Phosphorous Availability Influences the Dissolution of Apatite by Soil Fungi

Geobiology (2007), 5, 265–280 DOI: 10.1111/j.1472-4669.2007.00107.x

PhosphorousBlackwellFungalORIGINAL dissolution Publishing ARTICLES of Ltdapatite availability inﬂuences the dissolution of apatite by soil fungi A. ROSLING,1 K. B. SUTTLE,1 E. JOHANSSON,2 P. A. W. VAN HEES2 AND J. F. BANFIELD1 1Department of Earth and Planetary Science, University of California, Berkeley, Mc Cone Hall, Berkeley, CA 94720-4767, USA 2Man-Technology-Environment Research Centre, Department of Natural Sciences, Örebro University, SE-701 82 Örebro, Sweden

ABSTRACT

Apatite (Ca10(PO4)6(OH,F,Cl)2) is the primary inorganic source of phosphorus in the biosphere. Soil fungi are known to increase plant-available phosphorus by promoting dissolution of various phosphate minerals. Yet no apatite dissolution studies exist using fungi as weathering agents, and regulation of fungal weathering activity in response to different levels of phosphorus availability is largely unknown. Fungi were isolated from a grassland soil in northern California. Three pathways of tri-calcium phosphate

(Ca3(PO4)2) (TCP) dissolution in liquid culture were identified among biogeochemically active fungi: (1) acidification (pH 3.3 ± 0.16), (2) moderate acidification (pH 4.9 ± 0.11) and (3) no acidification. Isolates representing pathway 1 and 2 were Zygomycetes in the order of Mucorales. All non-acidifying isolates in pathway 3 were

Ascomycetes and cleared the media by altering TCP into hydroxyapatite (Ca10(PO4)6(OH)2) and sequestering it within mycelial spheres. One isolate representing each pathway was used in fluorapatite dissolution experiments either with the fungi present or under abiotic conditions using cell-free liquid media conditioned by fungal growth at different phosphorus and calcium availabilities. Both Mucorales isolates acidify their substrate when growing in the presence of phosphorus. Mucorales exudates were mainly oxalic acid, and conditioned cell-free media with phosphorus induced fluorapatite dissolution at a rate of 10–0.9±0.14 and 10–1.2±0.22 µmol P m–2 s–1. The ascomycete isolate on the other hand, induced fluorapatite dissolution at a rate of 10–1.1±0.05 µmol P m–2 s–1 by lowering the pH of the media under phosphorus-limited conditions, without producing significant amounts of low molecular weight organic acids (LMWOAs). Oxalate strongly etches fluorapatite along channels parallel to [001], forming needle-like features, while exudates from the ascomycete-induced surface rounding. We conclude that while LMWOAs are well-studied weathering agents, these do not appear to be produced by fungi in response to phosphorus-limiting growth conditions.

Received 27 October 2006; accepted 20 February 2007

Corresponding author: Anna Rosling. Tel.: +46-18-67 18 64, Fax: +46-18-67 35 99; e-mail: [email protected].

In vitro, phosphorus-solubilizing fungi and bacteria INTRODUCTION produce a clear zone around their growing colony when After nitrogen, phosphorus is the most frequently limiting grown on solid media enriched with a precipitated calcium macronutrient for plants. In soil, free phosphorus is rapidly phosphate. This in vitro solubilization method has been used immobilized by sorption onto minerals or hydrous oxides of for the last hundred years as an initial criterion by which to Al, Fe and Mn. To survive in soil, roots and microbes need identify isolates with the potential to release phosphorus from mechanisms to increase phosphorus availability and facilitate minerals (Whitelaw, 2000). Detection of fungal clearing is uptake (Raghothama, 1999). Many soil fungi are able to facilitated in liquid cultures since mycelial cover does not solubilize phosphorus directly from minerals such as apatite obscure detection of clear zones, but the method is not

(Ca10(PO4)6(OH,F,Cl)2). Both plant-symbiotic and free-living suitable for large-scale screening. Microbial solubilization of soil fungi have been demonstrated to increase phosphorus phosphorus in liquid cultures has been demonstrated to uptake in plants through their biogeochemical activity depend linearly on the induced acidiﬁcation of the growth (Wallander et al., 1997; Whitelaw, 2000). substrate (Gupta et al., 1994). This acidiﬁcation results from

