•Soil humidity and bulk water transport are essential for nutrient mobilization. Ectomycorrhizal fungi, bridging soil and fine roots of woody plants, are capable of modulating both by being integrated into water movement driven by plant transpiration and the nocturnal hydraulic lift.
•Aquaporins are integral membrane proteins that function as gradient-driven water and/or solute channels. Seven aquaporins were identified in the genome of the ectomycorrhizal basidiomycete Laccaria bicolor and their role in fungal transfer processes was analyzed.
•Heterologous expression in Xenopus laevis oocytes revealed relevant water permeabilities for three aquaporins. In fungal mycelia, expression of the corresponding genes was high compared with other members of the gene family, indicating the significance of the respective proteins for plasma membrane water permeability.
•As growth temperature and ectomycorrhiza formation modified gene expression profiles of these water-conducting aquaporins, specific roles in those aspects of fungal physiology are suggested. Two aquaporins, which were highly expressed in ectomycorrhizas, conferred plasma membrane ammonia permeability in yeast. This indicates that these proteins are an integral part of ectomycorrhizal fungus-based plant nitrogen nutrition in symbiosis.
Forests are of outstanding significance for mankind. Not only are they of great economic significance (wood production), but more importantly, they have a great share in carbon sequestration leading to substantial reduction in anthropogenic CO2 release. To guarantee an optimal activity of boreal and temperate forest ecosystems, the mutualistic interaction of tree roots with certain soil fungi (ectomycorrhiza) is essential (Smith & Read, 2008). This symbiosis enables the plant partner to optimize its nutrient acquisition and allows a privileged access to easily degradable carbohydrates for the fungus, thus paving the way for an efficient colonization of nutrient-limited environments, which is reflected by the dominance of ectomycorrhizal (ECM) forest ecosystems in the northern hemisphere.
Soil growing hyphae that explore litter or mineral layers for nutrients often constitute a large part of ECM fungal colonies. As ECM fungal mycelia can comprise up to 80% of the fungal and 30% of the total microbial biomass in forest soils (Wallander et al., 2001; Högberg & Högberg, 2002; Wallander, 2006), they are regarded as key elements of forest ecosystem processes (for recent reviews, see Read et al., 2004; Anderson & Cairney, 2007). Gadgil & Gadgil (1971, 1975) reported that the presence of ECM fungi in natural soils reduced decomposition of organic matter, presumably as a result of competitive or antagonistic interactions between ECM and saprotrophic mycelia. Although it has not been universally observed, Koide & Wu (2003) have provided further evidence for this ‘Gadgil’ effect and suggested that it might arise from competition for water between ECM and saprotrophic mycelia. Ectomycorrhizal fungal colonies can be spread over large areas (Dahlberg & Stenlid, 1990; Baar et al., 1994; Dahlberg, 1997; Anderson et al., 1998) and are directly linked to the fine root system of trees. Owing to transpiration, plant water demand is quite high and could be the driving force for soil water movement over considerable distances. Infected plant fine roots are covered by hyphae and the fungal symbiont is thus integrated into transpiration-driven bulk water movement in soil (Plamboeck et al., 2007). As extraradical hyphae are in tight interaction with soil particles, the fungal partner has access to pore water, which is especially valuable under drought conditions. In agreement with this, pine seedlings grown in drying soil were shown to be water-balanced for several weeks when associated with a well developed ECM system, while nonmycorrhizal controls died (Duddridge et al., 1980). Nevertheless, water transfer is a dynamic process and can also be directed from the plant towards the fungus. Querejeta et al. (2003) showed that ECM fungi were able to obtain water from deeper soil areas via their plant partners at night by a process called hydraulic lift. This water was further distributed in soil via mycorrhizal networks and did not only allow a survival of fungal hyphae but also improved nutrient mobilization under drought conditions. Additionally, water support of young seedlings by adult plants could be demonstrated via the hyphal network under drought conditions (Egerton-Warburton et al., 2007). Even if it is hard to generalize properties owing to the large heterogeneity of ECM fungi (Allen, 2007), it is commonly accepted that fungal mycelia modulate the water conductivity of soils (Augéet al., 2004).
Because cell walls of older hyphae often do contain hydrophobic wall layers (Allen, 2007), manipulation of membrane permeability at the entry (tips of soil growing hyphae) and exit points (hyphae of the symbiotic interface) of fungal colonies is essential for the directed water flux into and out of the fungal colony. With the exception of some bacteria, fast water transfer over membranes is thought to be mediated by integral membrane proteins termed major intrinsic proteins (MIPs) or aquaporins (Preston et al., 1992). It seems likely that these proteins can play a major role in water transfer in mycorrhizas, no matter in which direction water is transported.
Mycorrhizal symbiosis causes significant changes in aquaporin activity of host plants. Marjanovic et al. (2005b) could show that transcript abundances of two aquaporin genes were increased in poplar living in an established ECM symbiosis, while others were decreased. One of these aquaporins might be of special importance for mycorrhizal drought response, as it was the only one that was highly expressed under drought conditions but strongly repressed in nonmycorrhizal fine roots (Marjanovic et al., 2005a). In Medicago truncatulaUehlein et al. (2007) showed that two out of six aquaporin genes were up-regulated when the plant was associated with Glomus mosseae. Aroca et al. (2007) showed that arbuscular mycorrhizal (AM) formation influenced plasma membrane intrinsic protein (PIP) gene expression in Phasaeolus vulgaris roots, too.
