Grosmannia clavigera is a bark beetle-vectored pine pathogen in the mountain pine beetle epidemic in western North America. Grosmannia clavigera colonizes pines despite the trees' massive oleoresin terpenoid defences. We are using a functional genomics approach to identify G. clavigera's mechanisms of adaptation to pine defences.
We annotated the ABC transporters in the G. clavigera genome and generated RNA-seq transcriptomes from G. clavigera grown with a range of terpenes. We functionally characterized GcABC-G1, a pleiotropic drug resistance (PDR) transporter that was highly induced by terpenes, using qRT-PCR, gene knock-out and heterologous expression in yeast.
Deleting GcABC-G1 increased G. clavigera's sensitivity to monoterpenes and delayed development of symptoms in inoculated young lodgepole pine trees. Heterologous expression of GcABC-G1 in yeast increased tolerance to monoterpenes. G. clavigera but not the deletion mutant, can use (+)-limonene as a carbon source. Phylogenetic analysis placed GcABC-G1 outside the ascomycete PDR transporter clades.
G. clavigera appears to have evolved two mechanisms to survive and grow when exposed to monoterpenes: GcABC-G1 controls monoterpene levels within the fungal cells and G. clavigera uses monoterpenes as a carbon source. This work has implications for understanding adaptation to host defences in an important forest insect–fungal system, and potentially for metabolic engineering of terpenoid production in yeast.
The ascomycete Grosmannia clavigera (Gc) is a fungal pathogen of pine trees (Pinus spp.) and is vectored by the mountain pine beetle (Dendroctonus ponderosae, MPB). MPB, Gc and other vectored microorganisms form an interactive biological complex that has caused a rapid, large-scale decline of lodgepole pine (Pinus contorta) in western North America (Lee et al., 2005). Grosmannia clavigera can kill trees and stain the sapwood blue or black when it is inoculated manually into trees at a certain density; such discoloration reduces the commercial value of lumber. In British Columbia alone, the MPB epidemic has already killed over 16 million hectares of lodgepole pine forests. With the recent spread of the MPB epidemic into forests east of the Rocky Mountains and an expansion of its host range into Jack pine (Pinus banksiana) (Cullingham et al., 2011), the MPB–microbial complex now threatens the Canadian boreal forest. This large-scale disturbance has caused massive economic losses to forest-based industries, and has important implications for forest ecosystem stability and global atmospheric carbon balance (Kurz et al., 2008).
While the MPB–Gc complex can successfully colonize > 20 different pine species, its preferred host is P. contorta (Safranyik et al., 2010). Like all conifers, the pine hosts of the MPB epidemic have complex oleoresin-based chemical defences that protect these trees against most potential pests and pathogens (Keeling & Bohlmann, 2006a,b; Boone et al., 2011; Bohlmann, 2012). The oleoresin of most conifers consists predominantly of monoterpenes and diterpene resin acids, with smaller amounts of sesquiterpenes. These terpenes can be fungistatic or fungicidal. Small lipophilic monoterpenes diffuse easily into and through eukaryotic cell membranes, interact with membranes and membrane-bound enzymes, and can change membrane fluidity and ultrastructure (Parveen et al., 2004; Bakkali et al., 2008; Witzke et al., 2010). They can also cause fungal cells to swell, shrink and vacuolize (Soylu et al., 2006). While antimicrobial properties of monoterpenes are documented, little is known about the mechanisms used by some microorganisms, particularly fungi that colonize conifers, to survive and grow in the presence of monoterpenes. The highly specialized MPB–Gc complex, which colonizes the monoterpene-rich environment of pine phloem and sapwood, requires mechanisms to overcome host defence chemicals. To discover mechanisms involved in the ability of Gc to cope with host terpenes, we analysed the Gc genome and transcriptome to identify genes that are differentially expressed in response to terpene treatments (DiGuistini et al., 2011). We noted that a gene annotated as an ATP-binding cassette (ABC) transporter, GcABC-G1 (previously reported as GLEAN_8030), was highly up-regulated in the terpene-induced Gc transcriptome responses (DiGuistini et al., 2011).
Fungal plant pathogens evolve a combination of strategies to colonize and survive in the unfavourable conditions occurring in their living environment, the host. These strategies include transporting toxic chemicals out of the cell or sequestrating them in cellular organelles, detoxifying host defence compounds by converting or modifying them and interfering with host signalling (Morrissey & Osbourn, 1999). Mechanisms can involve enzymes such as cytochrome P450s and membrane proteins such as ABC or MFS transporters (Han et al., 2001; Coleman & Mylonakis, 2009). Fungal ABC transporters are well known for their roles in the secretion of harmful chemicals (Sipos & Kuchler, 2006; Coleman & Mylonakis, 2009; Coleman et al., 2011). Typical ABC transporters consist of two transmembrane domains (TMDs) and two nucleotide-binding folds (NBFs); ‘half-transporters’ contain only one TMD and one NBF. ABC transporters are classified into subfamilies according to sequence homology and domain topology (Sipos & Kuchler, 2006; Lamping et al., 2010). In eukaryotes, eight major subfamilies have been defined: ABC-A to ABC-H (Dean & Allikmets, 2001; Verrier et al., 2008). Among these, full-size ABC-B, ABC-C and ABC-G are, respectively, referred to as multidrug resistance (MDR), multidrug resistance-associated protein (MRP) and pleiotropic drug resistance (PDR) (Paumi et al., 2009; Kovalchuk & Driessen, 2010). Fungal PDR transporters are located in the cytoplasmic membrane and function as efflux pumps, contributing to drug resistance, chemical sensitivity and cellular detoxification (Coleman & Mylonakis, 2009; Lamping et al., 2010).