266 A. ROSLING et al. microbial release of organic acids, proton efflux over the between 3 and 5, organic dissolution rates exceed inorganic plasma membrane, and the formation of carbonic acid in the (Welch et al., 2002). In general, dissolution of apatite in biotic media from respiratory CO2 production (Burford et al., 2003). systems is slower than abiotic dissolution, but accelerates as Apart from lowering the pH, organic acids form strong microbial biomass and organic compound concentrations complexes with cations, thereby increasing phosphorus mineral increase (Welch et al., 2002). Although phosphorus-solubilizing dissolution rates. This makes organic acids more effective activity in fungi is well documented, apatite dissolution exper- weathering agents than inorganic acids at the same pH (Gadd, iments with fungi as weathering agents are missing. 1999). Fungal-induced phosphorus solubilization from phosphate Biochemical, morphological, and physiological responses to minerals appears to depend on the amount and composition phosphorus deficiency in plants have been well studied. In of organic acids produced by the fungi (Whitelaw, 2000). dicots, increased production and excretion of organic acids A substantial body of literature exists on the production of and chelators can increase P availability from inorganic organic acids by filamentous fungi, largely because Aspergillus sources. In all plants, the induction of both intracellular and niger is used for the commercial production of citric acid extracellular acid phosphatases production releases P from (Magnusson & Lasure, 2004). When grown under conditions organic sources (see Raghothama, 1999). In yeast and of excess carbon relative to other nutrients (e.g. phosphorus), filamentous fungi, regulation of genes encoding membrane P some fungi maintain a high carbon flux through their mycelia transporters and phosphatases in response to P deficiency has by exuding excess carbon as metabolic intermediates such as also been demonstrated (Oshima, 1999; Tasaki et al., 2004a, b). citric and oxalic acid (Gadd, 1999; Gallmetzer & Burgstaller, Yet, regulation of biogeochemical activity in fungi in relation 2002). Substrate acidification and organic acid production in to phosphorus limitation is poorly understood. Gibson and fungi are strongly affected by the source of nitrogen supplied Mitchell (2004) found no differences in phosphorus solubili- in the growth media. Response patterns to ammonium and zation by four ericoid mycorrhizal fungi grown on different nitrate sources vary in different fungi (Gadd, 1999). This levels of available phosphorus. Gharieb (2000) studied difference can be avoided by providing an organic nitrogen dissolution of gypsum by A. niger and found that dissolution source, for instance L-alanine, to the growing fungi occurred only when phosphorus was supplied to the substrate. (Mahmood et al., 2001). The mycelial growth of ectomycorrhizal fungi has been The biogeochemical significance of organic acid production demonstrated to increase in response to low phosphorus in fungi has been extensively studied with respect to both availability as a result of increased carbon allocation from the bioremediation of toxic metals (Gadd, 1999) and mineral host plant (Ekblad et al., 1995). weathering above and below ground (Burford et al., 2003). This paper examines patterns of apatite dissolution under Apart from substrate acidification and organic acid production, conditions induced by three biogeochemiclly active soil fungi. fungi may induce mineral weathering through nutrient uptake We describe how these fungi regulate weathering in response and physical force (Burford et al., 2003), production of to different phosphorus and calcium availabilities. Rotating extracellular organic polymers mediating processes at the liquid media batch experimental systems were used to measure mineral surface (Barker & Banfield, 1996), siderophore solubilization of phosphorus from apatite. Fungal weathering production (Callot et al., 1987) and cation biosorption to in such systems is limited to substrate acidification and exchange sites in the cell wall (Marschner et al., 1998). Further- activities of fungal exudates, since physical contact between more, the hyphal growth mode of filamentous fungi is well hyphae and minerals is limited. However, by determining adapted for exploitation of nutrient sources in the highly hyphal cation exchange capacities (Marschner et al., 1998) heterogeneous soil environment (Robson, 1999). The ability under different growth conditions, the current study provides of mycelia to connect distant mineral and carbon sources insight into how cell wall composition is regulated in response enables translocation of heterogeneously distributed nutrients to phosphorus-limiting conditions. In soil, fungal weathering and moisture through the mycelia (Hirsch et al., 1995). This may be controlled by processes at the hyphal surface as well as makes fungi potentially important agents of weathering and by the activity of exuded compounds. nutrient movement in the soil system (Finlay & Rosling, 2006). MATERIAL AND METHODS Although apatite is the predominant form of mineral phosphorus in the Earth’s crust, biological forms (i.e. dental Identification of biogeochemically active soil fungal isolates and skeletal apatite) have received most attention in kinetic studies. Dissolution experiments with fluorapatite under Fungi were isolated from soil and weathered rocks collected in geologically relevant conditions demonstrate that rates are 20- to 45-cm-deep soil pits in a meadow at the Angelo Coast strongly pH dependent, with increasing dissolution at decreasing Range Reserve in Mendocino County, California (39°44′N, pH for pH values between 2 and 5 (Guidry & Mackenzie, 123°39′W). Using the soil plate method (Parkinsson, 1994) 2003). At pH 2, inorganic dissolution of apatite is faster than with three different solid media: 2% malt agar, Czapek-Dox organic dissolution via acetate or oxalate. As pH is raised (Parkinsson, 1994), and 1/2 MMN (Marx, 1969), all with

