Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. 2006 Apr 25; 103(17): 6647–6652.
Published online 2006 Apr 13. doi:  10.1073/pnas.0601951103
PMCID: PMC1458935

A collagenous protective coat enables Metarhizium anisopliae to evade insect immune responses


The ubiquitous fungal pathogen Metarhizium anisopliae kills a wide range of insects. Host hemocytes can recognize and ingest its conidia, but this capacity is lost on production of hyphal bodies. We show that the unusual ability of hyphal bodies to avoid detection depends on a gene (Mcl1) that is expressed within 20 min of the pathogen contacting hemolymph. A mutant disrupted in Mcl1 is rapidly attacked by hemocytes and shows a corresponding reduction of virulence to Manduca sexta. Mcl1 encodes a three domain protein comprising a hydrophilic, negatively charged N-terminal region with 14 cysteine residues, a central region comprising tandem repeats (GXY) characteristic of collagenous domains, and a C-terminal region that includes a glycosylphosphatidylinositol-dependent cell wall attachment site. Immunofluorescence assay showed that hyphal bodies are covered by the N-terminal domains of MCL1. The collagen domain became antibody accessible after treatment with DTT, suggesting that the N termini are linked by interchain disulfide bonds and are presented on the cell surface by extended collagenous fibers. Studies with staining reagents and hemocyte monolayers showed that MCL1 functions as an antiadhesive protective coat because it masks antigenic structural components of the cell wall such as β-glucans, and because its hydrophilic negatively charged nature makes it unattractive to hemocytes. A survey of 54 fungal genomes revealed that seven other species have proteins with collagenous domains suggesting that MCL1 is a member of a patchily distributed gene family.

Keywords: collagen, like protein, virulence, cell wall proteins, fungal pathogen

As the most abundant and diverse land animals, insects have attracted a variety of pathogens, including viruses and bacteria. However, most insect disease is caused by fungi, and their impact on insect populations demonstrates the potential of microbial control of insects of medical and agronomic interest (13). However, the slow speed of kill and inconsistent results of biologicals in general compared with chemicals has deterred development. An understanding of fungal-induced immune responses would identify the insect defenses and fungal pathogenicity factors that overcome them, and hence identify fungal virulence determinants that could be manipulated to accelerate host death in a biological control scenario (4).

Unlike bacteria and viruses that need to be ingested to cause disease, fungi penetrate directly through the cuticle. About 1% of known fungal species are capable of breaching the cuticle of at least some insect species. These are then fought by the insect innate immune responses based on both cellular (5) and humoral (6) mechanisms. An immune response starts with recognition of pathogen-associated molecular patterns (PAMPs), and many of the molecules and receptors involved are homologous in insects and vertebrates (6). For both groups, PAMPs include β-1,3-glucans from fungal cell walls (7) as well as nonspecific mechanisms such as surface charge and wettability (8). Various pathways of the immune system then become activated (6), leading to the destruction of the pathogen and/or its removal by cellular reactions such as phagocytosis or encapsulation in many layers of hemocytes. To cause infection, the fungus has to avoid, subvert, or circumvent this system. Given the response of the human immune system to fungal β-glucans, it has been speculated that pathogens may avoid immune recognition by camouflaging or modifying their β-glucan (9). Indeed, Paracoccidioides brasiliensis displays a transition from β-glucan to α-glucan in the cell wall upon infection of the lung (10). Insect pathogens are also reported to engage in several “hiding” tactics that include changes in cell wall composition that eliminate cell surface components associated with non-self recognition, thus allowing hyphal bodies to circulate freely in the hemolymph (11, 12). However, the molecular basis of these changes has not been determined and it is not clear the extent to which they reflect de novo synthesis of proteins, or morphological and topological rearrangement of cell surface components.

Recent EST and microarray studies have provided abundant evidence that sets of functionally related genes are coordinately induced or repressed by Metarhizium anisopliae in response to host related stimuli (1315). Multiple mechanisms specifically involved in acclimatizing to hemolymph isolated from the lepidopteran model insect Manduca sexta include dramatic changes in lipid composition, the accumulation of solutes that increase internal osmotic pressure, and up-regulation of nonoxidative respiratory pathways. However, the most highly expressed gene in hemolymph (5.6% of all ESTs) encoded a collagenous protein of unknown function (13). In this study, we show that transcripts of Mcl1 (for Metarhizium collagen-like protein) can be detected within 20 min of the pathogen contacting hemolymph. Mcl1 encodes a protein with a hydrophilic N-terminal domain that is presented on the cell surface within 30–45 min of induction by an extended glycosylated collagenous region. MCL1 functions as an antiadhesive protective coat against phagocytosis and encapsulation because its hydrophilic negatively charged nature is unattractive to hemocytes and because it masks the immunogenic β-1,3-glucan cell wall structural components. Because hyphal bodies (short hyphal lengths and yeast-like blastospores) represent the principal stage of replication of the fungus within the host insect hemocoel, the inability to clear these cells allows the fungus to more easily establish itself and kill the host.


