Cvt)⁵ and vacuolar import and degradation (Vid).^6,7 The latter two processes are specific to yeast, not occurring in mammalian cells. By contrast chaperone-mediated autophagy (CMA), which occurs in mammals, does not occur in yeast.^8,9

An improved understanding of the molecular mechanisms of plant fungal pathogenesis and their interface with autophagic processes will ultimately lead to better control of plant fungal diseases. Diverse plant organs produce different obstacles to infection by potential fungal pathogens, and therefore successful pathogens have evolved specific strategies, especially infection structures, that are able to break through host plant roots, stems, leaves, flowers or other special tissues. The appressorium is an example of such a special infection structure, which has been well studied in Magnaporthe oryzae and Colletotrichum spp.^10-12 Appressorium development is initiated by a fungal spore landing on, and attaching to, the plant cuticle surface. From the attached spore a germ tube is produced that grows until it recognizes a suitable site and receives environmental stimuli to activate the intracellular molecular machinery for appressorium formation and penetration. Substantial mechanical force is generated in a mature appressorium and used to deliver a penetration peg through the plant surface. Fungal infection-related development is a starvation-induced process. Nitrogen limitation has been proposed as a key signal for activating the expression of virulence genes in plant pathogens.¹³ The noteworthy discovery of the role of macroautophagy (hereafter autophagy), following its induction by nutrient starvation, in the development of turgor in the appressorium represents a milestone emphasizing the importance of, and functional connections between, autophagy and the formation of infection structures in plant pathogenic fungi.^14-16 It appears likely that autophagy is a pivotal process in determining the outcome of the penetration stages by plant pathogenic fungi.¹⁷ Here we will review current knowledge of autophagy in plant pathogenic fungi and the functional links between autophagy and plant fungal pathogenesis.

Infection Structures of Pathogenic Fungi

Plant organs have evolved various hurdles to frustrate infection by potential pathogenic fungi, and in response pathogens have therefore developed some powerful strategies to gain entry to the underlying tissues of prospective host plants. Unlike plant bacterial pathogens, which usually overcome the barriers presented by plant surfaces through the utilization of natural openings, such as stomata, or existing wounds, many plant fungal pathogens are able to actively initiate breaching of the plant cuticle. It should be noted that some plant fungal pathogens, for example, rust fungi, can enter host cells via the natural opening of stomata. Thus, during evolution fungal pathogens have evolved specialized hyphal structures that are active during host invasion, including the infection cushion, the appressorium, and the haustorium.^18-21 The formation and development of such structures are of vital importance to the success of the fungal infection and the eventual development of disease symptoms by the host plant. These processes are controlled by several genes and influenced by environmental factors such as temperature, pH, etc.^12,22-25 In addition, various cell wall depolymerases at the infection site are also needed by some pathogenic fungi.²⁶ In particular, appressoria and associated melanized cell wall structures are critical for penetration of host cells.

The development of an appressorium occurs as follows: A conidium of a plant fungal pathogen, for example M. oryzae or Colletotrium spp, will land on and attach to the leaf surface of its host plant, then germinate under suitable environmental conditions. The germ tube may sense a broad variety of signals derived from topographic information on leaf surfaces, physical stimuli and chemical substances produced by its host during growth.²⁰ These environmental signals will be accepted by membrane receptors of the germ tube or hyphal tip, transmitted into the fungal cell to activate appressorium development and morphogenesis through the action of certain molecular machinery, such as the adjustment of different components of the cytoskeleton, and anionic and/or electric changes mediated by mechanosensitive ion channels.²⁰ A typical life cycle of M. oryzae, which activates appressorium development during infection, is shown in Figure 1.

Figure 1. Life cycle of Magnaporthe oryzae. Conidia germinate and develop a specialized infection structure, the appressorium. The structure produces a penetration peg, which will lead to invasive growth in and between host cells. This infection cycle is destroyed in the autophagy-blocked mutants.

Morphology and Physiology of the Appresorium

The defining feature of the appressorium is that it can generate huge turgor and directly rupture the plant cuticle, or even an artificial membrane that is not able to be degraded by biological enzymes, suggesting that a significant mechanical force is required for fungal infection.^10,27-29 Appressoria are usually clearly visible as dark differentiated structures at the tips of the germ tubes,²¹ although there are instances in which they are difficult to distinguish morphologically. Under some circumstances, they may form from the hyphae. When an appressorium is separated from the germ tube by a septum, as in many Colletotrichum spp, M. oryzae, and Uromyces spp, the germ tube and spore frequently undergo autophagy, and nuclear degradation, and are devoid of cytoplasm.^15,16,30 The nucleus undergoes at least one round of mitosis and migrates with the cytoplasm into the appressorium.^16,30 Verification of the role of autophagy is provided by analysis of the mgatg8∆ mutant, which is blocked in autophagy and forms nonfunctional appressoria.¹⁶ Transfer of conidial cytoplasm into the developing appressoria is delayed in the autophagic null mutants.¹⁵