Fungal dissolution of apatite 267

1.5% agar and 200 mg L–1 ampicillin. Plates were incubated in products from ANG3 were aligned using CLUSTALW at EMBL darkness at +16 °C and pure cultured twice within two weeks. (Thompson et al., 1994) to produce a consensus sequence After which, all 350 isolates were transferred to 2% malt for the isolate. Sequences were submitted to GenBank and agar without ampicillin and grouped by visual appearance. obtained the accession numbers DQ914418–31. Representative isolates were named ANG for the Angelo Reserve and numbered 1–42 (originating from soil) and Composition of crystals associated with the mycelia of 43–75 (originating from rock). biogeochemically active isolates Dissolution of tri-calcium phosphate (TCP) on solid (1.5% White powder was observed to accumulate within spherical agar) P- and Ca-free media (MES) (Mahmood et al., 2001) mycelia during TCP dissolution. The mineral composition of indicated biogeochemical activity in fungal isolates. Of 75 the powder was examined in three replicates of isolates ANG3, isolates, 14 cleared a zone around their mycelia. These were ANG55 and ANG74 grown for 12 days in 150 mL of further examined in 100 mL liquid MES with 4 g L–1 of MES-L with 0.375 g of TCP. Growth media pH was determined D-glucose (MES-L) with 0.25 g TCP in 250 mL E-flasks. in 5 mL liquid media samples collected every third day starting Mycelia growing on MES-TCP plates were used as inoculum at inoculation. Accumulated powder was separated from (6 pieces, 2 × 2 mm) and compared to non-fungal controls the mycelia at harvest and dried overnight at +60 °C. X-ray (no inoculum). Flasks were incubated on a rotary shaker table diffraction (XRD) analysis of collected powder and samples of (100 r.p.m.) at room temperature in darkness and visually TCP as purchased, autoclaved and from non-fungal control inspected over 2 weeks, after which pH of the media was flasks was carried out at room temperature with a X’Pert PRO measured (pH meter, Model 225, Denver Instrument, diffractometer (PANalytical, Almedo, The Netherlands), Göttingen, Germany) and drying the mycelia overnight at using CuKα radiation for an angle range of 0 to 120° 2θ, with +60 °C determined the dry weight (DW). Non-fungal a step size of 0.0167° for 60 s at each step. Spectral analysis controls maintained the initial pH 5.8. was performed using X’Pert HighScore Plus to identify peaks All fourteen isolates cleared their growth media within and compare to a mineralogical peak and profile database 2 weeks and are henceforth referred to as biogeochemically (X’Pert Database32). Furthermore, crystals associated with active isolates. Clearing of TCP proceeded along three differ- the mycelia of all 14 biogeochemically active isolates were ent pathways in relation to the measured growth parameters identified by XRD. Dry mycelia were manually ground, mycelial DW and media pH. Pathway 1: Isolates ANG 3, 4, 28 packed in a quartz sample holder (1-mm-deep central well, and 52 cleared the solution within one week, lowered the pH 1.5 cm diameter), and analyzed using the XRD described to 3.3 ± 0.2 and produced a final mycelial DW of 0.38 g ± 0.02. above. Collecting a spectra from 0 to 80° 2θ, with a step size Pathway 2: Isolates ANG 51 and 55 cleared the solution within of 0.0167° for 40 sec at each step. No XRD analysis was 2 weeks, lowered the pH slightly (4.9 ± 0.1) and produced a performed for mycelia of isolates ANG4 and ANG37. final mycelial DW of 0.23 g ± 0.01. Pathway 3: Isolates ANG 31, 34, 37, 45, 61, 70, 71 and 74 cleared the solution within Experiment 1: Long-term fungal growth using fluorapatite as 1 week without changing the pH (5.6 ± 0.2) while producing a source of P and Ca a final mycelial DW of 0.32 g ± 0.04. One isolate was selected to represent each of the three pathways, 1: ANG3, 2: ANG55 The three isolates, ANG3, ANG55 and ANG74, were grown and 3: ANG74. for 36 days in MES-L with fluorapatite as the only source of Biogeochemically active isolates were identified by DNA phosphorus and calcium to examine how patterns of TCP sequencing of the ITS region. DNA was extracted using the clearing correlated to fungal fluorapatite dissolution. Acid-washed Power Soil DNA Kit (MoBio Laboratories, Carlsbad, CA, E-flasks were autoclaved with 150 mg of fluorapatite in each. USA) from fresh or dried mycelia from liquid cultures. The Fluorapatite (batch 1) was prepared according to Welch et al. ITS region was amplified by polymerase chain reaction (PCR) (2002). We used the size fraction 150–500 µM with a surface using the universal primers ITS1 and ITS4 (White et al., area of 0.027 m2 g–1 (Welch et al., 2002). XRD analysis 1990). PCR products (50 µL) were purified using QIAquick confirmed that the mineral was fluorapatite but XRD spectra PCR purification Kit (250) (QIAGEN Science, Montgomery also contained additional peaks at 3.10 and 3.02Å. This minor County, MD, USA) before sequencing with ITS1 at the impurity was visible as occasional dark flakes among fluorapatite UC Berkeley DNA Sequencing Facility. PCR products from particles. MES-L (150 mL) was added to all flasks before two isolates (ANG3 and 28) were transformed and cloned inoculation with one of the three isolates using nine pieces (TOPO TA cloning kitpCR 2.1, Invitrogen, Carlsbad, CA, USA) (2 × 2 mm) of mycelia cut from the edge of fungal mycelia before sequencing. The software 4PEAKS (www.mekentosj.com, growing on MES-TCP plates. Three replicate flasks were NL) was used to edit obtained sequences and to compare them inoculated for each isolate, plus three inoculum-free controls. with sequences in the GenBank database at NCBI using the Flasks were sealed with aluminium foil and parafilm and BLAST program (Altschul et al., 1997). Eight sequenced PCR incubated for 36 days at room temperature, in darkness on a products from ANG28 were identical. Fifteen sequenced PCR rotary shaker table (100 r.p.m.).

268 A. ROSLING et al.

Mycelial growth, pH and elemental composition of the Eq. 1 Apatite IN = Sample OUT + Liquid END + media Mycelia END + Mineral END A 6 mL sample of the growth medium was taken at inoculation and subsequently every third day throughout the The budget presents amounts of P and Ca in the three experiment. From each sample, 5 mL was frozen for elemental determined pools (underlined) as a percentage of the total analysis and 1 mL was used to measure pH. To compensate for P and Ca in the known pool Apatite IN. Amounts of P and sampling, 30–40 mL of new media was added on days 12 and Ca in Apatite IN were calculated from the amount of ﬂuora-

