Title: Grass–endophyte interactions: a note on the role of monosaccharide transport in the <i>Neotyphodium lolii</i>–<i>Lolium perenne</i> symbiosis
Abstract: Symbiotic mutualistic associations of plants with ecto- and endomycorrhizal and endophytic fungi are very common in natural and agricultural ecosystems; these fungi improve nutrient acquisition and/or tolerance to biotic and abiotic stresses of their hosts (Smith & Gianinazzi-Pearson, 1988; Nehls et al., 2007; Schardl et al., 2007). A major feature of associations between the heterotrophic fungi and autotrophic plants is an exchange of fungus-derived benefits (e.g. phosphorous in ecto- and endomycorrhizal associations, or antiherbivorous alkaloids in endophytic Neotyphodium–grass associations) for plant-derived carbohydrates as carbon (C) and energy source for fungal growth and maintenance (Bago et al., 2000; Nehls et al., 2007). Neotyphodium/Epichloë spp. endophytic fungi occur in 20–30% of cool-season grass species and are of widespread interest to ecological and agricultural research (Schardl, 2001). These obligate symbiotic fungi are an additional C sink and affect host C metabolism and mobilization (Pan & Clay, 2004; Hunt et al., 2005; Rasmussen et al., 2008, 2009). Under stressful conditions, in particular, such as drought and high temperatures, endophyte symbionts have been shown to enhance host photosynthesis and potentially increase total C reserves (Richardson et al., 1993; Marks & Clay, 1996). It has also been shown that endophyte concentrations are reduced in Lolium perenne cultivars accumulating high amounts of water-soluble carbohydrates compared with control cultivars (Rasmussen et al., 2007; Liu et al., 2011). These studies indicate that endophytic fungi are important regulators of host carbohydrate metabolism; and that carbohydrate supply by hosts and C utilization by endophytes interact with each other. Carbon transfer from host plants to fungal symbionts is catalysed by transporter proteins, and several sugar transporters from mutualistic ecto- and endomycorrhizal fungi have been functionally characterized (Schüβler et al., 2006; Nehls et al., 2007; Helber et al., 2011). In general, it has been assumed that simple soluble sugars such as glucose and fructose are the major source of C for symbiotic fungi (Bago et al., 2000); however, some of the isolated sugar transporters have been shown to be able to catalyse the uptake of cell wall sugars such as mannose and xylose as well (Schüβler et al., 2006; Fajardo López et al., 2008; Helber et al., 2011; Doidy et al., 2012). To provide more information about sugar uptake by endophytic Neotyphodium/Epichloë fungi from their hosts, we report here the isolation and functional characterization of a monosaccharide transporter from Neotyphodium lolii (MSTN). Initial studies indicated that MSTN preferentially catalyses the uptake of mannose, a monosaccharide mainly found in polymeric cell wall carbohydrates. We therefore hypothesized that N. lolii might be able to hydrolyse cell wall carbohydrates. We tested this hypothesis by quantifying the expression of cell wall carbohydrate hydrolysing fungal mannosidase, cellulase and glucanase, as well as sucrose hydrolysing fungal and plant invertases in cultured N. lolii mycelia and in ryegrass cultivars and tissues differing in their endogenous sugar content. The total RNA of N. lolii (strain Lp19) was isolated from fungal mycelia cultured in liquid potato dextrose medium (BD Difco™, Becton, Dickinson and Company, North Ryde, NSW, Australia), and cDNA libraries were synthesized using the SMART™ RACE cDNA Amplification Kit (Clontech; Norrie Biotech, Auckland, New Zealand). A gene fragment was identified in an in-house endophyte expressed sequence tag (EST) database (AgR Ltd, Hamilton, New Zealand) using a keyword search ('sugar transporter'); sequence analysis confirmed high similarity with known fungal sugar transporters. Full-length mstN cDNA and genomic DNA sequences were amplified (for primers see Supporting Information, Table S1; mstN GB Acc.# HQ413097.1) and PCR products were sequenced (Allan Wilson DNA Centre, Massey University, Palmerston North, New Zealand). To build a phylogenetic tree, the MSTN protein sequence was analysed using the software programmes BlastP and ClustalW. A sequence identity dendrogram of fungal sugar transporters (Fig. S1) was obtained using NCBI Blast pairwise alignments by the neighbour-joining method. A construct containing mstN was heterologously expressed in a yeast (Saccharomyces cerevisiae) mutant strain (EBY.VW4000) devoid of any monosaccharide transporters (MSTs; Thompson et al., 1998; Schüβler et al., 2006). The mutant yeast strain is unable to take up and grow on hexoses, but it can be cultivated on maltose-containing media (Wieczorke et al., 