Here, we characterized the responses of Gc to a range of individual terpenes and terpene mixtures. We identified and classified all putative ABC transporter genes in the Gc genome. We then profiled their expression and characterized the function of GcABC-G1, which was the gene most strongly induced by terpenes. Deleting this gene in Gc, and expressing it heterologously in Saccharomyces cerevisiae demonstrated that GcABC-G1 confers tolerance to monoterpenes. GcABC-G1 appears to be specific to monoterpenes and to function as an efflux ABC transporter. Transcripts of GcABC-G1 were detected in stem tissues of young lodgepole pine trees inoculated with Gc. The development of the symptoms in young trees that were inoculated with the deletion mutant was delayed relative to trees inoculated with the wild-type Gc, suggesting that GcABC-G1 contributes to the fungus' ability to overcome host defence chemicals and survive in an environment that is highly unfavourable to most organisms. Plant terpenes, beyond being important in conifer defence (Keeling & Bohlmann, 2006a,b) and in the interactions of plants with other organisms (Gershenzon & Dudareva, 2007), are also being actively explored for metabolic engineering of biofuels and bioproducts in microbial hosts (Bohlmann & Keeling, 2008; Peralta-Yahya et al., 2011). The discovery of a role of GcABC-G1 in tolerance to monoterpenes may lead to applications for improved microbial production systems for terpenoids (Dunlop et al., 2011; Ignea et al., 2011).
Materials and Methods
Strains and plasmids
Grosmannia clavigera strain kw1407 (NCBI Taxonomy ID: 655863) was deposited at the University of Alberta Mycological Herbarium (UAMH1150). It is important to note that G. clavigera consists of two cryptic species, Gc and Gs (Alamouti et al., 2011). The name Gc should be reserved for the holotype described in 1968. While the species described here belongs to the Gs group, which is not yet fully described, for continuity with our recent genomic work (DiGuistini et al., 2011) we have used the name Gc in this manuscript. Saccharomyces cerevisiae wild type BY4741(MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) was provided by Dr Christopher Loewen, UBC. Agrobacterium GV3101, a laboratory stock, was used to transform Gc. Expression vector pESC-URA (Stratagene, Accession no. AF063585) was used for yeast expression.
Fungal media and growth condition
RNA-seq and quantitative reverse-transcription PCR analysis of Gc grown on complete medium (see the Supporting Information, Methods S1) treated with blends of monoterpenes and diterpenes were described by DiGuistini et al. (2011). For experiments to characterize monoterpene utilization, fungal spores were incubated for 3 d on 1% malt extract agar (MEA) overlaid with cellophane (cat#80611781; Amersham Biosciences), then the young mycelia were transferred to yeast nitrogen base (YNB) minimal medium (Methods S1) in glass plates, two (2 × 4 cm) strips of filter paper were placed inside the lid of the plate and 200 μl of individual monoterpenes (saturated, 1.4 g l−1 limonene) or a mixture of monoterpenes (MT) were added onto the filter paper. The plates were sealed with DuraSeal film (cat# 89031-573, Laboratory Sealing Film; VWR, Mississauga, Ontario, Canada) and incubated at 22°C in a sealed glass container, if necessary, monoterpenes were resupplied biweekly until the mycelia covered at least half of the media surface (3–4 wk). Fungal growth rate assays were carried out on MEA. For monoterpene treatments, we transferred plugs of actively growing fungal cultures into the centre of glass plates containing MEA medium and applied monoterpenes using the method described above; for other chemical assays, the chemicals were added after autoclaving the media. The optimal concentration of each chemical was determined by gradient tests. The final concentrations of the chemicals were: 3.3 μM for diterpenes (abietic acid, dehydroabietic acid and isopimaric acid), 2 μM for azoles (propiconazole, tebuconazole), 50 μM for flavonoids (quercetin and fisetin), 500 μM for antibiotics (cycloheximide, erythromycin), 7.5 μM for phenolic compounds (benzoic acid, salicylic acid, vanillic acid, gentisic acid) and 50 μM for phytoalexins (catechin, resveratrol and taxifolin).
Annotation of GcABC genes
HMMER3.0 (www.hmmer.janelia.org) was used to search the Gc genome and expressed sequence tags (ESTs) and retrieve gene models encoding proteins containing the conserved ABC motif sequences ‘ABC-transporter (PF00005)’, ‘ABC-2 transporters (PF01061)’ and ‘ABC transporter transmembrane region (PF00664)’. The resulting gene model annotations were curated manually, considering alternatively spliced isoforms, exon-intron boundaries and coding starts/stops. Potential pseudo-genes were excluded. The Magnaporthe grisea ABC transporter collection (Kovalchuk & Driessen, 2010) was used as a query for tblasn local searches and to confirm gene prediction and classification in Gc.