For many fungi that directly penetrate the plant cuticle, the cell wall of the appressorium undergoes extensive modifications.^10,31 In Colletotrichum spp and M. oryzae, the cell wall of the appressorium becomes thicker, multilayered, and highly melanized, but at the contact surface with the plant, the cell wall is usually less modified and thinner.³¹ The dome-shaped appressorium with highly differentiated cell wall structures is rich in chitin and contains a melanin layer on the inner side of the cell wall. The melanin layer prevents efflux of cellular solutes and ensures the appressorium is able to accumulate substantial turgor for fungal penetration at a later time.^31-33 The importance of melanin in infection structures is emphasized by the observation that knockout of genes in the melanin biosynthesis pathway leads to loss of the function of appressorium-mediated penetration and pathogenicity in M. oryzae.^31,34,35 Turgor pressure in the appressoria of melanin-deficient mutants is about 30–70% of that measured in fully melanized appressoria.¹⁰

The force exerted during initial penetration of a single appressorium is substantial. The average penetration force of the C. graminicola appressorium reaches 17 μN, ranging from 8 to 25 μN. The pressure within the appressorium has been calculated to be 5.35 MPa, ~50 times atmospheric pressure. By comparison, the pressure of the M. oryzae appressorium is estimated to reach only a more modest 8 MPa corresponding to a force of 8 μN.³⁶ In C. graminicola, exertion of force typically starts 100 min after appressorium formation and reaches a steady level after 300 min. This level of force does not change for several hours.³⁷ These forces should be enough to breach the cuticle and epidermal cell wall of most monocotylous or dicotylous plants.

The enormous turgor pressure that develops in the appressorium can be ascribed to the intracellular accumulation of soluble carbohydrates such as glycerol, erythritol, arabinitol, mannitol and glucose.³⁸ In M. oryzae, it has been shown that turgor in the appressorium is a consequence of the accumulation of huge quantities of glycerol in a central vacuole of the cell.²¹ Formation of a central vacuole filled with glycerol is not observed in M. oryzae mutants deleted for one of the MoATG2, MoATG4, MoATG5, MoATG9 or MoATG18 genes. Thus, it was confirmed that autophagy is essential for turgor generation in the appressoria of M. oryzae (Fig. 2).^14,15,39

Figure 2. Model for the relationships between autophagy, appressorium turgor and the MAPK pathway in M. oryzae. In this model, the PMK1 MAPK pathway controls the mobilization of glycogen and lipid reserves to the developing appressorium. Glycerol accumulation is regulated by autophagy, transfer of cytoplasm, and degradation of glycogen, lipid, organelles, or other materials. Appressorium turgor is generated and hydrostatic turgor is translated into the force required for cuticle penetration. The OSM1 and MPS1 MAPK pathways may regulate development of turgor pressure and application of turgor pressure to penetration, respectively. Dashed lines indicate the conjectural relationships.

More generally, the deletion of ATG genes leads to morphological or physiological changes in the fungal life cycle, accompanied by reduced pathogenicity for the host. Though different fungal species may have similar lifestyles and pathogenic strategies, the control of basic cellular processes during fungal infection can be substantially different among species.⁴⁰ Thus far, the influence of individual ATG genes has been studied in relatively few pathogenic fungi, foremost among these is M. oryzae. In a moatg1∆ mutant, the appressorial turgor pressure is much lower than that in the wild type; only ~2.8% of the appressoria form long penetration pegs 24 h after inoculation, compared with ~49.7% in the wild type.¹⁴ In a moatg4∆ mutant, both the conidial germination and appressorium formation rates are reduced. However, as time elapses the differences become slight, such that 24 h after inoculation, there is no significant difference in the rate of appressorium formation between the mutant and wild type.⁴⁰ Nevertheless, the mutant is more osmotically sensitive when exposed to 2 M glycerol for 10 min.⁴⁰ Transfer of conidial cytoplasm into the developing appressorium is delayed, but not blocked in the moatg5∆ mutant, which loses its pathogenicity.¹⁵

Signaling Mechanisms Contributing to the Infection Structure

Mitogen-activated protein kinases (MAPKs) regulate many cellular processes in eukaryotic cells in response to extracellular stimuli.^41-43 To date, five MAPK signaling pathways have been studied in S. cerevisiae.⁴⁴ Two MAPKs, Slt2 and Hog1, are required for mitophagy in S. cerevisiae.⁴⁵ The yeast Slt2 and several upstream components of its signal transduction pathway are necessary for pexophagy, but not for pexophagosome formation or other nonselective and selective forms of autophagy.⁴⁶ Three MAPK genes have been identified in M. oryzae: PMK1 (pathogenicity MAP kinase), MPS1 (MAP kinase for penetration and sporulation) and OSM1 (osmoregulation MAP kinase), that are homologous to S. cerevisiae FUS3 and KSS1, SLT2 and HOG1, respectively.^47-49 Targeted gene disruption of M. oryzae PMK1 has revealed that the cognate MAPK signaling pathway is important for regulation of formation of the infection structure. Thus pmk1∆ mutants lose their capability for forming appressoria on some surfaces and are unable to rupture the rice cuticle to cause rice blast lesions.⁴⁷ Two upstream regulatory components, of this pathway, a MAPK kinase (MAPKK) Mst7, and a MAPKK kinase (MAPKKK) Mst11, have been identified in M. oryzae.⁵⁰ Functional analyses confirmed that both are indispensable for appressorium formation with Mst11 acting upstream of Mst7. Mst7 is responsible for Pmk1 phosphorylation and interacts physically with Pmk1 during formation of the appressorium by means of a conserved MAPK-docking site on Mst7; deletion of this site abolishes appressorium formation.⁵¹ There is a sterile α-motif (SAM) domain found in the Mst11 MAPKKK, facilitating interaction with the SAM-containing protein Mst50, which acts as the adaptor or scaffold protein linking Mst11 with Mst7. Deletion of MST50 leads to inability to form appressoria and thus loss of pathogenicity.⁵² While it is clear that this signaling pathway contributes to the development of the infection structure, as yet there are no data available that specifically link it with autophagy in plant pathogenic fungi (Fig. 2).