Analysis of MCL1 from M. anisopliae.

Structural analysis of the predicted MCL1 protein indicates that it is composed of 605 residues (60.4 kDa) that includes an 18-aa secretory sequence and a three domain structure (A, B, and C; Fig. 1A) comprising an N-terminal domain (domain A) predicted to be globular, acidic (pI 4.9), and highly hydrophilic, a central collagenous domain (domain B), and a C-terminal region (domain C) that includes a site for attachment of a glycosylphosphatidylinositol (GPI) anchor deduced with the algorithm of Eisenhaber et al. (16). GPI anchors link to β-1,6-glucans that protrude from the fungal cell wall, suggesting that MCL1 is a component of the external protein layer that is covalently linked to the underlying skeletal layer of the wall (17).

Fig. 1.
A schematic structure of MCL1 (A) and the alignment (clustalw) of MCL1 domain B with collagenous regions from other fungal sequences (B). Up- and down-pointing arrows indicate N-glycosylation sites and cysteine residues, respectively. SP, signal peptide; ...

The great functional versatility of collagens originates from the combinational assembly of other domains with the collagen domain (18). Many collagens, including mammalian collagen type IV that comprises basal membranes, have globular noncollagenous domains (19). However, a search of databases showed no significant matches to the N-terminal domain of MCL1. It contains 14 cysteine residues consistent with multiple intra- and intermolecular bonds. The collagenous domain itself comprises 33 Gly-X-Y copies in which X and Y are frequently Ser, Asn, or Pro. There were six interruptions in the regular Gly-X-Y repeats of the MCL1 protein consisting of two or three residues (Fig. 1B). Such interruptions lead to flexible sites or kinks and are very common in collagens (18). Similar to many bacterial, viral, and invertebrate collagen-like proteins (20), domain B has many (n = 13) consensus N-glycosylation sites. Heavy glycosylation would be expected to increase rigidity of the domain and produce an elongated structure (21). However, the domain lacked the multiple cysteine residues found in many collagens, indicating that it does not form the intermolecular disulfide bridges required for the high tensile strength and functioning of structural fibers.

The collagen content of fungi has not been well established. Therefore, we conducted unfiltered searches of 54 finished and unfinished fungal genomes using the MCL1 collagenous region or (GPP)7 (22). Seven species (Clavisopora lusitania, Candida glabrata, Debaryomyces hansenii, Aspergillus nidulans, Aspergillus fumigatus, Coccidioides posadasii, and Coccidioides immitis) have sequences that include the characteristic Gly-X-Y repeat of collagen domain (Fig. 1B). Most of these sequences also had a three-domain structure, including a hydrophilic 5′ domain containing variable numbers of cysteine residues, a central domain of variable numbers (22–52) of G-X-Y repeats with multiple glycosylation sites, and a 3′ domain that is variable in length (absent in the D. hansenii protein). Proline is a major component of X and Y of most collagens. It comprised 21.2% of the X and 15.2% of the Y residues in MCL1, as compared with mammalian collagens that contain 28.2% Pro at X and 38.4% Pro at Y. The average Pro content of the fungal G-X-Y domains at the Y position is 33.8%, as compared with 12.5% for viruses and 4.2% for bacterial collagens (22). The percentage of Pro residues at the X site varied from zero in C. glabrata to 56% in A. fumigatus (Fig. 6, which is published as supporting information on the PNAS web site).

Induction of MCL1 by Hemolymph Constituents.

We performed RT-PCR analysis of Mcl1 expression by mycelia suspended in different media. Mcl1 transcripts were detected during growth in the hemolymph of a diverse array of insects, consistent with the broad host range of M. anisopliae. However, Mcl1 was not expressed in nutrient-rich artificial media or during starvation conditions, suggesting that it is only involved in pathogenesis (Fig. 2A). A time course demonstrated that Mcl1 transcripts began to appear within 20 min of transfer into M. sexta hemolymph and were still accumulating at 4 h (Fig. 2B).