24. Approximate reaction volumes were monitored by mass of patite added, assuming the structure formula Ca10(PO4)6F2 the flasks after sample removal and media addition. A total of (M = 1008.604). Continuous sampling removed both P and 15 samples were collected from each flask during the Ca from the systems (Sample OUT). These removed amounts experiment. Mycelial DW in each flask was determined at the were calculated from measured concentrations multiplied by end of the 36-day growth period. Fluorapatite particles were sampled volumes. The concentrations in missing samples rinsed twice in MilliQ water (Millipore, Billerica, MA, USA), (days 30 and 33, see above) were approximated based on an dried at +60 °C for 2 h, and stored at room temperature. The assumption of linear increase from days 27 to 36. At the end concentrations of Ca, Fe and P were determined by ICP-OES of the experiment, amounts of P and Ca in the liquid media (Optima 5X00™ DV ICP-OES, PerkinElmer, Boston, MA, (Liquid END) were calculated from measured concentrations USA) after acidification of the samples by diluting 1:2 at day 36 multiplied by the sample volume prior to final using 4% HNO3 in MilliQ water. The ICP standard was 2% sampling. P and Ca accumulated in fungal mycelia (Mycelia –1 –1 HNO3 in MilliQ water with 2.15 g L MES and 2.0 g L END) were calculated from the average concentrations per D-glucose, using a range of standards based on preliminary milligram of mycelia of three mycelial digests, multiplied by analysis of all samples. All samples from day 33, and two the mass of mycelia (DW) in each flask. Amounts of P and replicates from each isolate on day 30 were lost due to Ca in fluorapatite and secondary minerals at the end of the equipment failure. experiment (Mineral END) were not determined. Secondary minerals were identified by XRD but abundances were not Elemental composition of the fungal mycelia quantified. Three replicate subsamples (25 mg) of dry mycelia were The overall weathering rate in experiment 1 was calculated ° digested overnight in refluxing HNO3 at 130 C in 50 mL for each treatment based on the total amount of P released digestion tubes, following the procedure of Zarcinas et al. from apatite (Sample OUT + Liquid END + Mycelia END) (1987). Three blanks and three controls with 100 mg of normalized to the initial surface area of fluorapatite particles certified plant standard material (NIST 1547) were included and expressed over the entire growth period (36 days). The in the digest. The concentrations of Al, Ca, Fe, K, Mg, Na and rates are an average of three replicates ± SEM. P in the extracts were determined by ICP-OES using 16.7% HNO as a blank and a range of standards based on pre- 3 Experiment 2: Short-term regulation of fungal growth and liminary analysis of all samples. Concentrations in original activity in response to P and Ca availability plant and mycelial material were calculated after subtracting measured average concentrations of the blanks. The accuracy The growth and activity of isolates ANG3, ANG55 and of the extraction and concentration measurements was ANG74 were examined in relation to the availability of P and evaluated by comparing obtained values from reference plant Ca in the liquid growth media. Four different treatments were material to its certified content. Measured concentrations for established: (+Ca) calcium enrichment (4.3 mM (0.48 g L–1 + most elements were reproducible among replicates of the CaCl* 2H2O)), ( P) phosphorus enrichment (2.6 mM –1 + reference plant material with a relative standard deviation of (0.45 g L K2HPO4)), ( CaP) calcium and phosphorus 2.4% for Ca, 2.7% for P, 3.1% for Mg, 3.9% for Al, 4.0 for K, enrichment (4.3 mM Ca and 2.8 mM P, as above), and (+A) and 4.2 for Fe. Measurements for Na were more variable at calcium and phosphorus enrichment via mineral fluorapatite 11.8%. Extraction and detection were satisfactory in plant addition (1 g L–1). Each treatment was replicated three times. reference material for Ca, Fe, K, Mg and P with 109%, The second fluorapatite batch was prepared as above and this 93%, 96%, 110% and 113% efficiency relative to reference time XRD analysis demonstrated that the fluorapatite was concentrations. Since Al average efficiency was only 78% and pure. Acid-washed 250 mL E-flasks with 150 mg fluorapatite Na was as high as 189%, these elements were not included in were autoclaved. To each flask 150 mL MES-L was added further analyses. before inoculating and incubated for 12 days as described for experiment 1. Effects on the growing fungi, fungal biomass Phosphorus and calcium budget for the 36-day day and mycelial cation exchange capacity (CEC), and sub- experiment strate acidification were examined in all treatments and the For experiment 1, a weathering budget was constructed for dissolution activity and composition of fungal exudates were Ca and P with the pools described in Equation 1. examined in three treatments (+P, +Ca and +A).

Effects of the growing fungi dissolution batch experiments were performed using cell-free All isolates grown in +CaP were rust colored, suggesting that iron liquid media from treatments +Ca, +P, and +A. Batch was associating with the mycelia. While ANG3 and ANG74 experiments were set up in 250 mL Stericup (Millipore) by were white in all other treatments, ANG55 was rust colored in sterile filtering (0.22 µm) 300 mL of bulked media and all treatments except +P. To test for fungal regulation of hyphal storing at +4 °C. Liquid dissolution treatments were named by cell wall composition in response to P and Ca availability, mycelial combining growth media treatment (Ca, P or A) and fungal cation exchange capacity (CEC) was analyzed according to isolate number and non-fungal control (3, 55, 74, no). Marschner et al. (1998) except that mycelia were rinsed twice Fluorapatite powder (batch 2) was autoclaved and added in MilliQ water between exchange solutions and inverted for under sterile conditions to each liquid media reactor before 5 min for extraction at each step. Mycelial DW was determined attaching the lid. Stericup lid had been autoclaved after after extraction. The concentrations of Ca, Fe, K, Mg, Na, P and inserting a 10 cm plastic tube though a hole in the lid. A S of the second exchange solution were determined using ICP- 10 mL syringe with a 25 mm, 0.2 µm Syringe filter (Fisherbrand, OES with 0.05 M HCl as a blank for a range of standards based Fisher Scientific, San Jose, CA, USA) was attached to the outer on the preliminary analysis of all samples. Concentrations in end of the tubing. Reactors were covered in aluminum foil and blanks were two orders of magnitude lower compared to sample incubated on a shaker table (100 r.p.m.) for 24 days (+Ca) or concentrations for all elements and samples except K in treatment 23 days (+A and +P) at room temperature (17–24.5 °C). (+Ca). Cations other than Ca contribute to less than 10% of Using the attached syringe a total of 24 (+Ca) and 23 (+A and the total cation concentration in all samples but ANG74 in +P. +P) 4 mL samples were collected. The first day samples were Mycelial CEC is calculated from the Ca concentration in exchange collected every hour starting 1 min after apatite addition. The solution 2 and expressed as µmol Ca per g of mycelia DW. time between sampling occasions was successively increased throughout the experiment. Samples were stored at +4 °C. Organic acids, siderophores and pH in fungal exudates Sample volumes were determined from the weight of tubes The pH of the growth media was determined in 5 mL samples before and after adding the sample. The pH of each sample was of the liquid media collected before inoculation, every third measured before acidifying with a drop of concentrated HNO3. day of the experiment and at harvest. Liquid media was bulked The concentrations of Ca and P were then determined by ICP- from all replicate flasks of treatments +Ca, +P, and +A after OES. Concentration of Fe was determined at start only. The