Molecular Mechanisms of Autophagy in Pathogenic Fungi

Progress in understanding the mechanistic basis of autophagy has been greatly facilitated by the discovery of the ATG genes and characterization of the encoded proteins. Studies in S. cervisiae have identified 36 ATG genes required for autophagic processes.³ Most of these genes are also found in the filamentous fungi. The homologs of proteins involved in autophagic processes in several plant pathogenic fungi, including Phaeosphaeria nodorum (Pn), Sclerotinia sclerotiorum (Ss), Puccinia graminis (Pg), Botrytis cinerea (Bc), M. oryzae (Mo), and U. maydis (Um), are given in Table 1. These proteins are ubiquitous in plant pathogenic fungi. However, only a limited number of functional studies concerning autophagic processes have been performed in pathogenic fungi.

Table 1. Homologs of proteins involved in autophagy in plant pathogenic fungi

	Ss^a	Pg	Pn	Um	Bc	Mo
ATGs
Atg1	SS1T_11231	PGTT_11663	SNOT_03344	UM06363^c	BC1T_05432	MGG_06393T0^c
Atg2	SS1T_00880	PGTT_11479	SNOT_11489	UM01166	BC1T_09010	MGG_05998T0 ^c
Atg3	SS1T_11342	PGTT_04384	SNOT_13776	UM00169	BC1T_08793	MGG_02959T0 ^c
Atg4	SS1T_10962	-	SNOT_14613	UM05142	BC1T_11320	MGG_03580T0 ^c
Atg5	-	PGTT_04287	SNOT_03476	-	-	MGG_09262 T0 ^c
Vps30	SS1T_12254	PGTT_16671	SNOT_04250	UM04265	BC1T_01844	MGG_03694T0^c
Atg7	SS1T_05017	PGTT_07890	SNOT_14934	UM04880	BC1T_09811	MGG_07297T0^c
Atg8	SS1T_01602	-	SNOT_01275	UM05567^c	BC1T_02467	MGG_01062T1^c
Atg9	SS1T_07955	PGTT_03991	SNOT_03152	UM04380	BC1T_15751	MGG_09559T0^c
Atg10	-	-	-	-	-	-
Atg11	SS1T_01192	PGTT_14853	SNOT_08377	UM03341	BC1T_03242	MGG_04486T0^c
Atg12	-	PGTT_07878	SNOT_06553	UM03577	-	MGG_00598T0^c
Atg13	SS1T_13492	-	SNOT_06190	-	-	MGG_00454T0^b,c
Atg14	-	-	-	-	-	MGG_03698T0^b,c
Atg15	SS1T_00233	PGTT_06305	SNOT_02073	UM00271	BC1T_07555	MGG_12828T0^c
Atg16	-	-	-	-	-	MGG_05255T0^c
Atg17	SS1T_06363	-	-	-	BC1T_11270	MGG_07667T0^c
Atg18	SS1T_09586	PGTT_10397	SNOT_14035	UM04932	BC1T_12821	MGG_03139T0^c
Atg19	-	-	-	-	-	-
Atg20	SS1T_13719	PGTT_16707	SNOT_13865	UM04532	-	MGG_12832T0^c
Atg21	SS1T_09586	PGTT_10397	SNOT_14035	UM04932	BC1T_12821	MGG_03139T0^c
Atg22	SS1T_10007	PGTT_15681	SNOT_14513	UM02974	BC1T_07239	MGG_09904T0^c
Atg23	-	-	-	-	BC1T_09995	MGG_10579T0
Snx4	SS1T_06028	PGTT_15875	SNOT_04668	UM03920	BC1T_01500	MGG_03638T0^c
Atg25	-	-	-	-	-	-
Atg26	SS1T_07979	PGTT_05939	SNOT_03311	UM02058	BC1T_12424	MGG_03459T0^c
Atg27	-	-	-	-	-	MGG_02386T0^b,c
Atg28	-	-	-	-	-	MGG_08061T0^b,c
Atg29	-	-	-	-	-	MGG_02790T0^b,c
Atg30	-	-	-	-	-	-
Cis1	-	-	-	-	-	-
Atg32	-	-	-	-	-	-
Atg33	SS1T_04629	-	-	-	BC1T_00142	MGG_12948T0
Non-ATGs
Irs4	SS1T_03343	-	SNOT_10527	-	BC1T_09584	MGG_05273T0
Tax4	SS1T_03343	-	SNOT_10527	-	BC1T_09584	MGG_05273T0
Tor1	SS1T_12947	PGTT_03326	SNOT_08245	UM03216	BC1T_11880	MGG_15156T0
Tor2	SS1T_12947	PGTT_03326	SNOT_08245	UM03216	BC1T_11880	MGG_15156T0
Ccz1	-	-	-	-	-	-
Mon1	SS1T_02970	PGTT_03945	SNOT_12092	-	BC1T_02219	MGG_05755T0^c
Sec17	SS1T_05321	PGTT_08200	SNOT_07668	UM00160	BC1T_12159	MGG_01121T0
Sec18	SS1T_12225	PGTT_00084	SNOT_03195	UM01284	BC1T_01819	MGG_02418T0
Vam3	-	-	-	UM02338	-	-
Vam7	SS1T_07165	-	-	-	BC1T_10233	MGG_05428 T0^b,c
Vti1	SS1T_14360	PGTT_16392	SNOT_07627	UM05645	BC1T_15556	MGG_01124T0
Vps15	SS1T_08574	PGTT_15709	SNOT_14089	UM01033	BC1T_07434	MGG_06100T0
Vps34	SS1T_11324	PGTT_16800	SNOT_12889	UM00453	BC1T_08774	MGG_03069T0
Ypt7	SS1T_03128	PGTT_14043	SNOT_09938	UM05511	BC1T_07591	MGG_08144T0^c
Vam6	SS1T_08189	PGTT_00217	SNOT_15601	-	BC1T_13204	MGG_01054T0
Vps16	SS1T_00520	PGTT_13771	SNOT_10369	UM04073	BC1T_02005	MGG_05256T0
Vps33	SS1T_10000	PGTT_15043	SNOT_04405	UM02612	BC1T_07246	MGG_09914T0
Vps41	SS1T_12002	PGTT_06044	SNOT_12006	UM01312	BC1T_14017	MGG_03313T0
Pep3	SS1T_00608	PGTT_07854	SNOT_09114	UM05292	BC1T_06250	MGG_06325T0
Pep5	SS1T_13462	PGTT_17544	SNOT_06922	UM03820	BC1T_06444	MGG_03283T0