Fig. 2.
Mcl1 gene induction and protein localization. (A) RT-PCR analysis of Mcl1 expression by wild-type M. anisopliae transferred from Sabouraud dextrose broth (SDB) cultures to minimal medium (MM), fresh SDB, or hemolymph collected from Manduca sexta (MS), ...

For MCL1 to function in avoiding the host immune system, it must be located on the cell surface. This was verified by using an indirect immunofluoresence (IIF) assay with rabbit antibodies (abA) raised against a peptide sequence from the noncollagenous domain A. No fluorescence was detected on conidia or mycelia grown in nutrient broth or minimal medium, but MCL1 was detectable on hyphal tips 30–45 min after induction with hemolymph, and levels of staining increased over several hours (Fig. 2 C and D). The accessibility of the N-terminal domain to abA establishes that it is presented on the surface of the cell. In contrast, the collagenous domain of MCL1 was not available to antibodies (abB) raised against a constituent peptide sequence, unless cells were pretreated with DTT. This finding suggests that domain B is internal to domain A and that disulfide bonds interconnecting the N termini of MCL1 proteins are involved in creating a nonporous barrier. Without exception, IIF of several hundred hyphal bodies isolated at 10-h intervals from the hemolymph of infected insects showed strong, even surface staining with abA (Fig. 2E). However, like hyphae in hemolymph in vitro, they were not recognized by abB.

Characterization of the Posttranslational Modifications of MCL1.

Western blot analysis of extracts of mycelial cell walls, harvested from insect hemolymph in vitro, confirmed that both abA and abB recognize a polydisperse band with an apparent molecular mass ranging as high as 300 kDa (Fig. 2F). Disperse bands are typical of extensive glycosylation so the protein was treated with N-glycosidase F to remove all N-chains. Deglycosylation caused the protein to run as a single sharp 75-kDa band consistent with heavy glycosylation of the collagenous domain, where all but one of the consensus N-glycosylation sites are found. The difference with the molecular mass predicted from the amino acid sequence (60.4 kDa) is presumably due to the GPI tail (≈5 kDa) and O-mannosylation.

Behavior of Wild-Type M. anisopliae and a ΔMcl1 Null Mutant Within Infected Manduca.

To study the role of MCL1, we constructed a strain of M. anisopliae in which the Mcl1 gene was disrupted. No fluorescence with abA was detected on hyphal bodies of the null mutant in hemolymph in vitro or in vivo. Because conidia are of a similar size (≈7 μm) to blastospores but lack MCL1 protein, we injected conidia directly into the hemocoel of larval M. sexta to study the functional significance of their different surface properties. Within 10 min of injection, single conidium from the wild-type and ΔMcl1 strains had attached to hemocytes or become phagocytosed, whereas clumps of conidia were encapsulated, showing that they are readily recognized as foreign. It was not always possible to distinguish by microscopic observation between hemocyte attachment and subsequent phagocytosis, but propagules of M. anisopliae are known to survive phagocytosis and grow within host cells (12, 23). Survival of conidia during their initial interactions with host cells was not dependent on Mcl1, because >90% of both wild-type and mutant ΔMcl1 conidia germinated within 8–10 h. However, hyphae and hyphal bodies produced by ΔMcl1 conidia continued to recruit hemocytes and were repeatedly encapsulated, whereas wild-type hyphae emerged from capsules and budded off hyphal bodies that received little attention from the hemocytes (Fig. 3AD). Thirty hours after injection of conida, capsule diameters averaged 75.6 ± 18.1 μm (n = 82) and 200.3 ± 29.9 μm (n = 73) in insects infected with the wild type and ΔMcl1, respectively. This finding suggests that, during conidial germination, the wild-type can rapidly adapt the composition of the newly formed cell wall in response to the hemolymph environment, with resulting changes in ligands on cell surfaces from those present on its conidia or on the ΔMcl1 mutant. Encapsulation of the ΔMcl1 mutant continued 50 h after injection (Fig. 3D) and only ceased with the manifestation of obvious disease symptoms such as reduced food uptake and softening of the body.

Fig. 3.
Differences in the patterns of infection shown by wild type and ΔMcl1. Manduca larvae were injected with conidia and bled at 10-h intervals. (A) Wild-type germ tubes emerging from encapsulation 30 h after injection. (B) Heavy encapsulation of ...