For each RNA-seq library, we collected samples from three biological replicates, extracted RNA separately and pooled the samples for paired-end sequencing on an Illumina GAIIx (Canada's Michael Smith Genome Science Center, Vancouver, BC, Canada). Sequence filtering, trimming, mapping to the reference genome and RNA-seq analyses were conducted on CLC Genomic Workbench v4 (http://www.clcbio.com) software. Differential expressed genes were identified by comparing the number of reads per kilobase of exon model per million mapped reads (RPKM) between treatment and control data sets; statistical analysis was carried out by Kal et al.'s Z-test (1999). Five RNA-seq data sets were analysed. Two were generated in previous work (DiGuistini et al., 2011) and the third was generated for Gc growing on YNB with a mixture of monoterpenes as the sole carbon source. Results were normalized to the sample grown on mannose as a carbon source. To compare the response of Gc and the deletion mutant to a mixture of monoterpenes, both strains were grown on MEA and treated with a mixture of monoterpenes for 12 h before sampling, and gene expression was normalized to the untreated controls.
Quantitative reverse transcription PCR (qRT-PCR)
Differential gene expression was validated by qRT-PCR. Fungal mycelium samples were removed at 0, 6, 12, 36, 48 and 72 h from complete medium with mono/diterpenes (CM + T) and 7, 10, 14, 18 d from YNB minimal medium with a mixture of monoterpenes as sole carbon (YNB + MT); GcABC-G1 expression was also measured in phloem of young pine trees inoculated with Gc for 4, 7, 14 d (see inoculation of pine, later). Phloem tissues next to the inoculation points were removed and immediately frozen in liquid nitrogen. Extraction of total RNA, cDNA synthesis and qPCR were performed as described by Hesse-Orce et al. (2010). Three biological and technical replicates were used for each time-point. Data collection and statistical analysis were carried out on the Bio-Rad CFX96 real-time PCR detection system.
Producing gene deletion mutants
ATP-binding cassette transporter mutants were generated using an Agrobacterium-mediated gene deletion procedure (Wang et al., 2010). The whole gene open reading frame was replaced with the selective gene marker hygromycin B (hph). Gene replacements were verified by PCR amplification of adjacent regions, targeted region and selective marker gene; copy numbers were determined by Southern blot.
Heterologous expression of GcABC-G1 in S. cerevisiae
The full-length cDNA of GcABC-G1 was amplified and cloned into the yeast expression vector pESC-URA under the control of the GAL1 promoter, using conventional digestion/ligation methods (HinDIII and BamHI). Yeasts were also transformed with the empty vector pESC-URA that was used as a control. For spot tests, yeast cells were induced in synthetic galactose (SG) broth, diluted to an OD600 of 0.1 and spotted on SG glass plates. Four filter paper discs (0.5 cm each) were placed in the centre of the plate and were loaded with 60 μl mixture or individual monoterpenes to provide a saturated environment. The plates were sealed with DuraSeal film and incubated face-up at 28°C until colonies showed. For survival tests, induced yeast cells were diluted to an appropriate density and spread on SG glass plate. The yeast cells were treated for 1 h with a specific monoterpene (5 μl) diluted in 245 μl ethanol; the solution was applied on five small filter papers placed inside the lid and the glass plates were sealed with DuraSeal film. After a 1-h incubation, the chemical was removed and plates were further incubated for 4 d at 28°C. The numbers of surviving cells with and without monoterpene treatments were counted.
Inoculation of young lodgepole pine trees with Gc or its GcABC-G1 mutant
Five-year-old lodgepole pine trees were grown in the UBC greenhouse and maintained as described previously for other conifer saplings (Miller et al., 2005). Trees were inoculated at six points along the stem with plugs of actively growing fungal mycelium on MEA medium. Circular bark plugs were removed from the outer stem tissue using a 5 mm diameter metal cork borer. Inoculations were done on opposite sides of the stem at locations that were 5, 10 and 15 cm above the base of the stem. Fungal inoculums consisting of 5 mm diameter circular MEA/mycelium plugs were inserted into each circular bark hole and a bark plug placed on top to close the hole. The inoculated stem section was sealed with Parafilm and an outer layer of duct tape (Wang et al., 2010). For control treatments, we used MEA plugs without fungal mycelium. Symptoms, that is, wilting and discoloration of needles and growing shoot tips, were recorded weekly for 4 wk. After 4 wk after inoculation the stems were harvested, and needles, branches and outer bark tissue were removed. Discoloration of the inner stem tissue was recorded, and fungi (i.e. Gc or its mutant) were reisolated from the phloem and the inner stem. Replicate experiments were carried out in April 2011 and May and June 2012. For each replicate, seven trees were used for the control and 11 trees were inoculated with either Gc or the deletion mutant.