To gain an insight into ECM fungal aquaporin function, we took advantage of the currently sequenced Laccaria bicolor genome (Martin et al., 2008). By genome-wide expression analysis and investigation of transport capabilities of selected proteins, we identified three members of the aquaporin gene family that are mainly responsible for water transport in this model fungus. Furthermore, we demonstrated ammonia permeability for certain aquaporins. Together with the strong expression of two of the corresponding genes in functional ectomycorrhizas, this finding is a further piece in the jigsaw suggesting that ammonium is a primary nitrogen source in ECM symbiosis.
Laccaria bicolor (Maire) P.D. Orton (strain S238N) was cultured on modified Melin Norkrans (MMN) medium (Marx, 1969; adapted according to Kottke et al., 1987), with 1% glucose at 18°C. Mycelia grown in liquid culture were harvested by filtration using a Büchner funnel under suction, washed twice with deionized water, frozen in liquid nitrogen, and stored at −80°C.
Populus tremula × tremuloides Mich. mycorrhizas were obtained under axenic conditions at 17°C according to Hampp et al. (1996) with ammonium as the sole nitrogen source (final concentration of 300 μM). Mycorrhizas and nonmycorrhizal fungal hyphae (extraradical mycelium) were harvested, frozen in liquid nitrogen, and stored at −80°C.
Expression analysis was performed by quantitative reverse transcription polymerase chain reaction (RT-PCR) in a MyiQ real-time PCR system (Bio-Rad, Hercules, CA, USA). Isolation of total RNA was carried out either according to Nehls et al. (1998) or by using the RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. RT-PCR was performed according to Selle et al. (2005). Specific primers for L. bicolor 18S rRNA (Fajardo-Lopez et al., 2008) were used as reference. PCR was always performed in duplicate using at least three independent cDNA preparations for analysis. PCR efficiencies, as calculated by the MyiQ software package (Version 1.0, Bio-Rad) were between 85 and 95%. To compare the transcript abundances of different members of the aquaporin gene family, dilution series of photometrically quantified DNA fragments of each gene and the reference were prepared and used as PCR template together with the first-strand cDNA samples.
Primers used for expression analysis were named according to the respective protein IDs in the L. bicolor genome database (v1.0 http://genome.jgi-psf.org/Lacbi1/Lacbi1.home.html): Lacbi1:247946: TGCTTCTCTCTTTGCGAC and CGTTGAGAGCCAAGAACAC; Lacbi1:392091: CTCGGTGCAGGCTTCTAC and GGGCCACTGGATCTGAAG; Lacbi1:233199: ATGGCGCTCCGACAAATG and AGTCTCGCGCAGGATTGAG; Lacbi1:317173: CATACATGACGCAAGCATC and CCCATTTGTCGGAGAGTG; Lacbi1:391485: TGCATCCCTTGGAATGG and CGGAAAGCGAATACCTG; Lacbi1:387054: ATGCACACGCCAACCAC and AACCTCTACGCCATTGCC; Lacbi1:307192: TCGCTCTTTATCTTCTTCC and AGTAGGGTCAAGAGCAGC.
Even when located within the coding region, primer pairs were chosen to be specific for each member of the gene family by primer sequence alignment to the L. bicolor genome sequence. To prove the amplification of the correct member of the gene family, purified PCR products originating from expression analysis were used for direct sequencing.
Cloning of the entire coding region of L. bicolor aquaporins
For heterologous expression, the entire coding regions of aquaporin genes were PCR-amplified from first-strand cDNA using Taq (Qiagen) or Phusion DNA polymerase (Finnzymes, Espoo, Finland) with the following gene-specific primers (named according to the respective protein IDs in the L. bicolor genome database): Lacbi1:392091: TCGATGCATCCACAAGTTGC and TGAATAGACAGGCATTAGACC; Lacbi1:247946: CAGTCTCATCATGAAGTTAACC and TTTCCTCAGACAGCCTCAG; Lacbi1:317173: CCGCTCCTAACATGTCTGGC and CTCTCTCAAACAACCTCAGC; Lacbi1:391485: CCATGGACGACAAATTCGAC and CTGGTGCAAATTTAAGCTGG; Lacbi1:387054: CTGGTCCAACATGTCTAACGC and GACAGAGTTCTCAAACATGGG; Lacbi1:307192: CTCCGTCATGTCCGCTACTC and CAGTGGCTTGCTCATACAGG; Lacbi1:233199: TAATGTTCTCGCCCGTCATG and TCCCTCAAACAACCTCTGCC.
PCR products were subcloned into plasmid vectors and sequenced using the ABI Prism Dye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, CA, USA) on an automated ABI 3130 sequencer (Applied Biosystems) according to the manufacturer’s instructions. A minimum of two independent cDNA clones was sequenced.
Heterologous expression of L. bicolor aquaporins in Xenopus laevis oocytes
The cDNA-inserts of Lacbi1:247946, Lacbi1:307192 and Lacbi1:387054 were initially cloned into the pJET1/blunt vector, (Fermentas, St. Leon-Rot, Germany) while Lacbi1:391485, Lacbi1:392091 and Lacbi1:317173 were inserted into pCR2.1 (Invitrogen). Linearized (Kpn2I/XbaI, Fermentas) fragments of Lacbi1: 247946, Lacbi1:307192 and Lacbi1:387054 were inserted into the XmaI (New England Biolabs, Beverly, MA, USA)/XbaI (Fermentas) digested pGEMHE-Vector (Liman et al., 1992). BamHI/XbaI fragments of Lacbi1:391485 and Lacbi1:317173 and a BamHI/HindIII (Fermentas) linerized Lacbi1:392091 cDNA were cloned into the pGEMHE-vector digested with the respective enzyme combination.