^a Homologs were identified using the yeast sequences derived from SGD (except Atg25, Atg28 and Atg30, which are from NCBI). Other fungal sequences were derived from the Broad Institute (http://www.broadinstitute.org). Sequences of the best hit with an E-value < 10⁻⁶ are shown as candidates. Ss, Sclerotinia sclerotiorum; Pg, Puccinia graminis; Pn, Phaeosphaeria nodorum; Um, Ustilago maydis; Bc, Botrytis cinerea; Mo, Magnaporthe oryzae. ^bThese proteins do not meet the criteria, but are identified by direct experimentation. ^cThese proteins have been identified by direct experimentation.

1. Magnaporthe oryzae

Autophagy contributes to development and differentiation in M. oryzae.^14,15 Evidence for the involvement of 24 autophagy-related genes in the pathogenicity of M. oryzae has been provided by genome-wide functional analysis.⁵³ Nonselective autophagy plays a key role in the pathogenicity of M. oryzae. Deletion of any of the genes necessary for nonselective autophagy (e.g MoATG1 to MoATG10, MoATG12, MoATG15, MoATG16, and MoATG18) renders the mutant fungi unable to cause rice blast. In contrast, deletion of any of the genes necessary for selective autophagy does not affect pathogenicity. Thus, the moatg11∆, moatg24∆, moatg26∆, moatg27∆, moatg28∆, and moatg29∆ mutants are not severely impaired in their ability to cause disease.

We now consider studies concerning autophagic processes in M. oryzae in the context of established findings in the yeast S. cerevisiae.

A. Nitrogen and TOR signaling

A Ser/Thr protein kinase, TOR (target of rapamycin), acts as a central regulator of autophagy and mediates the response of cells to nutrient starvation. The function of TOR is mediated by two distinct multiprotein complexes, TOR complex (TORC) 1 and 2. It is TORC1 which is thought to couple growth signals to cellular metabolism.⁵⁴ This complex is particularly inhibited by rapamycin, and therefore rapamycin is an activator of macroautophagy that is commonly used in yeast (and other species since the structure and function of TORC1 is conserved in evolution). In yeast, TORC1 functions as a switch between macroautophagy and the Cvt pathway by altering phosphorylation of Atg13.⁵⁵ Under normal growth conditions TORC1 directly phosphorylates Atg13, which results in reduced Atg1 kinase activity. Inhibition of TOR by rapamycin causes a nutrient stress response, including inhibition of translation initiation, arrest in the G₁ phase of the cell cycle, glycogen accumulation, downregulation of glycolysis, and autophagy.⁵⁶ Nitrogen is a particularly important nutrient in TOR signaling. In M. oryzae, MPG1 encoding a hydrophobin required for pathogenicity is strongly upregulated under conditions of nitrogen limitation.⁵⁷ Invasive growth is affected by nitrogen sources. Many genes identified in screens for nitrogen starvation-induced transcripts are also upregulated during plant infection.⁵⁸ Thus, nitrogen source seems to act as a metabolic switch to trigger expression of infection-related genes in the rice blast fungus. However, the role of TOR in nutrient regulation of fungal virulence on plants has not been examined so far.