ABC transporters in the G. clavigera genome
We identified 39 putative ABC protein genes in the Gc genome (DiGuistini et al., 2011). All of the predicted proteins contained at least one NBF domain. These gene models were located in 16 contigs; 23 (59%) of those genes were also found in Sanger ESTs (Hesse-Orce et al., 2010) and 37 (95%) in the RNA-seq transcriptome data (DiGuistini et al., 2011). The transporter gene lengths varied from 999 to 5241 bp and intron numbers varied from zero to 14 per gene. Using functional domain predictions and tblastn searches of other fungal genomes, we classified these ABC proteins into subfamilies following the nomenclature used by the Human Genome Organization (HUGO), and named these subfamilies GcABC-A to GcABC-G. These subfamilies were further divided into groups based on sequence similarity and phylogenetic analyses (Kovalchuk & Driessen, 2010). Table 1 gives Genbank IDs, genome locations, intron numbers and predicted topology of all GcABCs. Among the 39 GcABC proteins, 24 were full transporters with more than one TMD or NBF, while eight were half transporters with either a TMD-NBF or an NBF-TMD domain arrangement. The remaining seven had one or two NBFs but lacked TMDs, and were considered not to be membrane proteins or not to have a transport function. We compared the numbers of proteins in each subfamily for Gc, the yeasts S. cerevisiae and Yarrowia lipolytica, the ascomycete rice pathogen M. grisea, and the ascomycete saprophyte N. crassa (Table S1). The total number of ABC transporters in these species was similar, ranging from 30 to 48. Gc has five group V members in both ABC-C and in ABC-G subfamilies, while M. grisea has only two ABC-C-group V members and one ABC-G-group V member. In contrast to mitochondrial and peroxisome transporters (GcABC-B, D), which were highly conserved across these species (70–80% identity), seven out of the 10 GcABC-C members shared < 50% amino acid identity with closely related fungal species.
Table 1. Inventory and key features of the ABC transporters from the genome of Grosmannia clavigera
Annotation was corrected from the published genome.
Two adjacent half transporters were annotated as one full transporter.
Expression of GcABC transporter genes in response to terpenoids
To identify the ABC transporters involved in Gc's response to terpenes, we compared RNA-seq expression profiles of all GcABCs for: (1) Gc grown on a complete medium with a blend of monoterpenes and diterpenes (CM + T) for 12 h and 36 h (DiGuistini et al., 2011); and (2) Gc grown on a YNB medium with a mixture of monoterpenes (MT) as sole carbon source (YNB + MT). Fig. 1(a) shows GcABC transporter genes whose transcripts were significantly upregulated in at least one of these conditions (Z test P-value < 0.05). For CM + T, six GcABC transporter genes were upregulated at 12 h, while 11 were upregulated at 36 h. For YNB + MT, 10 GcABCs were upregulated. Of five GcABCs that were upregulated in all three conditions, two (GcABC-G1, G2) belonged to the ABC-G-group I transporters, and three (GcABC-F1, F2, F3) belonged to the ABC-F subfamily, whose members are not considered to be true transporters because of the absence of TMD (Kovalchuk & Driessen, 2010). Three GcABCs were upregulated only on the more restrictive YNB + MT medium, in which monoterpenes are the only available carbon source. These included a PDR (GcABC-G3), a putative vacuolar transporter (GcABC-B5) and a putative peroxisome transporter (GcABC-D1) that may be involved in fatty acid metabolism.
GcABC-G1 was the most strongly upregulated of the 39 GcABCs. Its abundance of transcript relative to controls increased at least 100-fold in all three conditions, and transcripts were almost 1500-fold more abundant in YNB + MT than in YNB + mannose, which we used as a control (Fig. 1a). We validated these results by qRT-PCR analysis at times up to 72 h for Gc growing on CM + T. The change of GcABC-G1 transcript abundance was up to 115-fold after 6 h, reached a peak of 648-fold at 12 h, and was still above 50-fold at 72 h (Fig. 1b). GcABC-G1 also showed increased transcript levels (> 100-fold) throughout fungal growth on YNB + MT compared with YNB + mannose (Fig. 1c). We noted that GcABC-G1 was not induced by other stress treatments (e.g. oxidative, osmotic, nitrogen starvation, high temperature and lodgepole pine phloem extract).
GcABC-G1 occupies a unique position in the phylogeny of PDR transporters
GcABC-G1 belongs to the ABC-G-group I transporters; in this group, three members were identified in Gc (Table 1). Maximum likelihood (ML) phylogenetic analyses of the predicted amino acid sequences for these three proteins (GcABC-G1, GcABC-G2 and GcABC-G3) with sequences for 80 ABC-G-group I transporters from 23 ascomycete species (Table S2), resolved four distinct clades. These included two clades that appear to be yeast-specific, one of which contains only the yeast Yarrowia, and two Eurotiomycete–Sordariomycete clades (Fig. 2). GcABC-G3 was placed in a Eurotiomycete–Sordariomycete clade showing an orthologous relationship with Giberella zeae (FGSG_03882). GcABC-G2 was placed in a Sordariomycete-specific subclade, which includes several transporters that have been reported as pathogenicity factors (e.g. M. grisea MGG_13624) or exporters of plant defence chemicals (Nectria haematococca NECHADRAFT_63187) in plant pathogens (Urban et al., 1999; Coleman et al., 2011). Whether GcABC-G2 has comparable functionality for Gc remains to be confirmed. By contrast, GcABC-G1 was placed outside of these four clades, and was separated from other sequences included in our ML analysis.