12 days of fungal growth. A 50 mL sample was frozen and ICP standard was set up in 2% HNO3 in MilliQ water blank stored at –20 °C for analysis of composition of low molecular with 4.3 g L–1 MES buffer and 4 g L–1 D-glucose, using a range weight organic acids (LMWOAs). Particles were removed from of the standards based on preliminary analysis of all samples. the electrolyte by filtration through a 0.45 µm cellulose acetate Reacted fluorapatite particles were rinsed twice in MilliQ membrane (Millipore). The LMWOAs were determined by water, dried at +60 °C for 2 h and stored at room temperature. capillary zone electrophoresis (CZE) method as described by Initial P and Ca concentrations were different in all batch (Dahlén et al., 2000). Analysis was performed on a Agilent reactors since cell-free media from different treatments were 3DCE capillary electrophoresis system (Agilent Technologies, used and these had been modified by the preceding fungal Santa Clara, USA) (HP 3D capillary electrophoresis) instrument growth. Dissolution of P is illustrated by the accumulated equipped with a UV-detector (254 nm), using 75 µm (i.d.) amount (mmol) of released P (Fig. 5). The ionic activity fused silica capillary and hydrostatic injection. Separation of 12 product (IAP) was calculated from the fluorapatite formula different LMWOAs was possible: acetate, butyrate, citrate, Ca10(PO4)6F2, according to Guidry and Mackenzie (2003) formate, fumarate, lactate, malate, malonate, oxalate, propionate, and the solution saturation state, Ω, was calculated by dividing succinate and shikimate. Solutions containing 1–100 µM of pure IAP by the solubility constant of fluorapatite (Ksp = 3.26 10–60) LMWOAs were used as standards. (Stumm & Morgan, 1996) (Fig. 6). For these calculations, Hydroxamate siderophores were quantified as equivalents of fluorine concentrations were assumed to follow P concentra- deferriferrioxamine B (Sigma-Aldrich, St. Louise, MO, USA) tions in a stoichiometric manner. according to Raaska et al. (1993) except that samples were not Because apatite dissolution conditions change over time in concentrated and the standard (0, 0.1, 0.5, 1, 2.5 and 5 mM) was batch experiments, P release is non-linear (Welch et al., 2002). set up in filtered MES at pH 3. Continuous spectra were recorded Thus, incremental P release from fluorapatite was calculated between 350 –750 nm, 7, 18 and 30 min after mixing equal from the increase in dissolved P concentrations for each volumes of CAS reagent and sample by vortex for 10 s. Concen- sampling time point. Dissolution rates were calculated for the trations were estimated from the decreasing peak Raaska et al. timespan between each sampling and normalized to the initial (1993) defined as the average absorbance at 640–650 nm. surface area of fluorapatite particles and plotted against average pH for that timespan. The log rate was calculated for Dissolution activity of fungal exudates the time of linear increase in P concentration and compared To test for weathering capacity of fungal exudates produced to literature data for similar conditions (Welch et al., 2002) in response to different P and Ca availability, fluorapatite (Fig. 7). The time of linear increasing P concentrations was

0–100 h for A_74 (R2 = 0.99) and Ca_74 (R2 = 0.98). For solution after 1 week, the solution pH was 3.6 ± 0.12 after ANG3, P concentrations increased linearly in A_3 between 12 days and 40 ± 1 mg of mycelia had formed, from which 0–30 h (R2 = 0.97) and in P_3 between 0–80 h (R2 = 0.97). 2 ± 0.7 mg of powder was collected. ANG74 cleared the Media conditioned by ANG55 resulted in a linear P release solution within 3 days and maintained pH of 5.5 ± 0.07. It from fluorapatite between 20–200 h for A_55 (R2 = 0.94) formed 24 ± 2 mg of mycelia, from which 8 ± 0.7 mg of and between 0–125 h for P_55 (R2 = 0.84). powder was collected. ANG55 lowered pH to 5.2 ± 0.06 and did not clear TCP entirely from the solution within 12 days. It formed 5 ± 1 mg of mycelia, from which 7 ± 1 mg Characterization of reacted fluorapatite of powder was separated. XRD analysis of collected powder For the fluorapatite treatment (+A), dried mineral fractions revealed the presence of three different minerals: monohydrate from one replicate flask of each inoculation (ANG3, 55, 74 (whewellite: Whe) and dihydrate (weddelite: Wed) calcium and no fungi) from both experiments 1 and 2 were examined oxalate and hydroxyapatite (HAp). Mycelia of ANG3 con- by scanning electron microscopy (SEM), Hitachi S-5000 tained Whe, Wed and HAp, ANG55 contained Wed and HAp SEM. Fluorapatite particles were mounted on SEM stubs and ANG74 contained only HAp. Spectra from the powder using carbon tape and coated with approximately 6 nm collected from ANG74 were distinct from that of all analyzed platinum using a Bal-Tec MED020 Coating Sputter (Bal-Tac, TCP samples and matched all peaks in the hydroxyapatite Balzars, Lichtenstein, Germany). Dried mineral fractions from reference spectra. a second replicate flask from both experiments 1 and 2 were Mineral composition of biogeochemically active isolates used for XRD-based mineral identification. Dried fluorapatite differed between the three pathways and followed the same particles were ground by hand to a fine powder and packed in pattern as their representative isolates. Calcium oxalate crystals, the quartz dish described above and analyzed by XRD. Whe and Wed, were associated with mycelia of all isolates representing pathway 1 but not with any isolates representing pathway 3, which were only associated with HAp. Isolates Statistics representing pathway 2, were associated with both HAp and Whe. Statistical analysis was performed using Minitab® Statistical Software (Minitab Inc., State College, PA, USA). For experiment Experiment 1: Long-term fungal growth using fluorapatite as 1, one-way ANOVA with Tukey pairwise comparison was used a source of P and Ca to analyze mycelial element concentration of K, Ca, Fe, Mg and P in mycelial digests and % Ca and % P in the pools, Sample All three fungal isolates grew well on fluorapatite as a sole OUT, Liquid END, Mycelia END and the total P dissolution source of P and Ca. ANG3 formed dense white mycelia with rates (log µmol P m–2 s–1), for 36-day weathering budgets. an average DW of 186 ± 13 mg. ANG55 formed large, loose For experiment 2, General Linear Model with the factors fungi and white mycelia (DW 94 ± 2.3 mg). ANG74 had a more (ANG3, 55 and 74) and treatment (+A, +Ca, +CaP and +P) dense growth form, forming beige mycelial spheres, that and Tukey pairwise comparison were used to analyze CEC, greyed and darkened over time (DW = 153 ± 1.8 mg). All DW and pH. Linear regression and R2 values were calculated isolates lowered the pH of their growth substrate (Fig. 1A). in Excel. All presented values represent averages of three Inoculation with ANG74 resulted in a fast pH decrease which replicates, ± standard error of the mean (SEM). stabilized below 4 between days 12–24, after which pH increased again. Inoculation with ANG3 resulted in a slower pH decrease, but ultimately leveling out at ~pH 2.8, the RESULTS lowest levels measured in this experiment. ANG55 resulted in the slowest overall pH decrease, reaching 3.5 by the end of the Characterization of biogeochemically active fungal isolates experiment (Fig. 1A). Sequencing of biogeochemically active fungal isolates As pH decreased in the media, P became detectable in the demonstrated that all isolates representing TCP dissolu- solution at pH 4.5 for both ANG3 (day 12) and ANG55 (day tion pathways 1 and 2 (i.e. strong acidification and moderate 18) and at pH 4 for ANG74 (day 9). P concentrations acidification, respectively) were Zygomycetes in the order of increased linearly with decreasing pH in the pH range 4.5–3, Mucorales. Species identification of representative isolates for ANG3 (R2 = 0.95) (days 12–21) and ANG55 (R2 = 0.88) ANG3 (DQ914418) and ANG55 (DQ914427) were not (days 18–36). This linearity was less clear for ANG74 possible within GenBank. All isolates representing pathway 3 (R2 = 0.76) (day 6–21). Increasing pH during the second half (i.e. no acidification) were found to be Ascomycetes. The of the experiment with ANG74 did not effect P concentra- representative isolate ANG74 (DQ914431) was identified as tions in the media. These were stabile at 0.2 mM. P remained belonging to the family Trichocomaceae. below 0.02 mM in non-fungal controls (Fig. 1B). Ca concen- In the study of mycelial accumulation of crystals during trations increased to 0.4 mM by day 3 in all flasks including the 12 days of TCP dissolution in MES-L, ANG3 cleared the control. For ANG3, Ca concentrations remained slightly