The ability of ΔMcl1 to survive and cause disease indicates that M. anisopliae has additional mechanisms to cope with immune responses. However, LT50 values from injection assays showed that the ΔMcl1 mutant takes a significantly longer time (P = 0.0012) to kill insects than the wild type (Fig. 4). Insects were also dipped in suspensions of conidia to assay infections through the cuticle. At a high (2 × 107 conidia per ml) dosage, the ΔMcl1 mutant took significantly longer to kill than the wild type (P = 0.0006) (Fig. 4). At 8 × 106 conidia per ml, ΔMcl1 failed to achieve 50% mortality before pupation and, at lower (<5 × 106) dosages, mortality fell to <10%. In contrast, the wild type achieved >50% larval mortality at 1 × 106 conidia per ml, suggesting that it would be much more likely than ΔMcl1 to cause high mortality under field conditions, where concentrations seldom exceed 106 conidia per g of soil (2).

Fig. 4.
Kinetics of insect survivorship in bioassays. (A) Mortality of Manduca larvae after topical application with 2 × 107 conidia per ml suspensions of wild-type or ΔMcl1 mutant strains (control insects were dipped in water). LT50 values were ...

Hemocyte Monolayer Assays.

Given the differences in how hemocytes in infected insects behave to wild type and ΔMcl1, we also investigated the effects of deleting Mcl1 on cell surface properties and hemocyte responses in vitro. Cell surface hydrophobicity was measured by using a microsphere adhesion assay. Conidia from the wild type and ΔMcl1 were similarly hydrophobic with 15.37 ± 1.21 and 15.14 ± 1.59 beads attached per cell, respectively. Despite conidia and blastospores being very similar in size, this represents >3-fold more microspheres than adhered to blastospores of the wild type (3.28 ± 0.16 beads per spore) and ΔMcl1 (4.94 ± 0.48 beads per spore). The ≈30% fewer microspheres adhering to wild-type blastospores demonstrates that MCL1 produces a small but significant (t = 6.81, P = 0.0103) overall increase in cell surface wettability. This finding suggests that MCL1 disruption had unmasked components of blastospores that were also hydrophilic, but less so than MCL1. Hydrophobic Dynabeads exhibit a much greater attraction to insect hemocytes than do hydrophilic beads (t = 21.31, P = 0.0011) (Fig. 5), so loss of hydrophilic MCL1 would be expected to increase attractiveness to hemocytes. Indeed, hemocyte monolayer assays agreed with the infection studies in showing that ΔMcl1 blastospores are recognized at >3-fold higher efficiency than are wild-type blastospores (Fig. 5). We investigated whether this could be due to an ability by hemocytes to distinguish between different cell surface charges or degrees of hydrophilicity and/or to the exposure of PAMPs underlying the MCL1 layer.

Fig. 5.
Recognition of blastospores, conidia, and beads by Manduca hemocytes in vitro. Monolayers were exposed to wild-type (WT) or ΔMcl1 (MU) M. anisopliae cells treated with collagenase (Coll), proteinase K (Pro K), lyticase (Lyt), DTT, poly(l-lysine) ...

The surface charge of outer cell surfaces was assessed using FITC-labeled poly(l-lysine). Only a faint fluorescence was observed on conidia, but blastospores of the wild type and the ΔMcl1 mutant were negatively charged (Fig. 3E). Blastospores preincubated with unconjugated poly(l-lysine) before treatment with the FITC probe did not fluoresce, indicating that negative surface charges were neutralized. However, they remained hydrophilic (3.59 ± 0.23 microspheres per spore), suggesting that the hydrophilicity of blastospores is not due to their electronegativity. To confirm this possibility, cell surface electronegativity was also reduced by treating wild-type blastospores with dicyclohexylcarbodiimide and ethylenediamine to replace negatively charged carboxyl groups with positively charged ammonium groups (24). The derivitized wild-type cells still showed a lower degree of binding to microspheres (3.87 ± 0.39 beads per spore) than did ΔMcl1 blastospores (t = 4.37, P = 0.0243). The difference with untreated wild-type cells was not significant (t = 1.26, P > 0.05), and they were not more readily recognized than untreated cells (Fig. 5). Thus, the differential hemocyte response elicited by the ΔMcl1 mutant is not a nonspecific reaction to charge.