The Snf1 kinase complex belongs to a highly conserved family of serine/threonine protein kinases and is involved in glycogen biosynthesis, lipid biosynthesis, regulation of general stress responses and autophagy in yeast.⁵⁹ Snf1 and Pho85 act as positive and negative regulators of autophagy, respectively, via Atg1 and Atg13.⁶⁰ Homologs of each of the Snf1 kinase complex subunits (Snf1, Snf4, Sip1, Sip2, and Gal83) have been found in M. oryzae. MoSNF1, a homolog of yeast SNF1, contributes to growth of the colony, shape and germination of the conidia, formation of the appressoria and pathogenicity of the rice blast fungus.⁶¹ However, there are no reports about the relationship between the Snf1 kinase complex and autophagy.

B. The Atg1-Atg13 kinase complex

The yeast proteins Atg1, Atg13 and Atg17 form an Atg1 protein kinase complex.⁶² In M. oryzae, the MoATG1 gene encodes a Ser/Thr protein kinase, which is highly conserved among eukaryotes. Disruption of the MoATG1 gene significantly affects fungal ability to survive under starvation conditions, with the mutant showing reduced conidiation, conidial germination, lipid turnover, and appressorium turgor. As a result, the moatg1∆ mutant loses its penetration ability and pathogenicity for rice and barley.¹⁴ Notably, in each of the moatg13∆ and moatg17∆ mutants, autophagy was not completely blocked and pathogenicity was not affected (unpublished data). Furthermore, MoAtg9 localized to multiple punctate structures under control of the MoAtg1 protein, and these multiple dots tended to pool into one central structure within/near the vacuoles of the moatg1∆ mutant. However, such a central structure is not detected in proximity of the vacuoles of moatg13∆ and moatg17∆ mutants.⁶³ These findings are not consistent with observations in yeast.^64,65 It can be speculated that MoAtg13 and/or MoAtg17 has lost its function in autophagy in M. oryzae and therefore the Atg1-Atg13 complex might not be preserved across all fungi during evolution. Alternatively there may be functional homologs of ATG13 and ATG17 encoded in the M. oryzae genome, which are yet to be identified as such. Moreover, a gene encoding Atg31, which is also a constituent of the Atg1 kinase complex in yeast,⁶⁶ has not been identified in M. oryzae. Therefore, it is possible that novel specific molecular mechanisms, distinct from those in yeast, exist for the induction of autophagy in M. oryzae.

C. Ubiquitin-like conjugation systems

The yeast autophagy pathway requires two ubiquitin-like conjugation systems, Atg8 and Atg12–Atg5-Atg16, for autophagosome maturation.⁶⁷ It was reported that moatg8∆ mutants are unable to undergo conidial cell collapse and have lost their pathogenicity to the plant.¹⁶ MoAtg8 is also involved in the regulation of glycogen metabolism during conidiogenesis.^53,68 Analysis of the cellular localization pattern and flux of Atg8 is a reliable marker for autophagy in a variety of organisms.⁶⁹ Expression of an EGFP-MoAtg8 fusion is sufficient to complement the moatg8∆ mutant phenotypes, providing evidence that it is functional within M. oryzae, and therefore a reliable marker for analysis of the cellular pattern of autophagy.⁵³ Additionally, RFP-MoAtg8-PE has been established as a reliable marker for autophagosomes and autophagic vacuoles in aerial hyphae and conidiophores.⁶⁸

Atg8 undergoes C-terminal cleavage by the cysteine protease activity of Atg4 to expose a glycine residue in yeast.⁷⁰ Yeast complementation revealed that MoAtg4 can functionally complement the defects of a yeast ATG4 deletion mutant. The direct interaction between MoAtg4 and MoAtg8 is detected in both yeast two-hybrid and bimolecular fluorescence complementation (BiFC) assays. BiFC data indicated also that the MoAtg4-MoAtg8 interaction is enhanced only during nitrogen starvation. Recombinant MoAtg4 harboring a substitution of Cys to Ser at position 206 (MoAtg4^C206S) is unable to cleave MoAtg8 in vitro, either in the absence or presence of DTT. These data suggest that Cys206 is part of the active site of MoAtg4⁴⁰ and show that MoAtg4 is a conserved cysteine protease required for autophagy.