Complementary RNA (cRNA) was in vitro synthesized from purified PCR-products by T7 RNA polymerase using the mMESSAGE mMACHINE transcription kit (Applied Biosystems) according to the manufacturer’s instructions.
Either 50 nl ND96 buffer (negative control) or 50 nl cRNA (5 ng) of the different Laccaria aquaporins, rAQP1 (Rattus norvegicus) or PfAQP (Plasmodium falciparum) were injected into X. laevis oocytes (maturity stages V and VI). Water permeability was measured in osmotic oocyte swelling assays according to Beitz et al. (2006). For calculation of the osmotic water permeability (Pf, in μm s−1), the volume increase after 30 s was used. Pf-values were calculated according to Zhang & Verkman (1991). Solute permeability (PSol) was determined according to Hansen et al. (2002) by measurement of oocyte swelling in modified isoosmotic ND96 buffer containing 130 mM urea or glycerol instead of 65 mM NaCl. At least 20 replicates were made for every construct and the respective controls.
Heterologous expression of L. bicolor aquaporins in yeast
The PCR-amplified open reading frames of L. bicolor aquaporins were also inserted into the yeast expression vector pDR196 (Rentsch et al., 1995). Lacbi1:247946 and Lacbi1:307192 cDNAs were released with Kpn2I/PstI (Fermentas) and ligated into the PstI/XmaI (NEB, Frankfurt, Germany) digested pDR196 vector. Lacbi1:387054 was excised by XbaI/XhoI (Fermentas) digestion and introduced into BcuI/XhoI (Fermentas) linearized pDR196, while Lacbi1:391485 and Lacbi1:317173 were released by XhoI/BcuI (Fermentas) and Mph1103I/BcuI (Fermentas) digestion, respectively, and integrated into XhoI/BcuI respectively, Xhol /BcuI digested pDR196. Lacbi1:392091 was released from a pGEMHE construct by XmaI/PstI digestion and integrated into the XmaI/PstI linearized pDR196 vector. Yeast cells were transformed according to Gietz et al. (1995).
For methylamine toxicity assays, according to Beitz et al. (2006), overnight cultures of transformed BY4742Δfps1 (MATa his3-1 leu2Δ0 lys2Δ0 ura3Δ0 fps1::kanMX) cells were adjusted to an OD600 of 1 and 5 μl of each dilution series (1 to 10−3) were spotted onto pH-adjusted (pH: 5.5, 6.5 and 7.5) minimal medium agar plates either supplemented with 50 mM methylamine or not. The plates were incubated for up to 6 d at 30°C.
As methylamine is an artificial substrate, ammonium uptake experiments were also performed. Yeast (310Δfps1: mep1Δ mep2Δ mep3Δ ura3 fps1Δ, Zeuthen et al., 2006) transformants, expressing single L. bicolor aquaporins, were grown in 0.17% YNB (Difco), 2% glucose, supplemented with 1 mM NH4Cl. The pH of the growth medium was adjusted with 80 mM MOPS to pH 7.5. Fifty milliliters of medium were inoculated with cells from an overnight culture (pre-grown in YNB-medium containing 2% glucose and 0.1% arginine; start OD600 0.05). Cultures were incubated at 28°C under agitation and yeast growth (OD600) was monitored for up to 6 d.
To investigate urea conductivity, nitrogen-starved (pre-grown in YNB-medium containing 2% glucose and 0.01% arginine for 2 d) yeast (YNVW1: Δdur3 Δura3, Liu et al., 2003) transformants, expressing single L. bicolor aquaporins, were grown in YNB (Difco, Becton, Dickinson and Co, Heidelberg, Germany), 2% glucose, supplemented with 0.05 mM urea at pH 6 at 28°C under agitation. Yeast growth was monitored by OD600 measurement.
Construction of the phylogenetic tree
Sequence data were compared with gene libraries using BlastX (Altschul et al., 1997) and further analyzed using ClustalW (Thompson et al., 1997). The alignment was investigated by applying a Bayesian approach based on Markov chain Monte Carlo (MCMC) as implemented in the computer program MrBayes 2.02 (Ronquist & Huelsenbeck, 2003) to estimate phylogenetic relationships. Four incrementally heated chains with over 1.1 million generations were involved, starting from random trees and assuming a percentage of invariable alignment sites with gamma-distributed substitution rates of the remaining sites. Rather than specifying an amino acid substitution model, we allowed the Markov processes to sample randomly from the substitution models implemented in MrBayes. Trees were sampled every 200 generations, resulting in an overall sampling of 20 000 trees per run, from which the first 4000 trees of each run were discarded (burn-in). The remaining 12 000 trees sampled in each run were pooled and used to compute a majority rule consensus tree to obtain estimates of the posterior probabilities. Branch lengths were averaged over the sampled trees. Stationarity of the process was controlled using the Tracer software (Rambaut & Drummond, 2003), version 1.2.1.