The surface exposure of β-glucans was measured by the degree of binding of Calcoflour white. Strong fluorescence was observed on the ΔMcl1 blastospores, but the wild-type blastospores were barely visible using the same exposure time (Fig. 3 F and G). This finding demonstrates that MCL1 renders PAMPs, such as β-1,3-glucans, inaccessible to arriving hemocytes. Treatment of wild-type cells with collagenase or proteinase K prevented labeling with antibodies to MCL1 but produced strong fluorescence with Calcofluor and increased their recognition to the levels shown for ΔMcl1 (Fig. 5), confirming that a protein component of wild-type cells blocks hemocyte responses. Additionally, treatment of ΔMcl1 cells with lyticase (a β-1,3-glucanase) greatly reduced labeling with Calcofluor and phagocytosis compared with untreated cells (t = 15.67, P = 0.0051), confirming that hyphal bodies are recognized on contact with β-1,3-glucans.


In the course of infecting a host, pathogens are presented with a wide array of host environments. Cell surface proteins and secreted hydrolases will define the interactions between host and pathogen and together are likely to have a profound impact on the infection outcome (25). Because of the functional adaptations of GPI proteins to localization at the cell surface, they likely comprise the majority of Candida spp. surface proteins involved in human disease (26). The agglutinin-like protein (ALS) cell surface adhesins of C. albicans are particularly informative in light of their similarities and differences with MCL1. Although lacking a collagenous domain, ALS proteins have a relatively nonglycosylated cysteine rich N-terminal domain that is displayed on the cell surface by the remainder of the protein that is extended because of its heavy glycosylation (27). The central region is so heavily glycosylated that, like MCL1, ALS proteins migrate at three to five times their predicted molecular weights (28). However, unlike MCL1, they have hydrophobic N-terminal domains that facilitate adhesion to host tissues (28, 29). There is an abundance of literature identifying the hydrophobic effect as the driving force for the initial adhesion of pathogens to host surfaces that establishes infection (reviewed in ref. 30). However, the MCL1 protein has the opposite function of producing a nonadherent cell; to achieve this, it provides a hydrophilic antiencapsulation coat around the fungus that prevents recognition by hemocytes. Therefore, it appears analogous in function to the extracellular polysaccharide capsule produced by Cryptococcus neoformans to avoid recognition by phagocytes (31).

A major part of the PAMP, the MCL1 protein blocks is β-glucans in the cell wall, but it also contributes to properties of the cell surface that reduce attraction to insect hemocytes. These include physiochemical properties such as wettability, and it is also pertinent that MCL1 lacks the tripeptide Arg-Gly-Asp (RGD) sequence as hemocytes possess an RGD-dependent adhesion mechanism (32). Rapid encapsulation of hydrophobic conidia despite their failure to bind Calcofluor suggests that hemocytes respond to several criteria including nonspecific mechanisms that can independently induce an immune response. The wettability of blastospore surfaces has also been noted in other insect pathogens (11) and could be consequential for infection processes in many ways, besides just avoiding a direct hemocyte response. The negatively charged hydrophilic surface is likely to prevent clumping of cells and attachment to host surfaces thus facilitating dispersal through the insect. Hydrophilic surfaces are also much more resistant to non specific adsorption of proteins (33), lectins, and other opsonins (30), minimizing the possibility of opsonization by β-1,3-glucan binding proteins. The MCL1 coat probably protects against several β-glucan binding proteins in M. sexta. A laminarin-binding M. sexta lectin promotes encapsulation (34), and cell wall β-glucans activate the phenoloxidase cascade (35), so it is significant that encapsulated ΔMcl1 propagules were frequently melanized (Fig. 3B). Because host humoral responses are also triggered by PAMPs (7), MCL1 will likely be involved in avoiding these as well. However, both wild-type and ΔMcl1 mutant hyphae survive challenge with the antimicrobial cercopin A at a level (50 μM) sufficient to kill saprobic fungi (C.W. and R.J.S.L., unpublished data). This finding suggests that M. anisopliae has evolved multiple strategies that allow it to grow unhindered by the insect immune system.

A further unique feature of this research was the identification of a collagenous Gly-X-Y domain in a fungus. Collagens are the most abundant proteins in the vertebrate body and are essential structural elements that evolutionary models trace to fibrillar forms in early animals (18, 36). The presence of collagen in the cell walls of the human pathogen C. albicans had been inferred by immunological analysis (37), but we found no collagen sequence in the published genome of C. albicans. The fungus Ustilago violaceum was also inferred to contain collagenous protein (38), but no gene has been identified. In contrast, the characteristic triple-helical structure formed by Gly-X-Y motifs has been found in bacterial collagens (39). Demonstrating that MCL1 forms the trimers characteristic of collagen fibers is beyond the scope of this paper. However, the MCL1 sequence has all of the features expected of a collagen and apparently none that would preclude a multimeric helicoidal structure. The high proline content of MCL1 domain B that is characteristic of most collagens will prevent secondary structure. In addition, the heavy glycosylation also characteristic of collagens is likely to confer an extended conformation on that portion of the molecule (21). This will project the relatively glycosylation free N-terminal domain into the extracellular milieu. Thus, the experimental support for the surface location of domain A also makes intuitive sense. An interesting potential model for the structure of protein motifs in MCL1 is provided by the collectins, a family of animal lectins that recognize PAMPs. These also have a cysteine-rich N-terminal domain and a heavily glycosylated collagenous region with interruptions (kinks). The basic functional unit is a trimer, but the kinks provide flexibility, enabling the trimeric subunits to cover more area. The trimers are stabilized and assembled into larger oligomers via the cysteine residues (40). The possibility of interchain disulfide bridges in MCL1 is consistent with the ability of DTT to permeabilize the MCL1 sheath around the fungus.