Table 1 Elements (mg g–1) in fungal mycelia of ANG3, 55 and 74, grown for 36 days on apatite as a source of P and Ca

Element ANG3 ANG55 ANG74

Ca (mg g–1) 105.4 ± 26.3 (ab) 174.0 ± 9.6 (a) 54.7 ± 20.0 (b) P (mg g–1) 49.4 ± 9.7 52.3 ± 3.4 29.9 ± 9.2 Ca/P 2.1 ± 0.1 3.3 ± 0.04 1.8 ± 0.1 K (mg g–1) Not detected 1.1 ± 0.1 6.4 ± 0.4 Fe (mg g–1) 1.3 ± 0.1 3.0 ± 0.7 1.8 ± 0.3 Mg (mg g–1) 0.4 ± 0.05 0.55 ± 0.27 0.44 ± 0.02

Values are average of three ± SEM. Different letters indicate signiﬁcant differences (Tukey pairwise comparison).

with ANG74 increased to a maximum of 1.4 mM on day 21. The calculated Ca/P ratio in mineral fluorapatite is 1.67. All inoculated treatments maintained a lower Ca/P compared to control (Fig. 1C). A Ca/P ratio below 1 was maintained in liquid media after day 15 for ANG3 and after day 21 for ANG55 while ANG74 maintained a Ca/P ratio of 4 or below from day 12. Among replicates of ANG74, the Ca/P ratio of the media increased with higher pH towards the end of the experiment. The amount of Ca, Fe, K, Mg and P in mg element per gram of mycelial dry weight was determined at the end of experiment 1 (Table 1). Ca concentrations were highest and least variable in ANG55 and significantly lower in ANG74, with high variation between replicates. Ca/P ratios showed low variability within fungal type. Mycelial iron concentrations were highest in ANG55 compared to the other isolates but differences were not significant.

P and Ca budget for the 36-day experiment Fluorapatite was the only source of P and Ca in experiment 1 and corresponded to 892 ± 1.3 µmol of P and 1487 ± 2.1 µmol of Ca. Throughout experiment 1, P and Ca were released from fluorapatite to the solution where they were either taken up by the mycelia, remained free in the solution, or precipitated as secondary minerals. The % P and Ca respectively in each pool relative to added apatite is presented in Table 2. Total P dissolution rate (log µmol P m–2 s–1) over 36 days differed significantly by fungal isolate (P < 0.0001). Rates were the same at 10 × –3.6 µmol P m–2 s–1 for ANG3 and ANG74, while ANG55 resulted in significantly faster dissolution at 10 × –3.3 µmol P m–2 s–1 (Tukey P < 0.05). All fungal treatments resulted in significantly faster P dissolution compared to the non-fungal control at 10 × –5.0 µmol P m–2 s–1 (Tukey P < 0.05). Iron concentrations did not appear to affect phosphorus Fig. 1 Changes over 36 days in (A) pH, (B) concentration of P (mM) and (C) log concentrations in liquid samples collected over 36 days. Ca/P ratio, in experiment 1, for the fungal isolates ANG3, ANG55 and ANG74 and the no fungal control (no fun). In (C) the line 1.67 represents the stoichiometric Ca/P ratio in fluorapatite. Values present average of three Experiment 2: Short-term regulation of fungal growth and replicates ± SEM. activity in response to P and Ca availability below those in the control throughout the experiment. In Mycelia growth and composition in response to different flasks with ANG55, Ca concentrations were comparable to H1X P and Ca availability those of the control until pH decreased below 4, from which Mycelial DW at harvest (day 12) was significantly affected concentrations increased linearly. Ca concentrations in flasks by both fungal isolate (P < 0.0001) and mineral amendment

Table 2 P and Ca budgets for experiment 1. The pools Sample OUT (OUT), Liquid END (Liq END) and Mycelia END (Myc END) are expressed as % of introduced apatite

% P % Ca

OUT Liq END Myc END OUT Liq END Myc END

ANG3 4.0 ± 0.2 (a) 3.7 ± 2.2 (a) 5.7 ± 1.2 3.4 ± 0.6 (a) 1.1 ± 0.4 (a) 4.4 ± 1.2 (ab) ANG55 2.4 ± 0.2 (b) 20.0 ± 2.2 (b) 5.3 ± 0.4 2.9 ± 0.2 (a) 8.0 ± 0.8 (b) 6.5 ± 0.4 (b) ANG74 5.8 ± 1.4 (a) 3.2 ± 0.3 (a) 3.3 ± 1.1 13.5 ± 1.8 (b) 7.6 ± 0.2 (b) 2.2 ± 0.9 (a) No fun 0.2 ± 0.0 (b) 0.3 ± 0.0 (a) 4.5 ± 1.4 (a) 2.2 ± 0.2 (a)

Treatments (ANG3, 55 and 74) and the non-fungal (No fun) control. Letters indicate signiﬁcant differences within pools (columns) as determined by Tukey pairwise comparison.