Identification of potential aquaporins in the L. bicolor genome
By homology search using yeast and other fungal homologs as a source, seven open reading frames were identified in the L. bicolor genome, which could potentially encode aquaporins. The length of the deduced protein sequences ranged from 286 to 344 amino acids.
By using gene-specific primers based on genome data, the entire open reading frames of all identified potential aquaporins were PCR-amplified from cDNA and cloned into plasmid vectors. The amplified nucleotide sequences of Lacbi1:392091, Lacbi1:391485, Lacbi1:317173, Lacbi1:387054 and Lacbi1:307192 were in agreement with the predicted cDNA sequences of the JGI database (http://genome.jgi-psf.org/Lacbi1/Lacbi1.home.html). The annotated sequence of Lacbi1:247946, however, revealed a 21 bp insertion (position 130–151 in the JGI sequence) that was never found in PCR-amplified cDNA sequences.
For Lacbi1:233199, 16 individual clones were sequenced that frequently contained additional stop codons or introns as well as frame shifts so that we decided to exclude this gene from further analyses.
Phylogenetic analysis of L. bicolor aquaporins
When using Arabidopsis thaliana PIP1;1 as an outgroup, four main branches of the fungal aquaporin gene family can be distinguished (Fig. 1; the entire tree consisting of a total of 135 proteins are shown in Supporting Information, Fig. S1). In the first, the only so-called classical aquaporin-containing branch, one L. bicolor (Lacbi1:392091) homolog was found. Another three L. bicolor homologs (Lacbi1:391485, Lacbi1:317173, and Lacbi1:247946) belong to a second branch, called ‘Fps-like aquaglyceroporins’ as it includes all yeast-derived aquaglyceroporins (Van Aelst et al., 1991; Luyten et al., 1995). Pettersson et al. (2005) restricted the ‘Fps-like aquaglyceroporins’ to a group of aquaporins carrying a conserved regulatory N-terminal domain, which is exclusively found in yeasts. Nevertheless, as a number of nonyeast proteins are clearly more related to the Fps-like aquaglyceroporins from yeast than to other branches of the fungal aquaporin gene family, we will stick to the less stringent definition.
The third branch, comprising the remaining two L. bicolor aquaporins (Lacbi1:387054, Lacbi1:307192), was named ‘other aquaglyceroporins’.
Interestingly, Escherichia coli GlpF, a well investigated aquaglyceroporin, clusters together with two proteins of P. blankesleeanus between Fps-like aquaglyceroporins and ‘other aquaglyceroporins’ (Figs 1, S1), clearly suggesting both branches as aquaglyceroporins.
In silico analysis of the deduced Laccaria proteins
Aquaporins usually have six transmembrane helices, with N- and C-termini located inside the cytoplasm (Smith & Agre, 1991). The computer programs ConPred II (Arai et al., 2004) and HMMTOP (Tusnady & Simon, 1998, 2001) both predicted six transmembrane helices for the classical aquaporin and all Fps-like aquaglyceroporins, but seven transmembrane helices for the two ‘other aquaglyceroporins’ (Lacbi1:387054 and Lacbi1:307192). The N-terminus of all proteins was predicted to be located in the cytoplasm.
Two motifs of the pore-forming parts of aquaporins are of special interest for protein function, the asparagine-proline-alanine (NPA) motif and the so-called aromatic/arginine (ar/R) region.
Two connecting loops, B and E, of membrane-spanning helices have a hydrophobic character and thus fold back into the membrane. Each of the loops holds a characteristic NPA motif and both interacting motifs together form the central pore constriction (Murata et al., 2000). Only in the three ‘Fps-like aquaglyceroporins’ were both NPA motifs perfectly conserved, while the classical aquaporin (Lacbi1:392091, NPN and NSA) and the two ‘other’ aquaglyceroporins (Lacbi1:387054: NPC/NTA, and Lacbi1:307192 NPC/NSA) showed alterations. Alterations in NPA motifs are, however, quite common in fungi and changes found in the Laccaria aquaporins fit into the observed patterns (NPx and NxA; Pettersson et al., 2005).
The ar/R region is the narrowest constriction of the pore and alterations within this motif are expected to alter pore diameter and therefore account for pore selectivity (Beitz, 2005). Based on the well characterized AQP1 from Rattus norvegicus (Beitz et al., 2006), only position R195 (arginine) was conserved in all Laccaria aquaporins. With the exception of the classical aquaporin Lacbi1:392091 (altered into methionine) position F56 (aromatic amino acid) was also conserved. By contrast, position H180 (histidine) was only conserved in the classical aquaporin but was changed in all other L. bicolor proteins (glycine in ‘Fps-like’ aquaglyceroporins, isoleucine in the ‘other’ aquaglyceroporins). Position C189 (cysteine) was not conserved at all in Laccaria aquaporins (alanine in Lacbi1:392091, tyrosine in the ‘Fps-like’ aquaglyceroporins, and valine/isoleucine in the two ‘other’ aquaglyceroporins).
Together these changes offer some hints concerning the observed differences in water and solute transfer of the Laccaria proteins. However, whether these modifications of the ar/R region are responsible for the observed differences in water and solute permeability has to be further investigated in an appropriate model aquaporin (for example, AQP1).
Water permeability of L. bicolor aquaporins
All six L. bicolor aquaporins were investigated regarding their water permeability by heterologous expression in X. laevis oocytes (Fig. 2).