The presence of a collagenous domain is not associated with a specific lifestyle, i.e., they are present in saprobic fungi as well as pathogens. The patchy distribution of collagen-like proteins among bacteria and viruses is supposed to derive from horizontal gene transfer from multicellular animals (22). The apparently similar sparse distribution of the collagenous domain in fungi could also be explained by repeated independent instances of gene loss (41), or by convergent evolution, particularly as alignment of collagenous regions indicates high sequence divergence (Fig. 1B). However, it is interesting to note that three of four C. galabrata collagenous proteins (XP_447814, XP_447815 and XP_447816) have >50% overall similarity and locate in tandem within 12 kb on chromosome J, indicative of recent duplication events and diverging functions. Given that most fungi lack homologs for MCL1 or other proteins with a collagenous domain, they are evidently expendable in terms of maintaining the normal organization of fungal hyphae. The roles they play in the fungi that possess them will probably therefore have to be addressed on a case-by-case basis.

Materials and Methods

For further details, see Figs. 7–9, which are published as supporting information on the PNAS web site.

Gene Cloning and Deletion.

The cDNA of Mcl1 was fully sequenced, and the genomic DNA was acquired by using a primer walking kit (Seegene) from M. anisopliae strain ARSEF2575. The procedures for construction of gene knockout plasmids and fungal transformation are provided in Fig. 8.

Cell Wall Protein Isolation and Western Blot Analysis.

Thirty-six-hour Sabouraud dextrose broth (SDB) cultures were washed three times with sterile distilled water and then transferred either to minimum medium (MM) or to isolated Manduca hemolymph as described (13) for up to 24 h. Fungal cell wall proteins were extracted as before (17). In-solution deglycosylation was conducted by using a GlycoProfile II kit (Sigma). Two predicted antigenic regions, A1 (PGPNASPDQIKKHRD; residues 59–73 of the N-terminal domain) and B1 (NGKPGSGNNGANGSN; residues 421–435 of the collagenous domain) were synthesized and conjugated with keyhole limpet hemocyanin. The antibodies were raised in New Zealand White rabbits (Sigma) and designated as abA and abB. Western blot analyses were conducted as described (42).

Indirect Immunofluorescence (IIF).

Fungal cells grown in hemolymph in vitro or in vivo were prepared for IIF as described (37). Antibodies were diluted 500-fold and FITC-conjugated-goat anti-rabbit Ig G (Sigma) was used for secondary labeling. Control samples of cells were treated as above but minus either the primary or secondary antibodies.

Insect Bioassay.

Virulence of wild-type and ΔMcl1 mutant conidia was assayed against newly molted fifth-instar larvae of M. sexta as described (4). Thus, conidia were applied topically by immersion of larvae or by injecting the rearmost proleg with 10 μl of an aqueous suspension containing 5 × 105 spores per ml. Each treatment was replicated three times with 10 insects per replicate, and the experiments were repeated twice. Mortality was recorded every 12 h, and estimated lethal time values for 50% mortality (LT50) were used to compare speed of kill between strains with the t test as before (4). Additional infected insects were bled at 10-h intervals for microscopic observation of fungal development within the insect haemocoel.

Hemocyte Monolayer Assay.