Deleting GcABC-G1 increased fungal sensitivity to monoterpenes
To further functionally characterize GcABC-G1, we generated deletion mutants (Δgcabc-g1) using the method previously developed for Gc (Wang et al., 2010), and confirmed the deletion of the targeted gene by PCR and Southern blot (Fig. S1). On MEA, colony morphologies and growth rates were similar for Gc and Δgcabc-g1 (Fig. 3a). However, the mutant was more sensitive to monoterpenes than Gc. In the presence of individual monoterpenes the growth of Gc was not delayed by a racemic α-pinene, but was delayed by 1 d by (+)-3-carene, (+)-limonene, (−)-β-pinene and a mixture of monoterpenes (MT) (Fig. S2). In contrast, the mutant showed longer growth delays for all monoterpenes tested: 1 d for racemic α-pinene; 2 d for (+)-3-carene, (+)-limonene and MT, and up to 3 d for (−)-β-pinene (Fig. S2). We calculated fungal growth rates in the linear phase that followed such delays. Fig. 3(b) shows that the growth rates of Δgcabc-g1 and Gc were similar in the presence of racemic α-pinene (7 mm d−1) and only slightly different with (+)-3-carene (5–6 mm d−1). However, the growth rates of the mutant were 52% and 60% lower than Gc on (+)-limonene and (−)-β-pinene, respectively. The growth difference between Gc and Δgcabc-g1 was significant at all monoterpene levels tested (Fig. S3). These results indicate that GcABC-G1 supports the growth of Gc in the presence of certain monoterpenes.
To further assess GcABC-G1's role in the Gc response to monoterpenes, we compared the effects of monoterpenes on asexual spore germination in Gc and Δgcabc-g1 (Fig. 3c). For Gc, germination was not inhibited by racemic α-pinene or (+)-3-carene, and only partially reduced by (+)-limonene and (−)-β-pinene (70%). For the mutant, racemic α-pinene reduced spore germination by only 30%. However, (+)-3-carene, (+)-limonene, and (−)-β-pinene completely prevented spore germination; further, when we removed these monoterpenes after 6 d of incubation and continued the incubation, we found that 90% of Δgcabc-g1 spores had been killed.
In contrast to the monoterpenes, individual diterpenes (abietic acid, dehydroabietic acid, isopimaric acid) only slightly inhibited growth of Gc and Δgcabc-g1, and growth rates were similar for Gc and Δgcabc-g1 (Fig. S4).
We had previously reported that Gc was able to grow on YNB with a mixture of monoterpenes as a sole carbon source (DiGuistini et al., 2011). Here, we assessed which monoterpenes could support Gc growth and whether the utilization of specific monoterpenes was affected in Δgcabc-g1. Gc grew on YNB with 200 μl of (+)-limonene as a sole carbon source (Fig. S5) but not with the same amount of racemic α-pinene or (−)-β-pinene. To observe if the deletion of GcABC-G1 affects limonene utilization, we grew the GcABC-G1 mutant in the same condition as Gc and found that Δgcabc-g1 was killed by 200 μl of (+)-limonene on YNB medium (Fig. S5), while at 10 μl the mutant survived, but did not grow.
Deleting GcABC-G1 did not affect sensitivity to other potential PDR substrates
Based on sequence similarity and domain topology (Table 1), GcABC-G1 belongs to the PDR group of transporters, members of which are able to excrete a wide range of chemicals (Rogers et al., 2001; de Waard et al., 2006). To establish whether GcABC-G1 affects the tolerance of Gc to compounds other than monoterpenes, we assessed the growth rate of Gc and Δgcabc-g1 in the presence of other potential PDR substrates: azoles (propiconazole and tebuconazole), antibiotics (cycloheximide, erythromycin), flavonoids (fisetin, quercetin), simple phenolics (benzoic acid, salicylic acid, vanillic acid, gentisic acid) and phytoalexins (catechin, resveratrol, taxifolin). We selected these compounds based on the literature and determined experimentally which concentrations were affecting Gc growth rates. On MEA, none of the antibiotics, flavonoids or phytoalexins tested affected the growth of either Gc or the mutant. While the azoles and phenolics inhibited fungal growth, there were no inhibition differences between Gc and the mutant.
