- Top of page
- Materials and Methods
- Supporting Information
In eukaryotes, glycosylphosphatidylinositol (GPI) anchoring is a conserved mode of covalent attachment of numerous proteins to cell membranes (Orlean & Menon, 2007). The model yeast Saccharomyces cerevisiae has been proposed to contain two groups of GPI-anchored proteins according to their final location (De Sampaïo et al., 1999). In the main group, the GPI anchor is involved in a transglycosylation reaction which cross-links the protein to cell wall β-glucans, whereas the second minor group includes plasma membrane-resident proteins. GPI-anchored proteins must firstly acquire a GPI anchor as a posttranslational modification in order to be addressed from their site of synthesis to their ultimate destination (De Sampaïo et al., 1999). The pathway responsible for the synthesis and attachment of the GPI anchor to proteins is complex and plays a critical role in the proper targeting, transport and function of all GPI-anchored proteins. The complete GPI anchor structure is fully preassembled in a multistep pathway that sequentially adds the various GPI components. As suggested by the presence of an evolutionary conserved core region in GPI anchors, the biosynthetic machinery that assembles GPI anchors is also conserved among eukaryotes and consists of at least 10 reaction steps (Orlean & Menon, 2007). In all cases, GPI anchor biosynthesis is initiated by the transfer of N-acetylglucosamine (GlcNAc) from UDP-GlcNAc to phosphatidylinositol (PI). This reaction is catalysed by an enzymatic complex called GPI-N-acetylglucosaminyltransferase (GPI-GnT), which is unusually elaborate compared to other glycosyltransferases. The human GPI-GnT consists of at least seven known subunits – PIG-A, PIG-C, PIG-H, GPI1, PIG-P, PIG-Y and DPM2 (Kamitani et al., 1993; Miyata et al., 1993; Inoue et al., 1996; Watanabe et al., 2000; Tiede et al., 2001; Murakami et al., 2005) – with the first six subunits having structural and functional counterparts in S. cerevisiae, termed GPI3, GPI2, GPI15, GPI1, GPI19 and ERI1, respectively (Leidich et al., 1995; Schönbächler et al., 1995; Leidich & Orlean, 1996; Yan et al., 2001; Sobering et al., 2004; Newman et al., 2005). Of these subunits, PIG-A/GPI3, PIG-C/GPI2, PIG-H/GPI15 and PIG-P/GPI19 are essential for GPI-GnT, because cells with mutations in these genes are completely deficient in the surface targeting of GPI-anchored proteins.
The phytopathogenic Ascomycete Leptosphaeria maculans is the causal agent of stem canker of Brassica species, the most damaging disease on oilseed Brassica such as B. napus (Fitt et al., 2006). This hemibiotrophic fungus develops a complex life cycle, alternating saprophytic, biotrophic and necrotrophic phases. These traits are representative of infection strategies of fungi belonging to the class Dothideomycetes and also include other traits such as endophytic intercellular colonization of plant tissues and the lack of development of specialized infection structures (for a review, see Rouxel & Balesdent, 2005). Despite major economic losses in the main growing areas of the world (Fitt et al., 2006), the molecular determinism of L. maculans pathogenicity towards its host plant is currently poorly understood, with only three factors required for pathogenicity identified to date. Two of these factors, ICL1 and THIOL, are involved in lipid metabolism, whereas the third one has no putative predicted function (Idnurm & Howlett, 2002, 2003; Elliott & Howlett, 2006). In addition, two genes conferring avirulence on resistant genotypes of B. napus have been recently cloned, AvrLm1 and AvrLm6 (Gout et al., 2006; Fudal et al., 2007). Avirulence genes are suggested to participate in virulence or fitness of the pathogen as ‘effectors’ (Huang et al., 2006).
Insertional mutagenesis is a powerful tool for uncovering new pathogenicity genes in fungi, with Agrobacterium tumefaciens-mediated transformation (ATMT) being the tool of choice for such a purpose (Michielse et al., 2005; Elliott & Howlett, 2006; Blaise et al., 2007). A large collection of L. maculans random insertional transformants was generated via ATMT to identify novel pathogenicity factors (Blaise et al., 2007). A total of 53 transformants reproducibly affected in their virulence, and showing various altered phenotypes, were recovered from a subset of 1388 ATMT transformants (Blaise et al., 2007). Following formal genetic analyses, only 50% of the ‘loss of pathogenicity’ mutants were actually tagged by the T-DNA (i.e. were mutants where the T-DNA insertion was responsible for the altered phenotype).
Here, we report on the phenotypic and molecular characterization of one of the L. maculans nonpathogenic mutants, m20, previously shown by formal genetics to be a tagged mutant (Blaise et al., 2007). Our data show that the mutant exhibited a significantly reduced growth rate and an altered morphology in vitro. Following inoculation of oilseed rape cotyledons, it was blocked during the invasive growth phase of plant tissues. The T-DNA integration occurred in the overlapping promoter region of two head-to-tail genes, leading to a complex deregulation of their expression. Complementation experiments further identified one of the two genes as responsible for the mutant phenotype. The corresponding gene encoded an endoplasmic reticulum (ER)-localized protein, which is the putative functional homologue of the human PIG-H and S. cerevisiae GPI15 proteins, involved in GPI anchor biosynthesis. This study thus highlights the intimate connections between GPI anchoring, cell wall integrity and pathogenicity in L. maculans.