1999). Initial growth rates of yeast hxt-null recombinants complemented with mstN and empty-vector controls were determined in YNB (yeast nitrogen base) media supplemented with 0.1% of various hexoses. After inoculation, subsamples were collected during a time course (0–171 h) and cell concentrations (OD600) measured using a microplate reader (VERSAmax; Labspec, Hamilton, New Zealand). Least significant differences (LSD) were calculated for separating treatment differences in yeast cell growth. Uptake experiments using either 1 mM D-[U-14C] mannose (Amersham CFB26, specific radioactivity 111 kBq μmol−1) or D-[U-14C] glucose (Amersham CFB96, specific radioactivity 9435 kBq μmol−1) were performed according to Schüβler et al. (2006) and Helber et al. (2011). Radioactivity was measured with a liquid scintillation counter (Wallac μBeta Trilux 1450 LSC & Luminescence counter; Perkin Elmer, Melbourne, Victoria, Australia) after addition of a scintillation cocktail (Starscint; Perkin Elmer). Km values were estimated using the Michaelis–Menten equation (V = (Vmax[S])/(Km + [S])) with V, 14C-hexose uptake rate and [S], concentration of glucose (or mannose). The pH optimum of D-[U-14C] glucose uptake of mstN-expressing yeast cells was analysed in 100 mM potassium phosphate buffer adjusted to pH values 5.0, 6.0, 7.0 and 8.0. To test proton dependency of sugar uptake, yeast cells expressing mstN were preincubated for 2 min with 100 μM of the transmembrane proton gradient uncoupler carbonylcyanide m-chlorophenyl hydrazone (CCCP) before addition of D-[U-14C] glucose. For all experiments, three independent mstN-expressing yeast transformants were used as replication. In vitro cultured N. lolii (strain Lp19) mycelia from agar plugs were propagated in liquid potato dextrose broth for 2 wk. Washed mycelia were subsequently incubated for 4 d at 26°C in sterile H2O-based d-glucose, d-mannose and d-fructose solutions. Forty different endophyte-free or N. lolii (Lp19)-infected genotypes from two L. perenne L. cultivars, the high sugar grass 'AberDove' and a normal sugar grass 'Fennema' (Parsons et al., 2004) were grown in a controlled environment chamber as described previously (Liu et al., 2011). Mycelia, leaf blades and pseudostems were immediately frozen and ground in liquid nitrogen, and then stored at −80°C for RNA isolation. Total RNA was isolated and reverse-transcribed using the SuperScript®VILO™ cDNA Synthesis Kit (Invitrogen). Quantitative PCR was performed as previously described (Rasmussen et al., 2007). Primers (Table S1) for quantifying the expression of N. lolii and L. perenne genes were designed for β-1,6-glucanase (Bryant et al., 2007), cellulase, α-mannosidase, fungal invertase, and vacuolar, cell wall and cytosolic plant invertases. Sequences for mannosidase and fungal invertase were obtained from the Epichlöe festucae genome database (Epichloë_festucae_2368-scaffolds 2011-02 database; http://www.endophyte.uky.edu/). For accession and gene contig numbers, see Table S1. N. lolii gene expression data were normalized against the expression of β-tubulin (Bryant et al., 2007); L. perenne gene expression data were normalized against the expression of the protein elongation factor LpeEFα (Martin et al., 2008). The transcript abundances are expressed as the ratio of target gene copies to housekeeping gene copies; data were statistically analysed by ANOVA. Based on sequence similarity, conserved motifs and predicted membrane topology, the sugar transporter isolated from N. lolii is classified as an MST of the sugar porter family included in the major facilitator superfamily (Notes S1). The MSTN amino acid sequence is highly similar to many putative MSTs from Ascomycetes (Fig. S1), with an 81% amino acid sequence similarity to the hexose transporter CgHXT3 from the hemibiotrophic maize pathogen Colletotrichum graminicola (Lingner et al., 2011), but relatively distant from sugar transporters from the ectomycorrhizal Basidiomycetes Laccaria bicolor (CAQ53118; 46%) and Amanita muscaria (CAB06078; 44%). N. lolii is the asexual derivative of E. festucae and is nearly indistinguishable from it at the genetic level (Schardl et al., 1994; Schardl, 2001). Alignment of the mstN genomic nucleotide sequence with genes in the E. festucae genome database (Epichloë_festucae_2368-scaffolds 2011-02; http://www.endophyte.uky.edu/) resulted in the identification of EfM2.114080_A061106_1361_3407_5576_Alt0_PolyA (Supercontig 42:67179.688134), which was highly similar to mstN (99% identity). Only one sequence with high similarity to mstN was identified in the E. festucae database, suggesting that mstN is a single copy gene in Neotyphodium/Epichloë spp. A keyword search ('sugar transporter', 'sucrose transporter') resulted in several hits to putative other sugar/sucrose transporters, but