Heterologous GcABC-G1 expression enhanced survival of S. cerevisiae in the presence of monoterpenes
We used heterologous expression of GcABC-G1 in S. cerevisiae (Sc) to further assess the role of GcABC-G1 in monoterpene tolerance. GcABC-G1 was expressed under the control of the GAL1 promoter, which was induced in a synthetic galactose (SG) medium. Sc transformed with the vector only (Sc-V), or with the vector containing GcABC-G1 (Sc-ABC), grew at similar rates on yeast extract-peptone-dextrose (YPD) and SG media. However, when Sc-V and Sc-ABC were spotted onto SG plates and treated with 60 μl of a mixture of monoterpenes, only Sc-ABC had grown after 7 d of incubation (Fig. 4a). Monoterpene treatments with > 60 μl per plate killed all yeast cells. These initial tests established that both Sc-V and Sc-ABC were more sensitive to monoterpenes than the pine pathogen Gc. We further tested the survival of Sc-V and Sc-ABC with individual monoterpenes using a dilution plate assay. When 105 yeast cells were spread on SG plates and incubated with 5 μl of individual monoterpene for 3 d, neither Sc-V nor Sc-ABC survived. When the duration of monoterpene treatments was reduced to 1 h, a sufficient number of Sc-V and Sc-ABC cells survived to allow for comparative analyses (Fig. 4b). Under conditions of 1-h treatment with (+)-3-carene, 30-times more Sc-ABC cells survived than Sc-V; for (+)-limonene and (−)-β-pinene, seven- and three-times more Sc-ABC cells survived than Sc-V. In contrast, with racemic α-pinene, the numbers of surviving cells were low and not significantly different for Sc-V and Sc-ABC (Fig. 4b). These results showed that the heterologous expression of GcABC-G1 in Sc improved the survival of yeast cells in the presence of some but not all monoterpenes.
Together, the results obtained in independent experiments with Gc and Sc support a role for GcABC-G1 in tolerating (+)-limonene, (+)-3-carene and (−)-β-pinene, while effects varied for racemic α-pinene depending on the experimental system (Table 2).
Table 2. Summary of GcABC-G1-dependent differences in effects of specific monoterpenes on Grosmannia clavigera (Gc) asexual spores and transformed Saccharomyces cerevisiae (Sc) cells
Observation (A), Gc spore survival
Observation (B), transformed Sc survival
Comparison of observation (A) and (B)
+++, Difference between with and without GcABC-G1 is > 50%, P < 0.01.
+, Difference between with and without GcABC-G1 is < 50%, P < 0.01.
Δ, No significant difference between with and without GcABC-G1.
Pathogenicity and detection of GcABC-G1 transcripts in lodgepole pine inoculated with Gc or its GcABC-G1 mutant (Δgcabc-g1)
To test the effect of the deletion of the GcABC-G1 gene on the development of symptoms in lodgepole pine, we carried out greenhouse inoculations on stems of young lodgepole pines with Gc, mutant Δgcabc-g1 or controls without fungus (Fig. 5). Two weeks after inoculation, several trees inoculated with Gc showed early symptoms of infection, that is, wilting of growing shoots and browning of needles (Fig. 5a-2); similar symptoms were observed 2–3 d later in the pines inoculated with Δgcabc-g1. During the third and fourth week severe symptoms developed on branches and growing shoots for both Gc and Δgcabc-g1 (Fig. 5a-3). During the fourth week, all of the needles of several trees had become completely brown, and growing shoots had severely wilted (Fig. 5a-4). Although the numbers of trees that appeared healthy (Fig. 5a-1) was low for both treatments after 4 wk, trees treated with Δgcabc-g1 showed a 10–20% higher survival rate in all three experiments compared with trees inoculated with Gc (Fig. 5b). For both fungi, we also measured the typical blue/black discoloration in the stem cross-section of the young pine trees that is associated with fungal growth and melanin production. After 4 wk, we observed higher numbers of stem cross-sections with dark stains for Gc (81%) than for Δgcabc-g1 (21%) (Fig. S6). We were able to reisolate Gc and Δgcabc-g1 from stem cross-sections and from inner bark, but not from the controls inoculated with MEA agar plugs, confirming that the symptoms were caused by fungal growth and not by wounding.
To assess whether the GcABC-G1 gene was expressed in Gc upon inoculation of trees we measured its transcript levels in stem tissues inoculated with Gc or Δgcabc-g1 as well as in the controls treated with MEA plugs only. No GcABC-G1 gene transcripts were detected in the pine tissue treated with the Δgcabc-g1 or in the controls. However, in trees inoculated with Gc, GcABC-G1 transcripts were clearly detectable at 4, 7 and 14 d after inoculation. As transcript abundance was normalized to fungal β-tubulin transcripts, we were able to detect a temporal profile of increased relative abundance of the GcABC-G1 transcript with a maximum 178-fold change at 7 d relative to Gc grown on MEA (Fig. 6). The observed induction of GcABC-G1 gene transcripts when Gc grows in pine host tissue further supports a role of this gene in the pine–Gc interaction in vivo.