In yeast the Atg12–Atg5 complex interacts with Atg16 to form a complex of 2:2:2 stoichiometry based on homodimerization of Atg16,⁷¹ which localizes to the phagophore assembly site (PAS).^72,73 The Atg12–Atg5-Atg16 complex is not associated with the completed autophagosome,⁷⁴ dissociating just prior to, or after, completion of the fusion event that finalizes autophagosome formation.³ The MoATG5 gene encodes a conserved domain that is essential for autophagosome formation in M. oryzae. Deletion of the MoATG5 gene shows defective autophagy, shortened life span of aerial hyphae, reduced conidiation and perithecia formation, delayed germination of conidia and slower transfer of conidial cytoplasm, and reduction in appressorial turgor. As a result, the moatg5∆ mutant loses the ability to penetrate and infect rice and barley.¹⁵ Laser excitation epifluorescence microscopy showed that the GFP-MoAtg5 protein is distributed evenly throughout cells accompanied by the presence of some sharply defined puncta in the cytoplasm of conidia, mycelia and appressoria. MoAtg8 localizes to the PAS as reported previously.⁶³ When a DsRed2-MoAtg8 fusion protein is expressed in an moatg5∆ mutant also expressing GFP-MoAtg5, then the DsRed2-MoAtg8 protein colocalizes with GFP-MoAtg5 puncta. The GFP-MoAtg5 and DsRed2-MoAtg8 fusion proteins are colocalized in the perivacuolar region in M. oryzae (unpublished data) consistent with at least transient localization to the PAS.

D. Atg9 cycling system

An Atg9 complex that includes Atg2 and Atg18 also seems to be involved in the autophagy of pathogenic fungi. In yeast, Atg9 is a transmembrane protein localized not only at the PAS, but also at cytosolic punctate compartments whose identity remains uncertain,⁷⁵ but may be tubulovesicular clusters.⁷⁶ The recycling of Atg9 from the PAS to its storage sites was demonstrated to be blocked in an Atg1-deficient mutant, and loss of Atg2, Atg13, Atg14 or Atg18 appears to produce a similar phenotype with regard to their effect on Atg9 retrograde movement in yeast.⁶⁴ When GFP-MoAtg9 is expressed in wild-type conidial and appressoria cells of M. oryzae, multiple green puncta are detected. The distribution pattern of EGFP-MgAtg9 is changed in these cells in moatg1∆, moatg2∆ or moatg18∆ mutants with prominent large green puncta that colocalize with DsRed2-MoAtg8 being detected. By contrast, the expression pattern of the same fusion proteins is not changed in a moatg13∆ mutant.⁶³ It was concluded that cycling of MgAtg9 through multiple sites (as indicated by colocalization studies) to its storage pools in the cytosol of M. oryzae depends on MoATG1, MoATG2, and MoATG18, but not MoATG13. Loss of MoATG1, MoATG2 and MoATG18 might prevent MoATG9 from leaving the PAS-like structures in M. oryzae. Deletion of MoATG2 or MoATG18 also affects conidiation, the production of appressorium turgor, pathogenicity for rice and sexual reproductive ability.¹⁴

E. PtdIns3K complex I

The PtdIns3K complex I, which consists of Vps30/Atg6, Atg14, Vps34 (the only PtdIns3K in yeast) and Vps15, plays a pivotal role at the PAS in yeast.^77,78 Most of the core proteins in this complex are fully conserved in plant pathogenic fungi (Table 1). Intriguingly, the gene for one component of the complex, ATG14, remains to be definitively identified. A sequence having very weak similarity to yeast ATG14 has been identified in M. oryzae by performing a re-analysis of NCBI protein databases. This putative homolog of ATG14 contains a conserved Cys-rich motif at its N terminus.⁷⁹ Disruption of the putative ATG14 in M. oryzae produced a highly similar phenotype to the moatg1∆ mutant (unpublished data).

F. SNAREs and HOPs

In yeast, docking and fusion of autophagosomes with the vacuolar membrane requires the SNARE proteins Vam3 and Vti1 (found on the vacuolar membrane), Vam7 and Ykt6 (found on the autophagosome), NSF, SNAP, Sec17, Sec18 and Sec19, the Rab protein Ypt7, and members of the class C Vps/HOPS complex.⁸⁰ Components of this machinery also mediate other intracellular vesicle fusion events, an essential cellular process of eukaryotic cells.⁸¹ MoSec22 and MoVam7, the orthologs of yeast SNARE proteins Sec22 and Vam7, can also be demonstrated to have functions in vacuole assembly, which underlie the growth, conidiation, appressorium formation, and pathogenicity of M. oryzae.^82,83 Similarly, MoMon1 and MoYpt7 are the orthologs of the yeast proteins Mon1 and Ypt7. The phenotypes of momon1∆ and moypt7∆ mutants share common features with the mosec22∆ mutant; they could not develop appressoria and had lost the ability to infect plants (unpublished data). Clearly, components of docking and fusion play key roles in vacuole formation and membrane trafficking that is linked to fungal pathogenicity.

2. Colletotrichum spp

In the cucumber anthracnose fungus C. orbiculare (syn. C. lagenarium) peroxisome degradation (pexophagy) occurs when it infects host plants. A random insertional mutagenesis screening was performed to identify genes involved in the pathogenesis of this fungus. A homolog of P. pastoris ATG26, which encodes a sterol glucosyltransferase that enhances pexophagy in this methylotrophic yeast, was isolated. The coatg26∆ mutant develops appressoria, but exhibits a specific defect in the subsequent host invasion step. Analysis using a GFP-tagged fusion protein suggests that CoAtg26 is localized at putative PAS sites. It has been proposed that CoAtg26 is involved in the regulation of the efficiency of pexophagy in the mature appressoria.⁸⁴ These data show that selective autophagy might be required during plant infection by C. orbiculare.