Hemolymph was collected in prechilled saline buffer (43) from day 2 fifth-instar larvae and applied as a suspension of 2 × 106 cells per ml onto glass coverslips (10-mm diameter). The coverslips were incubated in Grace's medium at 28°C for 2 h and then washed twice with 0.5 ml Grace's medium (43). Wild-type and ΔMcl1 blastospores (harvested from fungal cultures grown in hemolymph in vitro for 72 h) or conidia were washed twice with PBS, and 2 × 103 cells were applied to the hemocyte monolayers to assay recognition. In some experiments, fungal cells were fixed in 4% formaldehyde or pretreated for 1 h in PBS containing either DTT, poly(l-lysine), lyticase, proteinase K, or collagenase (Sigma) (the enzymes at 200 μg/ml) before assaying. Hydrophobic (M280) and hydrophilic (M270) Dynabeads (2.8 μm in diameter, Dynal) were used as references to test the effects of nonspecific surface properties on hemocyte responses. After incubation of the monolayer coverslips for 1 h at 28°C, the percentage of test particles recognized by the hemocytes (the number of particles bound or ingested by hemocytes relative to the number added) was determined in five different fields of vision using the 40× objective. Data shown were calculated from 600 or more cells or beads/monolayer/insect and six insects per treatment.

Characterization of Fungal Cell Surface Properties.

A microsphere adhesion assay of cell surface hydrophobicity was conducted by using 0.6-μm latex polystyrene beads (Sigma) (44). The spores were suspended in 0.1 M KNO3 solution (2 × 107 cells per ml, pH 6.5), and the suspensions were mixed with microspheres suspended in the same buffer in a ratio of 20 beads/one spore. Three replicates for each treatment were performed, and a total of 50 spores were counted for each replicate. To modify cell surface charge, blastospores were treated with dicyclohexylcarbodiimide and ethylenediamine. By this method, carbodiimide-activated carboxylate groups are substituted with positively charged ammonium groups from the ethylenediamine (24). FITC-labeled poly(l-lysine) was used to assay surface charge (8). Staining with Calcofluor White was used to measure the exposure of β-glucans on the cell surfaces of wild-type and mutant hyphal bodies (45).


Mycelia from 36-h SDB cultures (0.1 g wet weight) was washed twice with sterile water before transfer into minimal medium, fresh SDB, or hemolymph collected from seven insect species. At different time points, RNA (0.5 μg) was extracted and converted into single-strand cDNA using an anchored oligo(dT) primer (ABgene, Surrey, U.K.). Complementary DNA samples diluted 500-fold were used as template for PCR. Primers designed for the small subunit ribosomal RNA were used as the reference.

Statistical Analysis.

Student's t test was used for the pairwise comparisons of means given in the text. Dunnett's least significance difference multiple comparison method in the program spss (version 11.0.0) was used to compare the different treatments shown in Fig. 5.

Supplementary Material

Supporting Figures:


This work was supported by National Research Initiative of the U.S. Department of Agriculture Cooperative State Research, Education, and Extension Service Grant 2003-353-02-13588.


PAMPpathogen-associated molecular pattern
SDBSabouraud dextrose broth.


Conflict of interest statement: No conflicts declared.

Data deposition: The sequences reported in this paper have been deposited in the GenBank database (accession nos. DQ238488 and DQ238489).