Fig. 2 Mycelial dry weight (DW) in mg for isolates ANG3, ANG55 and ANG74 when grown for 12 days in MES-L enriched by one of the following: mineral Fig. 3 Mycelial cation exchange capacity expressed as mmol Ca per g of fluorapatite (A), both calcium and phosphorus (CaP), phosphorus (P) or calcium mycelia DW. Values present average of three replicates ± SEM. The fungal (Ca). Values present average of three replicates ± SEM. isolates ANG3, ANG55 and ANG74 were grown for 12 days in MES-L enriched with one of the nutrient amendments: mineral fluorapatite (A), calcium (Ca), treatment (P < 0.0001), with a significant interaction between both calcium and phosphorus (CaP) or phosphorus (P). Stars (*) indicate significantly higher values (P < 0.05) as determined by Tukey pairwise the two factors (P < 0.0001) (Fig. 2). On average, ANG3 comparison. produced significantly greater biomass than ANG55 (Tukey P < 0.05) and ANG74 (Tukey P < 0.05). In turn, ANG74 formed greater biomass than ANG55 (Tukey P < 0.05). and treatment (P < 0.0001) with a significant interaction Accumulated biomass was significantly lower in +Ca conditions between the two factors (P < 0.0001) (Fig. 4). Growth of than in other treatments (Tukey P < 0.05). ANG74 grew most ANG3 resulted in significantly lower pH than the other on +A, forming 118.3 ± 5.5 mg DW, which was not signi- isolates and non-fungal control (Tukey P < 0.05) and growth ficantly different from its growth on +P (Fig. 2). of both ANG55 and ANG74 resulted in pH lower than the The mycelial CEC was significantly affected by both fungal non-fungal control (Tukey P < 0.05). isolate (P < 0.0001) and treatment (P < 0.0001), and there was ANG3 represented isolates in pathway 1, which dissolved a significant interaction between the two factors (P = 0.025) TCP within a week through strong acidification of the media. (Fig. 3). Over all treatments, CEC of ANG55 was higher than When grown with soluble P (+P or +CaP), ANG3 caused a that of both ANG3 (Tukey P < 0.05) and ANG74 (Tukey rapid pH decrease (Fig. 4A). When P was supplied as fluorapatite P < 0.05). Treatment +Ca gave the overall highest CEC (+A), acidification was slower but ultimately reached a (Tukey P < 0.05) compared to all other treatments. None of comparable value around pH 3. ANG55 represented isolates the other treatments were significantly different. Based on Tukey in pathway 2, which dissolved TCP through moderate pairwise comparison, only mycelial CEC for ANG3 and ANG55 acidification of the media. At harvest, media pH for ANG55 in treatment +Ca was significantly higher (Tukey P < 0.05) was significantly lower (Tukey P < 0.05) under conditions than other combinations of fungi and treatments (Fig. 3). of soluble P amendment than under Ca and fluorapatite addition (Fig. 4B). ANG74 represented isolates in pathway 3, Substrate conditioning in response to different P and which cleared TCP from MES-L without acidifying the media. Ca availability For this isolate, pH showed significant decrease in the +Ca Substrate acidification, measured as pH at harvest (day 12), treatment (Tukey P < 0.05) (Fig. 4C). Fluorapatite addition was significantly affected by both fungal isolate (P < 0.0001) resulted in a less pronounced pH decrease, reaching

Table 3 Concentration (µM) of LMWOAs in bulked cell-free liquid media used for ﬂuorapatite dissolution in experiment 2

Reactor Oxalate (µM) Malonate (µM) Malate (µM) Acetate (µM)

A_no 5.0 Ca_no 9.5 P_no 11.7 A_3 3044.1 520.3 Ca_3 553.7 217.1 P_3 6460.4 645.8 116.7 A_55 18.0 3.6 14.4 Ca_55 5.4 2.4 5.6 P_55 96.6 1.7 36.7 A_74 4.0 6.6 2.6 Ca_74 4.1 6.7 2.1 4.2 P_74 7.0 7.4 28.8

Sample names combine the letter of treatment (A, Ca or P) with isolate number (no, 3, 55 or 74).

concentrations were detected in treatment +P where malate was also present. Growth of ANG55 produced the same organic acids as ANG3 but at lower concentrations, with the highest concentration when ANG55 was grown in +P. Growth of ANG74 resulted in low concentrations of LMWOAs. In contrast to the two other isolates, ANG74 produced acetate (2–30 µM) (Table 3). The CAS assay indicated the presence of hydroxamate siderophores only in media collected from ANG3 grown in +P. The concentration in P_3 was estimated to be equivalent to 2 mM deferriferrioxamine B.