With permeability coefficients (Pf) between 10 μm s−1 (buffer-injected oocytes) and 275 μm s−1 (rAQP1; Rattus norvegicus), the values of the controls ranged within previously published results (Hansen et al., 2002; Beitz et al., 2009).
Lacbi1:317173 (Pf = 147 μm s−1) turned out to be the most efficient Laccaria water channel, but was only half as efficient as rAQP1. Lacbi1:391485 (Pf = 120 μm s−1) and Lacbi1:392091 (Pf = 62 μm s−1) followed, with decreasing water permeability, while Lacbi1:387054 (Pf = 30 μm s−1) and Lacbi1:247946 (Pf = 18 μm s−1) showed only low but still significant water conductivity. No water permeability at all was observed for Lacbi1:307192. However, as the protein content of heterologously expressed Laccaria aquaporins was not quantified, the observed water permeabilities of X. laevis oocytes have to be taken with some caution.
Urea permeability of L. bicolor aquaporins
All six L. bicolor aquaporins were examined for urea permeability by heterologous expression in X. laevis oocytes. To avoid misleading results because of an only limited water permeability of some of the examined Laccaria proteins, rAQP (which does not conduct urea and glycerol) was coexpressed with the respective channel of interest as described by Hansen et al. (2002). By using highly urea-permeable PfAQP from Plasmodium falciparum (Hansen et al., 2002) as a positive control (PSol = 0.53 μm s−1) and rAQP1 as a negative control (PSol = 0.0119 μm s−1), only Lacbi1:387054 (a predicted aquaglyceroporin) enabled a low but statistically significant (P =2.2 × 10−8) urea permeability (PSol = 0.08 μm s−1; Fig. S2). Owing to its low transfer rate in the oocyte system, complementation of the urea uptake deficient yeast strain YNVW1 (Liu et al., 2003) was applied to verify these result. Again, only Lacbi1:387054 restored, in a similar manner to the positive control, the growth defect of the yeast mutant (data not shown).
Glycerol permeability of L. bicolor aquaporins
Glycerol transfer is a mechanism for osmoprotection in yeasts (Tamas et al., 1999) that may also play a role in filamentous fungi. Hence, Laccaria aquaporins were tested for capability of glycerol conductance in the oocyte system. As described for urea permeability assays, rAQP (which does not conduct glycerol, PSol = 0.018 μm s−1) was coexpressed with the respective channels of interest.
Out of the six investigated members of the Laccaria aquaporin gene family, Lacbi1:391485 performed a strong glycerol conductance, which was only c. 2.5 times lower than that of the highly glycerol-permeable aquaporin PfAQP (PSol values of 0.34 and 0.88 μm s−1, respectively). Additionally, Lacbi1:387054 showed a rather weak (PSol value of 0.036 μm s−1) glycerol conductance (supported by only a weak statistical significance; P =0.08), while none of the other Laccaria aquaporins enabled glycerol permeability (Fig. S3).
Ammonium/ammonia permeability of L. bicolor aquaporins
NH4+/NH3 permeability of L. bicolor aquaporins was investigated in yeast by using two different assays, which gave similar results for most Laccaria aquaporins, except for Lacbi1: 247946.
For the investigation of methylamine conductance, yeast BY4742Δfps1 cells expressing L. bicolor aquaporins were used as a result of the high stringency of the test (Beitz et al., 2006). The test is based on the uptake of methylammonium (protonated form) into yeast cells by proton-coupled ammonium importers. Owing to the deletion of the methylamine transport facilitator Fps1 (an aquaglyceroporin) in the BY4742Δfps1 mutant, methylammonium will accumulate in the cytoplasm and yeast cells will die or stop growth. However, as soon as a functional methylamine transport facilitator is present in the plasma membrane (e.g. a L. bicolor aquaporin), methylamine can be released by yeast cells, which can be easily detected as yeast growth (Wu et al., 2008). The driving force for methylamine release is a pH gradient between cytosol (c. pH 7.2) and apoplastic space (pH 5.5 or 6.5 in the experiment). At physiological pH, a fraction of methylammonium becomes deprotonated in the cytosol (the ratio between protonated methylammonium and deprotonated methylamine is c. 2500 : 1). Methylamine, which is released by aquaglyceroporins, becomes immediately protonated in the apoplast (pH 5.5; ratio between methylammonium and methylamine is c. 125 892 : 1) and is thus continuously removed from the equilibrium. On agar plates, containing 50 mM methylammonium, untransformed BY4742Δfps1 cells, transgenic BY4742Δfps1 yeasts containing the empty pDR196 vector or expressing the classical aquaporin Lacbi1:392091 showed no growth at all (Fig. 3). By contrast, yeast cells expressing Lacbi1:387054, Lacbi1:391485 or Lacbi1:317173 were able to grow even better than the parental yeast strain (BY4742) containing yeast endogenous aquaglyceroporin Fps1 (Fig. 3). Lacbi1:247946 and Lacbi1:307192 enabled a clearly weaker yeast cell growth that was just above background level. The assay is expected to work best if the difference in cytosolic and apoplastic pH is high. In agreement with the transport of the deprotonated methylamine by Laccaria aquaglyceroporins, yeast growth was much weaker at pH 6.5 than at pH 5.5 (Fig. 4).