none of these sequences was homologous to mstN (data not shown). The monosaccharide transporter from Neotyphodium lolii complemented the sugar uptake deficiency in a yeast strain devoid of all MSTs (Wieczorke et al., 1999), and initial growth rate assays (Fig. 1a) as well as kinetic studies (Fig. 1b) indicated that MSTN has a higher affinity for mannose (Km = 0.47 mM) than for glucose (Km = 5.77 mM). Competition assays with 14C-glucose (Fig. 1c) showed that, at high sugar concentrations, glucose was a better competitor than mannose, while the cell wall sugars galactose and xylose were intermediate competitors. Neither fructose nor the sugar alcohols mannitol and myo-inositol competed with uptake of glucose. The optimum pH for glucose uptake was 7.0 (Fig. 1d), and, like many other MSTs, MSTN was inhibited by CCCP (Fig. 1e), indicating it is an energy-dependent H+-symporter. These results demonstrate that MSTN has the capacity to catalyse the transport of monosaccharides, including the cell wall sugars mannose and xylose. At this stage we cannot exclude the possibility that MSTN is also able to catalyse transport of sucrose, which would require the analysis of MSTN in an invertase deficient yeast strain. Functional characterization of monosaccharide transporter from Neotyphodium lolii (MSTN) in complemented yeast hxt-null mutants. (a) Initial growth curves in 0.1% hexose supplied media. Squares, d-mannose; diamonds, d-fructose; triangles, d-glucose; circles, d-galactose. Line with crosses, empty vector controls in d-mannose. (b) Uptake kinetics for 14C-glucose and 14C-mannose. (c) Substrate competition of various sugars (3.5 mM) with 14C-glucose (1 mM). (d) Uptake of 14C-glucose at different pH values. (e) Uptake of 14C-glucose in the presence of the protonophore CCCP. Error bars indicate ± SE of three replicates. Different letters in (c and d) denote means that are significantly different. Differences in mstN transcript abundance were associated with soluble carbohydrate concentrations both in culture and in planta, showing a lower expression at high carbohydrate concentrations. In the absence of any carbohydrate substrates, mstN was highly expressed in cultured N. lolii mycelia, while addition of hexoses resulted in a strong down-regulation of mstN expression (Fig. 2a; F3,11 = 25.29, P < 0.001 for glucose; F3,11 = 5.16, P < 0.05 for mannose; F3,11 = 44.72, P < 0.001 for fructose), comparable to expression patterns seen for hexose transporters from the ectomycorrhizal fungi Tuber borchii and L. bicolor (Polidori et al., 2007; Fajardo López et al., 2008), as well as for CgHXT3 from C. graminicola (Lingner et al., 2011). Expression of β-tubulin was not different in mycelia from N. lolii grown in hexose-free water compared with hexose containing water as indicated by the Ct values (data not shown). Expression analysis of mstN and carbohydrate hydrolases in culture and in planta. (a, b) Expression of mstN in cultured mycelia of Neotyphodium lolii supplied with various hexoses (a) (black bars, 0 mM; dark grey bars, 0.1 mM; light grey bars, 1 mM; white bars, 10 mM hexose), and infected perennial ryegrass (Lolium perenne) pseudostems and blades (b) (black bars, Fennema; white bars, AberDove). (c, d) Expression of fungal α-mannosidase (MAN), cellulase (CEL), β-1,6-glucanase (GLU; data scale reduced 10-fold) and invertase (INV) in cultured fungal mycelia (left-hand bars) and in infected Lolium perenne plants (right-hand bars) (c), and L. perenne specific vacuolar invertase (vINV, black bars) and apoplastic/cell wall invertase (cwINV, white bars) in uninfected (left-hand bars) and infected (right-hand bars) plants (d). Different letters denote means that are significantly different. Vertical bars indicate ± SE. In endophyte-infected plant blades and pseudostems, the relative transcript abundance of mstN were significantly lower in the high sugar cv 'AberDove' than in the control cv 'Fennema' (in blades: F1,77 = 8.84, P < 0.01; in pseudostems: F1,77 = 4.18, P < 0.05); relative expression was also reduced in pseudostems compared with blades (F1,155 = 9.47, P < 0.01; Fig. 2b). The total water-soluble carbohydrate concentrations were significantly higher in 'AberDove' than in the control 'Fennema'; they were also significantly higher in pseudostems than in blades (Notes S2). It is possible that mstN is primarily important for sugar uptake at low carbohydrate availability (hence its relatively high expression under these conditions), while at high availability, other sugar uptake mechanisms might be activated, as was reported for Acremonium typhinum (syn. Neotyphodium typhinum) sugar uptake (Lam et al., 1994). The prevalent hypothesis for grass–Neotyphodium/Epichloë associations is that the endophyte takes up glucose and fructose derived from hydrolysis