Unlike the coatg26∆ mutant, which is able to form appressoria, the coatg8∆ mutant is defective in the entire autophagic pathway and cannot form normal appressoria in the earlier steps of morphogenesis. By contrast, the moatg8∆ mutant of M. oryzae retains the ability to develop appressoria. This shows the diversity of function of some autophagy proteins and of the autophagy process as a whole in different plant pathogenic fungi.

Clk1, a homolog of yeast ATG1, was identified by a random insertional mutagenesis screen in C. lindemuthianum. For the clk1∆ mutant, very few lesions are produced on the intact leaves. However, a marked increase in the number of lesions is clearly visible 7 d after inoculation on wounded leaves, compared with intact leaves. However, even where the number of lesions on veins is much higher on wounded leaves, their extension is still limited and no maceration is observed. Clk1 transcripts are present both in pure cultures of the fungus and during the time-course of host infection, indicating the gene may be constitutively expressed. Furthermore, it was suggested that the signal transduction pathway involving the Clk1 protein kinase may be regulated either at a post-transcriptional level or through its substrates.⁸⁵

3. Fusarium graminearum

Autophagy plays critical roles in the pathogenicity of the Fusarium head blight fungus Fusarium graminearum. The FgATG15 gene encodes a lipolytic enzyme that plays an important role in the development of the fungus. Deletion of FgATG15 leads to defects in conidiogenesis, conidial shapes, germination, growth rate, and aerial hyphae formation. Under nutrient starvation conditions, the wild type degrades stored lipid droplets while the mutant loses this ability. Wheat head blight is severely attenuated in the fgatg15∆ mutant. Disease severity in the mutant is 9%, compared with 92% and 88% in the wild type and ectopic strains, respectively. The concentration of the mycotoxin deoxynivalenol (DON, a type B trichothecene, an epoxy-sesquiterpenoid) produced by the fgatg15∆ mutant is 55% less than that of the wild-type strain. These results imply that FgATG15 is involved in numerous developmental processes and could be exploited as an antifungal target.⁸⁶

Recently, a fgatg8∆ mutant was generated by gene replacement and found to be unable to form autophagic compartments. The fgatg8∆ mutant shows no formation of fruiting bodies (perithecia), reduced conidia production and collapse of its aerial mycelium after a few days in culture. The mutant contains lipid droplets indicative of nitrogen starvation and/or an inability to use storage lipids, suggesting autophagy-dependent lipid utilization (lipophagy) in this fungus. The capacity to reallocate nutrients and support nonassimilating fungal structures is reduced in the mutant. Although the ability to infect barley and wheat remains normal, the mutant is unable to spread from spikelet to spikelet in wheat. Collectively the data were interpreted to mean that autophagy provides a mechanism for supplying nutrients to nonassimilating structures necessary for growth and is important for plant colonization.⁸⁷

4. Ustilago maydis

Autophagy genes are important for the development and virulence of the corn smut fungus U. maydis. In U. maydis, umatg8∆ or umatg1∆ deletion mutants prevent the vacuolar accumulation of autophagic bodies and dramatically reduce survival under carbon-starvation conditions. Deletion of UmATG8 affects the budding of haploid sporidia, gall formation and teliospore production in ears of mature maize.⁸⁸ The umatg1∆ deletion mutants have phenotypic similarity to umatg8∆ deletion mutants, but with lower severity, such that umatg1∆ mutants are only slightly less pathogenic than the wild type, and teliospore (the thick-walled resting spore of U. maydis) production is not affected.⁸⁸ Moreover, in the double-deletion mutant, plant gall induction is completely suppressed.⁸⁸

5. Other filamentous fungi

Autophagy has been monitored in the filamentous fungus Podospora anserina. In this organism, autophagy can be induced by starvation or rapamycin treatment, or by heterokaryon incompatibility genes.^89,90 Analysis of paatg1∆, paatg8∆, and pspA∆ mutants revealed that autophagy is essential for formation of aerial hyphae and for female organ differentiation, and is involved in spore germination.^91,92

In the opportunistic mold pathogen, Aspergillus fumigatus, starvation-associated foraging has been studied. Hyphal plugs were transferred from rich medium onto starvation medium, thereby forcing the organism to use autophagy to fuel any further growth. When conidia from a strain expressing GFP-AfAtg8 are incubated in Aspergillus minimal medium the accumulation of autophagic bodies in vacuoles is observed by laser confocal microscopy.⁹³