In most cases when channels were permeable for methylamine, they were also permeable for NH4+/NH3. However, methylammonium is an artificial substrate that does not necessarily reflect the capability of aquaporins to conduct NH4+/NH3. Therefore ammonium/ammonia uptake of 310Δfps1 yeast cells (Zeuthen et al., 2006) expressing L. bicolor aquaporin cDNAs was used to prove the methylammonium results (Fig. S4).
With the exception of Lacbi1: 247946, the results of the methylammonium assay were confirmed for all Laccaria aquaporins. Again, Lacbi1:392091-expressing yeast cells grew as poorly as the empty vector control, indicating no or rather weak NH4+/NH3 conductance. Lacbi1:387054-, Lacbi1:391485- and Lacbi1:317173-transformed yeast cells grew well, clearly indicating NH4+/NH3 conductance. According to the results of the methylamine assay, Lacbi1:307192-transformed yeast cells showed only a rather limited growth rate. In contrast to the methylammonium assay, where Lacbi1: 247946-transformed yeast cells revealed weak methylamine conductivity, Lacbi1: 247946 enabled no NH4+/NH3 growth.
Ectomycorrhiza formation and aquaporin gene expression
When compared with the extraradical mycelium, ectomycorrhiza formation considerably changed the transcript abundances of three aquaporin genes (Fig. 5). The transcript abundances of Lacbi1:317173 and Lacbi1:307192 were enhanced fourfold and twofold, respectively, upon ectomycorrhiza formation, while Lacbi1:387054 expression was reduced fourfold. Only minor (below twofold) or even no change in gene expression was observed for the other investigated members of the aquaporin gene family.
In RNA samples isolated from extraradical mycelium (grown on agar plates), the number of mRNA molecules varied for the different aquaporin genes by a factor of 100. Compared with 1 × 106 18S ribosomal RNA molecules (reference), nine mRNA molecules of Lacbi1:391485, four of Lacbi1:317173 and Lacbi1:307192, three of Lacbi1:392091, 0.2 of Lacbi1:387054 and 0.1 of Lacbi1:247946 were observed. The transcript abundances of all genes except Lacbi1:317173 were further reduced when hyphae were grown in liquid culture instead of agar plates.
The impact of osmotic potential of the growth medium on aquaporin gene expression
To investigate the impact of gentle changes of the osmotic potential of the growth medium on aquaporin gene expression, fungal mycelia were transferred from liquid MMN medium (0.2 MPa) to MMN media adjusted to 0.1 MPa (hypo-osmotic), 0.2 MPa (iso-osmotic) or 0.6 MPa (hyperosmotic) and gene expression was followed from 5 min to 2 d (data not shown). All Laccaria aquaporin genes displayed a similar temporary increase in gene expression under all osmotic regimes (including iso-osmotic conditions) within the first hour after medium exchange, which declined to the initial transcript abundance afterwards. As no difference in gene expression profiles was observed for iso-, hypo-, or hyperosmotic conditions, the observed temporary changes in aquaporin expression can be interpreted as a general cellular response to the exchange of the growth medium.
Impact of growth temperature on aquaporin gene expression
As the conductivity of lipid membranes for water is low at low temperatures, aquaporins might be of special importance keeping up water supply under these conditions. Thus, aquaporin gene expression was examined in fungal mycelia grown at different temperatures (5–33.5°C; Fig. 6). Compared with temperatures between 15 and 20°C, where similar growth rates were observed, fungal growth was reduced by 26% at 10 and 25°C and by 78% at 5°C. Nearly no fungal growth was observed at 30°C and higher. For all aquaporin genes, high expression levels were observed when fungal mycelia were grown at 15°C. Lacbi1:307192 and Lacbi1:391485 showed a comparable or even higher expression level when Laccaria mycelia were grown below 15°C, but transcript abundances declined at elevated temperatures. Transcript abundances of Lacbi1:317173, Lacbi1:247946 and Lacbi1:387054 were highest at 15°C and declined at any other investigated temperature. Only the classical aquaporin Lacbi1:392091 revealed similar high expression levels when grown between 15 and 33.5°C but a 50% reduction below 15°C.
All fungi investigated so far contain at least one aquaporin gene (Hill et al., 2004), with a clear tendency to higher numbers in filamentous fungi (e.g. one in Saccharomyces castellii, five in Fusarium gramineum; Pettersson et al., 2005). In agreement with this and the fact that the ECM model fungus L. bicolor shows enlargement of many multigene families (Martin & Selosse, 2008; Martin et al., 2008), seven putative aquaporin genes were identified. But why do filamentous ECM fungi with a high surface-to-volume ratio seem to be dependent on a sufficient large number of water and solute channels?
Plasma membrane permeability for water
Low temperatures lead to a tighter package of lipids and thus to lowered membrane water permeabilities. Accordingly, there are two ways to overcome water transfer limitations of membranes: modification of lipid constitution and/or increase of adequate water-permeable channels in the membrane. In boreal but also many temperate forest ecosystems, soil temperature (below 10 cm depth where ectomycorrhizas are frequently found) rarely exceeds 16°C with a frequent average between 5 and 10°C (Carreiro & Koske, 1992; Davidson et al., 1998; Davies-Colley et al., 2000). Low soil temperatures together with a large water flux from the fungus towards the plant and vice versa, driven by plant respiration and the nocturnal hydraulic lift, respectively, can explain the necessity of aquaporin function in ECM fungi.