Comparing the monoterpene response transcriptomes of Gc and Δgcabc-g1
GcABC-G1 appears to play a critical role in Gc's tolerance to certain monoterpenes. To assess broader effects of the deletion of GcABC-G1 in the Gc's response to monoterpenes, we sequenced mRNA libraries of Gc and Δgcabc-g1 grown for 12 h on MEA, with and without a mixture of monoterpenes. In response to monoterpenes, RNA-seq analyses in Δgcabc-g1 identified transcripts for 1312 genes as significantly upregulated and 3459 as downregulated, while in Gc 961 genes were upregulated and 2501 genes were downregulated (P-value < 0.05) (Fig. S7). Gc and Δgcabc-g1 shared 846 upregulated genes (88% of the total upregulated in Gc; 64% of the total upregulated in Δgcabc-g1) and 2218 downregulated genes (88% of the total downregulated in Gc; 64% of the total downregulated in Δgcabc-g1). Using KEGG (http://www.genome.jp/kegg/) and InterProScan (http://www.ebi.ac.uk/Tools/pfa/iprscan/), we determined that the 466 genes that were upregulated only in the transcriptome of the mutant were mainly involved in metabolism (e.g. carbohydrate/lipid metabolism, macromolecular biosynthesis), genetic information processing (transcription, protein/RNA folding, sorting and degradation, DNA replication and repair), environmental information processing (transportation and signal transduction) and stress responses (Fig. S8). Further, in Δgcabc-g1, 10 ABC transporters showed expression changes; nine of them were upregulated more highly in the mutant than in Gc (P-value < 0.05) (Table 3). Genes that were upregulated in both Gc and Δgcabc-g1 included an acetyl-CoA-acyltransferase, an alcohol dehydrogenase and genes involved in fatty acid metabolism that may be involved in degrading hydrophobic compounds. These genes were also upregulated when Gc was grown on YNB with monoterpenes as the sole carbon source.
Table 3. Comparison of GcABC transporter genes showing transcript abundance differences in Grosmannia clavigera (Gc) and Δgcabc-g1
To survive and become established in a pine tree, the MPB symbiont Gc has to overcome preformed or induced host defence chemicals (Bohlmann, 2012). Terpenoids, and specifically monoterpenes, are among the most abundant antimicrobial pine defence chemicals. In previous work, we reported that terpenes induce a stress response and activate a cluster of Gc genes that may be involved in detoxification or tolerance of host terpenes, and that monoterpenes can serve as a sole carbon source for Gc (DiGuistini et al., 2011). In the present work, we demonstrated a role for GcABC-G1 in tolerance to certain monoterpenes, using a combination of growth experiments with a genetic deletion in Gc and heterologous expression in S. cerevisiae. We also inoculated young pine trees with Gc and the Δgcabc-g1 mutant to compare their pathogenicity and their expression of GcABC-G1 in the host. We propose that Gc employs a combination of mechanisms to cope with host defence monoterpenes. The pathogen may control intracellular levels of monoterpenes by the induced expression of an efflux ABC transporter GcABC-G1, and it can metabolize monoterpenes and use them as a carbon source.
The Gc genome contains all ABC transporter subfamilies found in closely related species. As indicators of Gc having mechanisms for processing xenobiotics, the Gc ABC transporter subfamilies include the ABC-B, C and G subfamilies, whose members confer drug resistance in other fungi. RNA-seq expression analysis showed that transcript levels of members of the GcABC-F and GcABC-G subfamilies were upregulated when Gc was exposed to terpenes or grown in the presence of monoterpenes as the sole carbon source, but not under other stress conditions tested. The ABC-F subfamily has been reported as essential for cell viability and involved in ribosome biogenesis and translation, which would be activated while the fungus responded to sudden exposure to toxic compounds (Kovalchuk & Driessen, 2010). There are three ABC-G group I transporters in Gc (GcABC-G1, GcABC-G2 and GcABC-G3) and they are all upregulated in Gc in response to terpenes; GcABC-G1 was the most strongly induced in response to terpenes (i.e. > 100-fold change) while G2 (fourfold changes) and G3 (eightfold change) were expressed at a much lower level. GcABC-G2 has orthologues that are described as pathogenicity factors in M. grisea (MGG13624), Gibberellapulicaris (GpABC1), and N. haematococca (NECHADRAFT_63178) (Urban et al., 1999; Fleissner et al., 2002; Coleman et al., 2011), but its function remains to be tested in Gc in the future. The GcABC-G1 protein sequence had the full (NBF-TMD)2 domain organization that is common in PDR efflux transporters, which are localized in the cytoplasmic membrane in other fungal species; for example, the camalexin exporter BcatrB in B. cinierea and the pisatin exporter NhABC1 in N. haematococca (Stefanato et al., 2009; Coleman et al., 2011). The literature suggests that fungal PDRs evolved broad substrate specificities to export diverse antifungal compounds, including compounds derived from plants (Sipos & Kuchler, 2006; de Waard et al., 2006; Cannon et al., 2009). However, GcABC-G1 conferred resistance to monoterpenes but not to typical PDR substrates such as azoles, antibiotics, flavonoids, phytoalexins or simple phenolics. GcABC-G1 also had no close orthologues in the large set of ascomycete sequences analysed in this work, suggesting a specialized function as opposed to general xenobiotic transport. When GcABC-G1 was deleted, the mutant's transcriptome showed an elevated stress response to monoterpenes that included upregulation of other ABC transporters. However, none of these were upregulated to the same extent as GcABC-G1 in Gc, and none appeared to substitute functionally for GcABC-G1 in the mutant.