of sucrose by invertases in the apoplastic space in which the fungal hyphae reside (Thrower & Lewis, 1973; Lam et al., 1994). In culture studies by Lam et al. (1994), it was shown that A. typhinum invertase activity was induced by sucrose and the authors suggested that the majority of the high invertase activity seen in infected plant tissues is the result of fungal invertases. In support of this, we show here that relative expression of a putative fungal invertase identified in the E. festucae genome was higher in planta than in culture (F2,29 = 26.67, P < 0.0001; Fig. 2c). We also show that two of the three host plant invertases tested, a vacuolar and an apoplastic/cell wall invertase were significantly down-regulated in N. lolii-infected plants compared with uninfected plants (F1,39 = 19.15, P < 0.0001 for vacuolar invertase; F1,39 = 20.49, P < 0.0001 for cell wall invertase; Fig. 2d), while a cytosolic invertase was not affected. These results possibly indicate that the infected host plant can reduce sucrose hydrolysis (and thereby hexose availability in the apoplastic space), while the fungus has the ability to 'override' plant control of sucrose hydrolysis. However, as the genome of L. perenne has not been sequenced yet, we cannot exclude the presence of additional host invertases which might be regulated in a different manner. Based on the uptake characteristics of MSTN, particularly its high affinity for the cell wall carbohydrate mannose, we hypothesized that N. lolii might be able to hydrolyse polymeric plant cell wall carbohydrates as a supplementary source of C. The analysis of transcript abundance of fungal genes coding for putative cell wall carbohydrate hydrolases showed that relative expression of α-mannosidase, cellulase and β-1,6-glucanase (Bryant et al., 2007) was significantly higher in planta than in culture (F2,29 = 15.63, P < 0.0001 for mannosidase; F2,29 = 6.36, P < 0.01 for cellulase; F2,29 = 8.126, P < 0.005 for glucanase; Fig. 2c), which might suggest that N. lolii has the potential to hydrolyse cell wall carbohydrates. In this context it should be noted that a metabolite-profiling study demonstrated that N. lolii-infected L. perenne plants had reduced amounts of neutral detergent fibres, congruent with a reduction in cell wall hemicelluloses caused by carbohydrate hydrolysis (Rasmussen et al., 2008). In support of this, a recent transcriptome study comparing an E. festucae mutant with a wildtype strain in planta revealed that the disruption of a mitogen-activated kinase (sakA) resulted in a switch from mutualism to pathogenesis (Eaton et al., 2010). Proliferative branched growth of the fungus was accompanied by an up-regulation of fungal cell wall hydrolases. Even though electron micrographs do not show any major damage on cell walls of infected plants (Christensen et al., 2008), it is possible that cell wall degradation is only limited and localized, as is seen in biotrophic fungal infections (Münch et al., 2008). Carbohydrate exchange between host plants and symbiotic fungi is critical for a successful mutualistic association, and the present study contributes some new information on sugar transfer from grass hosts to endophytic Neotyphodium fungi. The apparent high affinity of MSTN for mannose and the high expression of putative endophytic cell wall hydrolases in planta present additional supportive evidence for the hypothesis of limited cell wall carbohydrate utilization in symbiotic plant–fungal associations, clearly calling for further investigations. This work was partially funded by a postdoctoral fellowship granted by the NZ Agricultural and Marketing Research and Development Trust (AGMARDT; grant #890) and the NZ Foundation of Research, Science and Technology (FoRST; C10X0203, PROJ 10333-ECOS-AGR). We thank Dr A. Schüßler (University of Munich, Germany) for kindly providing the sugar uptake-deficient S. cerevisiae strain EBY.VW4000. We also thank Dr M. Cao, S. Harrison and W. Simpson (AgR Grasslands, NZ) for technical assistance, and Drs R. Johnson and C. Jones (AgR Grasslands, NZ) as well as four anonymous referees for reviewing earlier versions of this manuscript. We acknowledge the Epichloë festucae Genome Project (Prof. C. Schardl, University of Kentucky, USA), which was funded by grants from NSF (EF-0523661), USDA-NRI (2005-35319-16141) and NIH (2 P2O RR-16481). Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office. Fig. S1 Sequence identity dendrogram of the MSTN protein and sugar transporters from various fungi. Table S1 Primer sequences used in this study for gene isolation and RT-qPCR Notes S1 Sequence analysis of MSTN. Notes S2 Water soluble carbohydrate concentrations in AberDove compared to Fennema and blades compared to pseudostems. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.