Three out of six investigated L. bicolor aquaporins (the classical one and two Fps-like channels) enabled water permeabilities of X. laevis oocyte plasma membranes that are strong enough to be of physiological impact. Interestingly, the highest oocyte water permeabilities were mediated by the Laccaria Fps-like aquaglyceroporins, while the classical aquaporin Lacbi1:392091 was only half as efficient. One explanation for this phenomenon could be seen in a modification of NPA-motifs in Lacbi1:392091, which is unusual for a classical aquaporin and does not occur in the Fps-like aquaglyceroporins.
All aquaporins, inducing high water permeability of X. laevis oocytes, also revealed high transcript abundances in mycelia grown in liquid culture at 15°C. At other temperatures, however, their expression pattern showed clear differences. While the classical aquaporin Lacbi1:392091 was expressed at a constant level at temperatures ≥ 15°C, its transcript abundance was reduced at lower temperatures. By contrast, Lacbi1:391485 showed exactly the opposite pattern while the third water channel (Lacbi1:317173) revealed a strong increase in gene expression only at temperatures of c. 15°C.
Protein function and gene expression together clearly indicate that these three aquaporins are of great impact for water permeability of the fungal plasma membrane. Furthermore, the complementary, temperature-dependent expression profiles of Lacbi1:392091, Lacbi1:317173 and Lacbi1:391485 can point to a temperature-dependent optimization of plasma membrane function by these aquaporins. Soveral et al. (2006) suggested that water transfer in yeasts is mediated by aquaporins mainly at temperatures < 23°C, while direct membrane transfer dominates at higher temperatures. The strong expression of Lacbi1: 391485 at ≤ 15°C, together with high water permeability of X. laevis oocytes expressing the respective protein, may thus indicate that Lacbi1:391485 is key to fungal plasma membrane water permeability at low temperatures.
In addition to temperature, ectomycorrhiza formation was also affecting aquaporin transcript abundance. The mycorrhiza-dependent strong increase in gene expression of Lacbi1:317173, the most effective water channel in X. laevis oocytes indicates an outstanding impact of this protein on water permeability of fungal plasma membranes in symbiosis. However, even when not affected at the transcriptional level, the overall high expression of Lacbi1:392091 and Lacbi1:391485 also points to an important function of the respective proteins in ectomycorrhizas.
Modulation of gene expression has been shown to be the main regulatory level for controlling protein activity for a number of aquaporins (Luu & Maurel, 2005). However, even when not yet shown for filamentous fungi, post-translational modification of aquaporins (affecting their permeabilities) was demonstrated for certain plant, mammal, and yeast aquaporins (Nemeth-Cahalan et al., 2004; Tornroth-Horsefield et al., 2006; Mollapour & Piper, 2007; Zhang et al., 2007). Furthermore, the swelling behavior of X. laevis oocytes does not necessarily reflect the situation in the fungal plasma membrane. Thus, for final conclusions about the impact of selected Laccaria aquaporins on plasma membrane water conductivity, gene inactivation must be performed in future experiments to prove the postulated aquaporin function.
Solute permeability mediated by L. bicolor aquaporins
Glycerol transfer via aquaglyceroporins has been shown to be a mechanism for osmoprotection in yeasts (Tamas et al., 1999) and may also play a role in ECM fungi. However, substantial glycerol permeability is only conferred by Lacbi1:391485 and glycerol accumulation has never been observed in ECM fungi. Therefore, it can be speculated that glycerol permeability of Lacbi1:391485 is a relict of the affiliation of the protein to the group of aquaglyceroporins, and that other osmotic active compounds (e.g. trehalose or mannitol) have taken over yeast glycerol function in ECM fungi.
An essential part of ECM function that is still widely unknown is the mechanism behind the fungal nitrogen support of its plant host. Recently, ammonium came into focus as a potential primary source of plant nitrogen in ECM symbiosis (Selle et al., 2005; Chalot et al., 2006). In AM fungi (Bago et al., 2001; Govindarajulu et al., 2005; Cruz et al., 2007), arginine catabolism (via arginase and urease activity) is believed to be involved in generating intracellular ammonium in hyphae of the symbiotic interface. Genes coding for proteins essential to the catabolic part of the ornithine cycle were also detected in the genome of the ECM fungus L. bicolor and turned out to be expressed in ectomycorrhizas (Lucic et al., 2008; Martin et al., 2008). Furthermore, in ectomycorrhizas of different ECM fungi, including L. bicolor (Martin et al., 2008), urease gene expression turned out to be strongly increased.
While data concerning ammonium concentrations in the fungal cytosol and the common apoplast of the plant–fungus interface of ectomycorrhizas are lacking, ammonium concentrations in the mM range were frequently observed in hyphae of saprotrophic fungi (Burgstaller, 1997). Because ECM poplar plants strongly induce certain high-affinity ammonium importer genes (Selle et al., 2005; Couturier et al., 2007), net export from the fungal cytosol into the common apoplast can thus be driven by an ammonium concentration gradient. As shown for animals and plants, ammonia-permeable aquaporins have the potential to facilitate a nitrogen transfer over the plasma membrane. Ammonia flux, mediated by a member of the aquaporin superfamily, has recently been demonstrated for another symbiotic interface, the symbiosome membrane that encloses nitrogen-fixing bacteroids in root nodules (Hwang et al., 2010). Therefore, Laccaria aquaporins were investigated for their ammonia/ammonium permeability.