1. Carruthers R. I., Soper R. S. In: Epizootiology of Insect Diseases. Fuxa J. R., Tanada Y., editors. New York: Wiley; 1987. pp. 357–416.
2. Roberts D. W., St. Leger R. J. Adv. Appl. Microbiol. 2004;54:1–70. [PubMed]
3. Blanford S., Chan B. H. K., Jenkins N., Sim D., Turner R. J., Read A. F., Thomas M. B. Science. 2005;308:1638–1641. [PubMed]
4. St. Leger R. J., Joshi L., Bidochka M. J., Roberts D. W. Proc. Natl. Acad. Sci. USA. 1996;93:6349–6354. [PMC free article] [PubMed]
5. Lavine M. D., Strand M. R. Insect Biochem. Mol. Biol. 2002;32:1295–1309. [PubMed]
6. Royet J. Mol. Immunol. 2004;41:1063–1075. [PubMed]
7. Kanost M. R., Jiang H., Yu X. Q. Immunol. Rev. 2004;198:97–105. [PubMed]
8. Lackie A. M. J. Cell Sci. 1983;63:181–190. [PubMed]
9. Brown G. D., Gordon S. Immunity. 2003;19:311–315. [PubMed]
10. Borges-Walmsley M. I., Chen D., Shu X., Walmsley A. R. Trends Microbiol. 2002;10:80–87. [PubMed]
11. Pendland J. C., Boucias D. G. Eur. J. Cell Biol. 1998;75:118–127. [PubMed]
12. Gillespie J. P., Bailey A. M., Cobb B., Vilcinskas A. Arch. Insect Biochem. Physiol. 2000;44:49–68. [PubMed]
13. Wang C.-S., Hu G., St. Leger R. J. Fungal Genet. Biol. 2005;42:704–718. [PubMed]
14. Wang C.-S., St. Leger R. J. Eukaryot. Cell. 2005;4:937–947. [PMC free article] [PubMed]
15. Freimoser F. M., Hu G., St. Leger R. J. Microbiology. 2005;151:361–371. [PubMed]
16. Eisenhaber B., Schneider G., Wildpaner M., Eisenhaber F. J. Mol. Biol. 2004;337:243–253. [PubMed]
17. Frieman M. B., Cormack B. P. Mol. Microbiol. 2003;50:883–896. [PubMed]
18. Engel J. Science. 1997;277:1785–1786. [PubMed]
19. Paulsson M. Crit. Rev. Biochem. Mol. Biol. 1992;27:93–127. [PubMed]
20. Sylvestre P., Couture-Tosi E., Mock M. Mol. Microbiol. 2002;45:169–178. [PubMed]
21. Jentoft N. Trends Biochem. Sci. 1990;15:291–294. [PubMed]
22. Rasmussen M., Jacobsson M., Bjorck L. J. Biol. Chem. 2003;278:32313–32316. [PubMed]
23. Vilcinskas A., Matha V., Gotz P. J. Insect Physiol. 1997;43:475–483.
24. Goldberg S., Doyle R. J., Rosenberg M. J. Bacteriol. 1990;172:5650–5654. [PMC free article] [PubMed]
25. Charnley A. K. Adv. Bot. Res. 2003;40:241–321.
26. Frieman M. B., McCaffery J. M., Cormack B. P. Mol. Microbiol. 2002;46:479–492. [PubMed]
27. Zhao X., Oh S. H., Yeater K. M., Hoyer L. L. Microbiology. 2005;151:1619–1630. [PMC free article] [PubMed]
28. Hoyer L. L. Trends Microbiol. 2001;9:176–180. [PubMed]
29. Sheppard D. C., Yeaman M. R., Welch W. H., Phan Q. T., Fu Y., Ibrahim A. S., Filler S. G., Zhang M., Waring A. J., Edwards J. E., Jr. J. Biol. Chem. 2004;279:30480–30489. [PubMed]
30. Doyle R. J. Microbes Infect. 2000;2:391–400. [PubMed]
31. Kozel T. R. Trends Microbiol. 1995;3:295–299. [PubMed]
32. Lavine M. D., Strand M. R. Insect Mol. Biol. 2003;12:441–452. [PubMed]
33. Rabinow B. E., Ding Y. S., Qin C., McHalsky M. L., Schneider J. H., Ashline K. A., Shelbourn T. L., Albrecht R. M. J. Biomater. Sci. Polym. Ed. 1994;6:91–109. [PubMed]
34. Yu X. Q., Tracy M. E., Ling E., Scholz F. R., Trenczek T. Insect Biochem. Mol. Biol. 2005;35:285–295. [PubMed]
35. Ma C., Kanost M. R. J. Biol. Chem. 2000;275:7505–7514. [PubMed]
36. Aouacheria A., Cluzel C., Lethias C., Gouy M., Garrone R., Exposito J. Y. J. Biol. Chem. 2004;279:47711–47719. [PubMed]
37. Sepulveda P., Murgui A., Lopez-Ribot J. L., Casanova M., Timoneda J., Martinez J. P. Infect. Immun. 1995;63:2173–2219. [PMC free article] [PubMed]
38. Celerin M., Ray J. M., Schisler N. J., Day A. W., Stetler-Stevenson W. G., Laudenbach D. E. EMBO J. 1996;15:4445–4453. [PMC free article] [PubMed]
39. Xu Y., Keene D. R., Bujnicki J. M., Hook M., Lukomski S. J. Biol. Chem. 2002;277:27312–27318. [PubMed]
40. van de Wetering J. K., van Golde L. M., Batenburg J. J. Eur. J. Biochem. 2004;271:1229–1249. [PubMed]
41. Hu G., St. Leger R. J. J. Evol. Biol. 2004;17:1204–1214. [PubMed]
42. St. Leger R. J., Frank D. C., Roberts D. W., Staples R. C. Eur. J. Biochem. 1992;204:991–1001. [PubMed]
43. Wiesner A., Götz P. J. Insect Physiol. 1993;39:865–876.
44. Hazen K. C., Hazen B. W. J. Microbiol. Methods. 1987;6:289–299.
45. Pistole J. Annu. Rev. Microbiol. 1981;35:85–112. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...