In Aspergillus oryzae, a aoatg8∆ mutant was constructed and autophagy monitored by the expression of DsRed2-AoAtg8 and EGFP-AoAtg8 fusion proteins.⁹⁴ Under normal growth conditions the fusion proteins are localized in dot PAS-like structures, whereas starvation or rapamycin treatment cause their accumulation in vacuoles. Cytosolically expressed DsRed2 (not fused to AoAtg8) is also taken up into vacuoles under starvation conditions, or during the differentiation of conidiophores and conidial germination. The aoatg8∆ mutant does not form aerial hyphae and conidia, and DsRed2 is not localized in vacuoles under starvation conditions.⁹⁴ Subsequently aoatg4∆, aoatg13∆ and aoatg15∆ mutants were examined by following EGFP-AoAtg8 fluorescence.⁹⁵ In the aoatg13∆ mutant only limited accumulation of EGFP-AoAtg8 at the PAS, or in autophagosomes and vacuoles was observed. In the aoatg4∆ mutant no vacuolar uptake is detected and only PAS-like structures are detected, whereas in the aoatg15∆ mutant autophagic bodies accumulate in vacuoles. Conidiation in the aoatg4∆ and aoatg15∆ mutants is not detected, and it is decreased in the aoatg13∆ mutant compared with the wild-type strain. The aoatg15∆ mutant also displays a marked reduction in development of aerial hyphae. Thus, it is suggested that autophagy functions in both the differentiation of aerial hyphae and conidial germination in A. oryzae.⁹⁴

Autophagy plays key roles in fruiting-body development of the homothallic ascomycete Sordaria macrospora.⁹⁶ SmATG7 is markedly upregulated under amino acid starvation conditions and during late stages of sexual development. SmATG7 is essential for fungal viability, and autophagy is disturbed in Smatg7-RNAi mutants of S. macrospore, which produce aberrant fruiting bodies. These data provide a starting point for probing the diverse functions of autophagy in a sophisticated microbial genetic system.⁹⁶

Tricherdoma reesei (teleomorph Hypocrea jecorina), a soil-borne, green-spored ascomycete fungus has the ability to secrete large amounts of cellulolytic enzymes (cellulases and hemicellulases). The functions of TrAtg5, the homolog of yeast Atg5, were studied in a tratg5∆ mutant generated using targeted gene disruption. The mutant is blocked in autophagy and produces very few hyphal branches, with abnormal conidiophores and no more than two conidia on single abnormal conidiophores. By contrast, conidia aggregate into fascicles and pustules with plump conidiophores in the wild-type fungus. In addition, the tratg5∆ mutant shows reduced conidiation and sensitivity to nutrient materials.⁹⁷

Penicillium chrysogenum, is used for industrial production of penicillin (PEN) a β-lactam antibiotic. The penicillin biosynthesis pathway takes place in the cytosol and in peroxisomes in this fungus. Deletion of the P. chrysogenum ATG1 gene impairs autophagy and causes sporulation defects as in other filamentous fungi. Extensive autophagy-related degradation of cytosolic components and peroxisomes are observed to occur in vacuolated, late hyphal elements of P. chrysogenum. Cell degeneration is delayed and the levels of PEN biosynthesis enzymes are increased in the Pcatg1 null mutant. Collectively the data demonstrate that impairment of autophagy in P. chrysogenum leads to significantly increased production levels of PEN.⁹⁸

Candida albicans is considered a commensal organism of humans, colonizing the oral cavity, gastrointestinal, and reproductive tracts. However, when host defenses are compromised, C. albicans can transform into a tissue invasive pathogen. The role played by autophagy in facilitating asymptomatic host colonization, persistence, and transition of C. albicans into its pathogenic form has not been fully explored. Resistance to nitrogen starvation has been studied in two mutants, caatg9∆ and cavps11∆. While the former shows defects in autophagosome formation, the latter is blocked in vacuolar fusion. The caatg9∆ mutant survives within and kills a mouse macrophage-like cell line as efficiently as control strains, which suggests that autophagy plays little or no role in C. albicans differentiation, or during its interaction with mammalian host cells.⁹⁹

Cryptococcus neoformans is a yeast-like fungus that causes a lethal meningoencephalitis in a broad spectrum of immunocompromised patients. Survival of the fungus within the hostile and nutrient-deprived environments of the host has recently been shown to depend on the induction of autophagy.¹⁰⁰ PtdIns3K signaling of autophagy is required for starvation tolerance and virulence of this fungus. A cnatg8∆ mutant demonstrates markedly attenuated virulence in a mouse model of infection.¹⁰¹

Concluding Remarks

Much has been learned about the morphological, physiological and biochemical characteristics of appressorium development and function during recent years. Autophagy is necessary for the formation of conidia and appressoria and for normal development and pathogenicity of Magnaporthe or Colletotrichum spp. It is likely that many factors influence the control of autophagy during pathogenesis. A problem in identifying ATG genes and their related functions in filamentous fungi is that functionally redundant genes exist in many fungal genomes. The significance of multiple gene copies in these fungi remains to be established.

The importance of autophagy for plant pathogenic fungi has a fundamental difference with its importance for yeasts because of the special penetration structure utilized in plant fungal infection. There are still many questions that need to be answered even concerning the autophagy machinery itself in pathogenic fungi. How might autophagy contribute to the development of turgor in the appressorium? Is the degradation of lipids and glycogen important for this process? What signaling pathways control autophagy in plant pathogenic fungi? Does the selective autophagy of organelles such as mitochondria, peroxisomes and endoplasmic reticulum contribute to the infectivity of plant pathogenic fungi?

Acknowledgment

This study was supported by grants (No. 30925029 and 31000077) from the National Natural Science Foundation of China to Fu-Cheng Lin and Xiao-Hong Liu, and by a grant (No. LQ12C14003) from the Zhejiang Provincial Natural Science Foundation of China to Fei Xu.