A number of researchers (Beitz et al., 2006; Zeuthen et al., 2006; Wu et al., 2008) report that the pH-dependent growth of BY4742Δfps1 (methylammonium experiment) and 310Δfps1 (ammonium experiment) yeast cells expressing single aquaporin genes can indicate whether methylamine/ ammonia (NH3) or methylammonium/ammonium (NH4+) transfer is mediated by the respective protein. Regarding the artificial substrate methylammonium, the main argument for methylamine export is a pH-dependent concentration gradient between cytosol (pH 7.2) and apoplast, which is c. 50 times at pH 5.5 but only five times at pH 6.5. As aquaporins can only facilitate a concentration gradient-driven export, the growth of yeast cells expressing a functional facilitator should be better at a lower external pH. Indeed, BY4742Δfps1 yeast cells expressing single Laccaria aquaporins grew much better at pH 5.5 than at pH 6.5. In contrast to methylamine, the methylammonium concentration does not differ significantly at the respective pH values. Furthermore, because of the high extracellular methylammonium concentration, methylammonium transport by aquaporins (which has never been shown) would result in a cellular methylammonium concentration of c. 50 mM. Even higher cytosolic concentration would be necessary to generate an outward directed gradient. The observed pH-dependent changes in the growth of Laccaria aquaporin-expressing yeasts thus strongly support methylamine export. Furthermore, to our knowledge, only one aquaporin capable of channeling charged molecules has yet been characterized (AQP6, selective for NO3−; Yasui et al., 1999; Ikeda et al. 2002), while the majority of aquaporins are restricted to a transfer of uncharged molecules (like methylamine or NH3).
Both Fps-like aquaglyceroporins (Lacbi1:391485, Lacbi1:317173), and one ‘other aquaglyceroporin’ (Lacbi1:387054) clearly showed NH3/NH4+ permeability in yeast cells. In addition, Lacbi1:387192 (the second ‘other aquaglyceroporin’) also mediated a weak ammonia (but no methylamine) permeability. As the transcript abundances of Lacbi1:391485 and Lacbi1:317173 were high in ectomycorrhizas (together comprising up to 65% of aquaporin transcripts), it can be speculated that the respective proteins are responsible for the supposed ammonia release by fungal hyphae at the plant/fungus interface of ectomycorrhizas.
As claimed by Chalot et al. (2006), ammonium export by hyphae at the plant/fungus interface only works if the fungal ammonium reuptake and assimilation are low. Otherwise, the activity of ammonium retrieval systems would result in an energy-demanding ammonium cycling across the membrane. Down-regulation of glutamine synthetase gene expression in ectomycorrhizas (the main ammonium assimilatory pathway) has been shown for Paxillus involutus (Wright et al., 2005) and Amanita muscaria (U. Nehls, unpublished). Additionally, Willmann et al. (2007) reported the mycorrhiza-specific repression of a fungal ammonium importer in A. muscaria. However, expression data from Laccaria (whole-genome oligoarray analyses) indicated a strong up-regulation of some putative ammonium importers in ectomycorrhizas (Lucic et al., 2008). As a closer analysis of function and kinetic properties of the respective transporters is still missing, a final assessment of these data is prevented. Thus, further investigations are necessary to unravel the function of ammonium as nitrogen source for host plant nutrition in ectomycorrhiza symbiosis.
In addition to ammonium, urea, which is an intermediate of arginine degradation (Govindarajulu et al., 2005), could be a potential plant nitrogen source in symbiosis. The (low) urea permeability of the aquaglyceroporin Lacbi1:387054 would make urea permeability of the fungal plasma membrane possible. At any rate, urea, liberated within fungal hyphae, is expected to be degraded by urease activity (resulting in an ammonium release). Indeed, an increased urease gene expression was observed in ectomycorrhizas formed with different ECM fungi, including L. bicolor (Martin et al., 2008). This fact, the low urea permeability of Lacbi1:387054, and the low (and decreasing even further upon ectomycorrhiza formation) transcript abundance of the corresponding gene lead to the conclusion that aquaporin-mediated urea leakage by L. bicolor hyphae can be ignored at the plant/fungus interface.
In summary, for the first time a comprehensive study of a basidiomycotic aquaporin gene family, covering gene expression as well as protein function, was accomplished. Two aspects of aquaporin function were the focus of this investigation: water and solute permeability. While nearly all L. bicolor aquaporins mediated water permeability of X. laevis oocyte plasma membranes, only three proteins revealed reasonably high rates of physiological significance. Protein function (in X. laevis oocytes) and gene expression data indicated that three proteins (Lacbi1:317173, Lacbi1:391485, and Lacbi1:392091) are key in relation to water permeability of the Laccaria plasma membrane. Moreover, ammonia permeability of Lacbi1:317173 and Lacbi1:391485 can be involved in fungal nitrogen (ammonium) support of the plant host in symbiosis. To prove the postulated function of Laccaria aquaporins in water and solute transport, suppression of aquaporin gene function has to be addressed in future experiments.
We are indebted to Margret Ecke and Andrea Bock for excellent technical assistance and to Dr Imke Bodendiek for critical reading of the manuscript. We also want to thank the anonymous reviewers for their helpful suggestions to improve the manuscript. This work was financed by the German Science Foundation (NE 332/11-1).