- Top of page
- Materials and Methods
- Supporting Information
In this paper, we describe identification of the L. maculans orthologue of the yeast gpi15 gene and human PIG-H, termed Lmgpi15, thus providing the first report on a functional characterization of a PIG-H/GPI15 orthologue in a filamentous fungus. We further show that a deregulation of its expression has drastic effects on fungal morphogenesis, cell wall integrity and pathogenicity. The GPI anchor biosynthesis pathway has already been demonstrated to be involved in overall morphogenesis and virulence of human pathogenic fungi such as in Candida albicans and Aspergillus fumigatus (Richard et al., 2002; Li et al., 2007). Here, we report for the first time that a GPI anchor biosynthetic component-encoding gene is also instrumental in determining pathogenicity to plants, consistent with the known importance of GPI-anchored proteins in fungal virulence.
The homology of LmGPI15 with the human PIG-H and the yeast GPI15 was suggested here via bioinformatics analyses, cell biology experiments and functional studies. Bioinformatics searches for homologues and putative function were initially uninformative, with only part of a C-terminal region of LmGPI15 showing limited, but significant, homology with a domain present in GPI15 and PIG-H. However, the weak sequence similarity between organisms seems to be the rule for this gene, with 26% identity at the amino acid level between LmGPI15 and GPI15 being consistent with the low 20% identity between PIG-H and GPI15. Even when comparing LmGPI15 with homologous proteins in more closely related filamentous fungi, sequence identities only range between 38 and 46% for Neurospora crassa and Aspergillus terreus homologues, respectively, and reach 54% with the closely related species S. nodorum. This low sequence conservation probably explains why efforts to complement the inducible null mutant of gpi15 from S. cerevisiae (Yan et al., 2001) with Lmgpi15 were unsuccessful (P. Orlean, pers. comm.). This lack of cross-complementation has already been observed for some of the genes encoding other components of the GPI anchor biosynthesis pathway in unrelated species. For instance, the S. pombe and C. albicans gpi3 homologue genes were inefficient in restoring the phenotype of the corresponding S. cerevisiae mutant, whereas the fission yeast gpi1 homologue complemented the corresponding S. cerevisiae mutant (P. Orlean, pers. comm.). Whereas sequence conservation was low between all fungal GPI15 homologues, the overall topography of the proteins showed conserved features, including the presence of two transmembrane domains at a similar position and the presence of a N-signal anchor.
The GPI anchor biosynthetic pathway takes place into the ER membrane in all organisms examined so far (Orlean & Menon, 2007). In addition, PIG-A/GPI3, PIG-C/GPI2, PIG-H/GPI15 and PIG-P/GPI19 have been shown to form a complex localized in the ER membrane (Tiede et al., 2000; Newman et al., 2005). Bioinformatics analysis predicted that LmGPI15 is an ER-targeted protein. In addition, the ‘KKXX’ motif identified at the C-terminal region of LmGPI15 is common to the C-terminal region of many membrane proteins known either to reside or to cycle through the ER or the Golgi complex (Jackson et al., 1993). Cell biology experiments aiming at identifying the subcellular location of the LmGPI15 protein substantiated localization of a LmGPI15-GFP fusion protein in the ER. Additional localization of LmGPI15-GFP fluorescence in large vacuoles throughout the fungal cell could be attributed to recycling of ER material (Seiler et al., 1999), but also to a possible more rapid turnover of the fusion protein as a result of overexpression of the LmGPI15 protein under the control of the strong promoter of the Lmpma1 gene (Remy et al., 2008).
Functional studies included biological analysis of the effect of deregulation of expression as well as analysis of the effect of (bio)chemical agents known to interact with cell wall morphogenesis or used as a substrate for GPI anchor biosynthesis. The mutant was affected in its growth rate, as are most of fungal GPI-anchoring mutants thus far characterized (Richard et al., 2002; Sobering et al., 2004; Bowman et al., 2006; Li et al., 2007). Screens designed to isolate cell wall-defective fungal mutants often identify genes encoding either GPI-anchoring components or GPI-anchored proteins (Bowman et al., 2006). In this respect, the product of Lmgpi15 gene is also involved in overall morphogenesis in L. maculans. Similar morphological defects were observed in GPI anchor biosynthesis mutants of N. crassa, C. albicans and A. fumigatus (Richard et al., 2002; Bowman et al., 2006; Li et al., 2007). Additionally, the mutant possesses a weakened cell wall as compared with the WT isolate, since it exhibits an enhanced sensibility to hyperosmotic conditions. Consistent with this hypothesis, the growth defect was relieved by the osmotic stabilizer sorbitol (Sobering et al., 2004). The higher sensitivity to Congo red displayed by m20.4.21 also suggests that its cell wall integrity is affected, since this dye binds to the cell wall glucan network and disturbs the assembly of microfibrils of β-1,3-glucan (Kopeckà & Gabriel, 1992). Based on the assumption that the GPI anchor biosynthesis pathway plays a similar role as in other fungi in cell wall integrity and morphology in L. maculans, we thus propose that Lmgpi15 encodes for the functional orthologue of S. cerevisiae GPI15. Consistently, addition of glucosamine to the culture medium, suggested to increase the intracellular concentration of UDP-N-acetylglucosamine, also suppressed the growth defect of the mutant. At least two GPI anchor biosynthesis mutants of S. cerevisiae have also been shown to complement growth defect with exogenous glucosamine, which assists the mutants in their effort to increase GPI-anchor production (Sobering et al., 2004; Newman et al., 2005).