GcABC-G1 appears to have evolved as a specialized monoterpene transporter that may allow Gc to better colonize a unique ecological niche: the monoterpene-rich tissues of living pine hosts. As Witzke et al. (2010) has shown that limonene can freely diffuse into biological membranes, it is plausible that monoterpenes can enter fungal cells by diffusion. While measuring both this volatile monoterpene and the fungal biomass were challenging, we were able to detect (+)-limonene in the young germinating spores of Gc or its mutant when exposed to this compound for 1 h, which is too short an exposure time for the fungus to have produced ABC transporter proteins. However, after 18 h, while the level of (+)-limonene remained high in the cells of the mutant, we detected only trace amounts of this monoterpene in Gc (Table S3). The data at 1 h are inconsistent with the ABC transporter importing monoterpenes, while the data at 18 h are consistent with induced GcABC-G1 playing a role in reducing the concentration of monoterpene in the cells. We suggest that a similar process for controlling cellular monoterpene levels would occur in the pine phloem when the MPB vector disperses and so exposes fungal spores and young colonizing mycelia to monoterpenes, where strong inducible expression of GcABC-G1 should provide an adaptive advantage in survival and growth.
ATP-binding cassette transporters, and particularly the ABC-G group I (PDR) transporters, play a role in plant infection by fungal pathogens by protecting pathogens from exogenous toxic compounds produced by hosts (Urban et al., 1999). Given this, we compared the pathogenicity of Gc and its mutant on young pine trees, and measured the transcript abundance of GcABC-G1 in pine stem tissue inoculated with Gc or the mutant. While both Gc and the mutant affected the health of young pine trees, symptoms were delayed by 3 d for the mutant, and the survival rate was higher for the mutant than for Gc. These results suggest that Gc survives in and colonizes pine trees more efficiently than the mutant, which lacks the ABC transporter. It should not be surprising that the difference in effects caused by Gc or its mutant was not more pronounced, given that pine trees combat fungi with a range of chemical defences (Kolosova & Bohlmann, 2012). The results are consistent with our in vitro data for monoterpene treatments on MEA, in which the GcABC-G1 mutant survived the treatment but required 2–3 d more than Gc to adapt to the chemical, and then grew more slowly than Gc. A role for the GcABC-G1 transporter in pine colonization is further supported by the induction profile of the fungal GcABC-G1 gene in young pine trees inoculated with Gc. The absence of transporter expression in young pine inoculated with the deletion mutant, and the early and peak expression at 7 d of the GcABC-G1 gene, indicate that this transporter plays a role in the early phase of fungal colonization.
The unique ecological pine host niche colonized by Gc has high levels of monoterpenes and so would be unsuitable to most microorganisms. For example, in the broad range of niches in which Sc strains are found in nature, including grapevine berries, concentrations of terpenes are typically low, and, to this point, no mechanisms for coping with high concentrations of monoterpenes have been reported for Sc. The heterologous expression of GcABC-G1 in Sc conferred increased resistance to monoterpenes, consistent with this transporter potentially being an efflux pump that may remove toxic monoterpenes from cells. In the current work, the four monoterpenes assessed were far more toxic to Sc than to Gc mycelia or germinating Gc spores. When exposed for shorter periods of time to certain monoterpenes, more cells survived for Sc transformed with GcABC-G1 than for Sc transformed with only the vector. Metabolic engineering of Sc and other microorganisms is being actively pursued for the production of monoterpenes and other terpenoids of plant origin as high-value bioproducts and advanced biofuels (Kirby & Keasling, 2009; Fischer et al., 2011; Zerbe et al., 2012). While many plant species, and in particular the resin-producing conifers, develop specialized anatomical structures for sequestering large amounts of low molecular weight terpenoids (Bohlmann & Keeling, 2008), in engineered single-cell production systems, the toxicity of monoterpenes and biofuels may limit yield and performance (Dunlop et al., 2011). As GcABC-G1 may be the first reported eukaryotic ABC transporter with a role in enhanced tolerance against monoterpenes, this gene may be of interest for protein and metabolic engineering geared towards improved terpenoid production in Sc and other systems.
Finally, we also demonstrated that Gc could process monoterpenes as a carbon source for growth. While conversion of monoterpenoids and diterpenoids into nontoxic compounds has been shown for Penicillium caseifulvum, Aspergillus niger and Botryococcus braunii (de Carvalho & da Fonseca, 2006), to our knowledge, fungal utilization of monoterpenes as carbon sources has not been reported in filamentous fungi. It is important to note that the GcABC-G1 mutant did not grow at very low dosages of monoterpenes but was able to survive. Except for GcABC-G1, when Gc and the mutant were grown on MEA with monoterpenes, RNA-seq data showed that similar genes were upregulated (for example, acetyl-CoA-acyltransferase, alcohol dehydrogenase and genes involved in fatty acid metabolism). This suggests that the same enzyme-mediated metabolism pathways were induced in both Gc and its mutant. In ongoing work we are creating additional mutants in order to characterize the functions of these genes and their roles in monoterpene utilization.
The authors thank Karen Reid for excellent project and laboratory management in the laboratory of J.B. and in support of the Tria project, and the undergraduate students who have worked in the laboratory of C.B. on this project. We thank Ms Lina Madilao for characterizing the chirality of monoterpenes used in this work. This work was funded by grants from the Natural Sciences and Engineering Research Council of Canada (NSERC; grant to J.B. and C.B.), and funds from Genome Canada, Genome BC and Genome Alberta (grant to J.B. and C.B.) in support for the Tria project (www.thetriaproject.ca).