Whereas intrinsic function of the altered gene could be easily established and validated, it was less clear how the T-DNA insertion affected its expression or regulation, and how the altered expression resulted in pathogenicity defects. The T-DNA was inserted in a short bidirectional promoter, which substantiates the postulate that most of the integrations take place in regulatory regions rather than within coding sequences in the genome of L. maculans (Blaise et al., 2007), as previously established for yeast, fungi and plants (Bundock et al., 2002; Michielse et al., 2005; Pan et al., 2005; Choi et al., 2007). In this respect, ATMT is a powerful tool to study the incidence of lethal genes on biological traits. gpi15 in S. cerevisiae, and more generally the GPI anchor biosynthesis pathway in S. cerevisiae and N. crassa are required for viability (Leidich et al., 1995; Yan et al., 2001; Bowman et al., 2006). In this study, both unsuccessful attempts to create a null mutant for Lmgpi15 by homologous recombination and the very low rate of silenced transformants recovered following the Lmgpi15 RNA silencing experiment (usually more efficient in L. maculans (Fudal et al., 2007; Remy et al., 2008)) could indicate that Lmgpi15 is also essential for L. maculans viability. The T-DNA integration event led to a complex deregulation of Lmgpi15 expression, thus suggesting that the T-DNA integration has modified or suppressed some still uncharacterized regulatory motifs present in the overlapping promoter region, such as those responding to environmental signals. However, promoter sequence analysis did not allow us to detect such previously characterized motifs. The apparent overexpression detected in vitro could be attributed to the very low level of expression in these conditions, thus causing ‘background noise’ in the real-time PCR analysis. In addition, overexpression of the gene under control of the strong Lmpma1 promoter did not lead to pathogenicity defects, as complementation of the mutant by the LmGPI15-GFP fusion protein restores full pathogenicity. During the infection process, the Lmgpi15 expression is significantly reduced in such a way that there is no induction of expression during the first days of plant infection. This suggests a special need for induced GPI anchor biosynthesis (either a special target or a generally high amount of GPI anchoring activity) during the first stages of tissue colonization.
On these bases, two nonexclusive effects can explain altered pathogenicity to oilseed rape in the mutant. First, slow progress of m20.4.21 and silenced transformants in host tissues, along with an altered cell wall, can lead to an increased sensitivity to plant defence reactions. Second, a fungus-specific algorithm, based on sequence characteristics of known GPI-anchored proteins from various Ascomycetes, recently revealed 66, 104, 33 and 97 putative GPI-anchored proteins in whole genome sequences of S. cerevisiae, C. albicans, S. pombe and N. crassa, respectively (De Groot et al., 2003). In these species, GPI-anchored proteins are mainly involved in cell wall biosynthesis and remodelling (Yin et al., 2005). The cell wall perturbations generated by their inactivation can in turn lead to a loss of cell wall-associated enzymatic activities. In addition, a subset of GPI-anchored proteins of Ascomycete pathogenic fungi have been reported to play a role in virulence, both in human pathogenic fungi, such as A. fumigatus (Mouyna et al., 2005) and C. albicans (Richard et al., 2002), and in phytopathogenic fungi, such as Magnaporthe grisea (Ahn et al., 2004) and Fusarium oxysporum (Caracuel et al., 2005). As GPI anchors are involved in cell wall targeting of a number of proteins, including pathogenicity determinants, failure of GPI anchoring could thus be responsible for reduced virulence as a result of a reduction of available pathogenicity determinants (De Sampaïo et al., 1999). We thus suggest that pathogenicity defects of m20 and silenced transformants can, in part, be the result of a reduced amount of cell surface or secreted proteins implicated in pathogenicity of L. maculans, in addition to cell wall and growth defects. The mutant m20 thus constitutes a unique resource for proteomic approaches aiming at comparing cell wall and secreted protein profiles between m20 and WT isolate, therefore allowing a critical assessment of GPI-anchored proteins as pathogenicity or virulence factors in L. maculans and other pathogenic fungi.