Fluorapatite dissolution in cell-free liquid media The impact of exudates from each fungal isolate on fluorapatite dissolution was analyzed in twelve reactors with different cell-free liquid media. Five of these (Ca_74, P_3, P_55, A_3 and A_74) resulted in more than 50 µmol of P being released from fluorapatite (Fig. 5). These five reactors all had an initial pH of 4 or below (Fig. 4). At the end of the experiment, P release was leveling out in reactors with media Fig. 4 The change in substrate pH over time as a result of 12 days of growth in experiment 2, by the fungal isolates (A) ANG3, (B) ANG55 and (C) ANG74, conditioned by ANG74 and even decreasing in reactors with depends on the P and Ca availability in the media. MES-L was enriched by media conditioned by ANG3 or ANG55 (Fig. 5). Even in mineral fluorapatite (A), calcium (Ca), phosphorus (P) or both calcium and reactors in which initial pH was above 5 (i.e. reactors P_74, ± phosphorus (CaP). Values present average of three replicates SEM. Ca_no and A_55), P was released from apatite, in total 23, 19 and 3.7 µmol P, respectively. P was released in reactor Ca_no after a dramatic pH decrease (5.3–3.6) between days 3–6, 4.0 ± 0.03 at harvest, which was still lower compared to probably due to contaminations of this reactor. treatments to which soluble P was added (Tukey P < 0.05). Of all reactors, only the solution in reactor A_74 demon- No substrate acidification was observed in any treatment for strated a Ca/P ratio close to the stoichiometric ratio of non-fungal controls (data not shown). Lower pH was fluorapatite (1.67). The reactor Ca_74 released Ca and P to correlated to higher mycelial DW over all treatments for the solution at a ratio close to 1. In reactors with media ANG3 (R2 = 0.94) and for ANG55 (R2 = 0.90). No such conditioned by ANG3 or ANG55, Ca concentrations were correlation was apparent for ANG74. low or even undetectable throughout the experiment, no Ca/ The LMWOAs oxalate, malonate, malate and acetate were P ratios could thus be calculated. XRD analysis of the dried identified (Table 3) in cell-free liquid media collected after mineral fraction after dissolution demonstrated that calcium 12 days of inoculation. Growth of ANG3 resulted in mM oxalate crystals formed during fluorapatite dissolution in concentrations of oxalic acid and malonate. The highest reactors P_3, A_3 and P_55. The saturation state of the

Fig. 5 Accumulated P (mmol) over time (thousands of s), during ﬂuorapatite dissolution in cell-free media conditioned by fungal growth. Liquid dissolution treatments are named by combining growth media treatment (Ca, P or A) and fungal isolate number (3, 55, 74). Only reactors that released more than 50 µmol of P from apatite are included in the ﬁgure.

Fig. 6 Calculated saturation state (IAP/Ksp) of cell-free liquid media over time (thousands of s) with respect to fluorapatite. Reactors with detectable levels of both Ca and P are included. The solution is saturated (bold line = 1) with respect to fluorapatite when IAP = Ksp, (log Ksp = –60). solution (Fig. 6) with respect to fluorapatite was calculated their respective pH, the log dissolution rates of all reactors based on solution concentrations of Ca and P as well as except A_55 are above those documented by Welch et al. extrapolated F concentrations. Based on these calculations, (2002) in abiotic dissolution of fluorapatite using 1 mM oxalic the solution of reactor A_74 became slightly oversaturated acid at different pH (Fig. 7). There is a negative correlation over the course of the experiment while reactor Ca_74 became (R2 = 0.85) between iron concentrations and pH over all oversaturated at an early stage of the dissolution experiment cell-free media at the start of the dissolution experiment, when (Fig. 6). Incremental dissolution rates were highest for excluding the media from P_3 where siderophores were reactors conditioned by the zygomycete ANG3 when grown detected. Subsequently, higher initial Fe concentrations in both +P and +A and by the zygomycete ANG55 when correspond with higher incremental rates. However, we have grown in +P (Fig. 7), followed by reactors conditioned by the no indications that Fe concentrations effected P solubility in ascomycete ANG74 when grown in +Ca. Dissolution rates the reactors. decrease with increasing pH for all reactors conditioned by ANG74 (rate = –0.34*pH –0.16, R2 = 0.58) as well as for Reacted fluorapatite particles ANG55 (rate = –2.03*pH +7.25, R2 = 0.86). Dissolution rate for exudates from ANG3 are independent of pH. Over all, Scanning electron microscopy demonstrated that mineral there is a strong correlation between the documented particles from the non-fungal control in both experiment 1 incremental rate and initial pH of the reactor (R2 = 0.73). At and 2 were smooth (Fig. 8I). Particles reacted with ANG3

Fig. 7 Incremental rate of fluorapatite dissolution versus pH determined in cell-free liquid media reactors. Linear representation of correlation data for fluorapatite dissolution rate are taken from Welch et al. (2002) and were determined at different pH in 1 mM NaCl, 1 mM acetate or 1 mM oxalate. had no visible apatite surface since secondary crystals covered external Ca concentrations (Casarin et al., 2003). Never the all particles (Fig. 8A,B). This was true both for particles less, their production induces mineral dissolution (Arvieu et al., reacted for 12 and for 36 days. The crystals were identified as 2003) and may be considered a passive weathering pathway. whewelite by XRD. Etching of the apatite surface was evident On the other hand, fungal substrate acidification in on particles reacted with ANG55 for 12 days (Fig. 8C). response to phosphorus-limited growth conditions, as exem- Particles exposed to ANG55 for 36 days were strongly etched plified by the ascomycete isolate ANG74 suggests an active along channels parallel to [001] apatite, resulting in a needle pathway for P deficiency-induced mineral weathering in some like surface structure (Fig. 8D,E). After 36 days particles were fungi. Substrate acidification by ANG74 is independent of partially covered in secondary crystals identified as whewelite mycelial biomass with low pH (Fig. 4C) at both the lowest and by XRD, no crystals were observed on particles exposed to the highest produced biomass (Fig. 2). Similarly, substrate ANG55 for 12 days. Circular precipitates were observed on acidification by the basidiomycete Mycena galopus was found fluorapatite particles reacted with ANG74 for 12 days (Fig. 8F), to be independent of mycelial density when tested in response and sharp etching was observed along channels parallel to to different mineral amendments (Rosling et al., 2004). For [001] (Fig. 8G). After 36 days ANG74 resulted in a bumpy ANG74 less substrate acidification is observed when P is surface structure (Fig. 8H). No secondary mineral phases added as mineral fluorapatite, compared to conditions with no were detected by XRD on particles reacted with ANG74. P (Fig. 4C). This is either a result of protons being consumed in the dissolution process or a result of a feedback mechanism reducing the acidifying activity of ANG74 when P is released DISCUSSION from the fluorapatite. Interestingly, P deficiency-induced substrate acidification is not a result of fungal production of Active and passive weathering pathways in fungi LMWOAs (Table 3). However, Ascomycetes are known to This paper examines fungal growth and exudation responses produce a variety of LMWOAs (Magnusson & Lasure, 2004), to different levels of phosphorus availability and provides new but growth conditions may be critical for fungal production of insight into passive and active weathering pathways in fungi. organic acids. Gharieb (2000) observed that the ascomycete The biogeochemical importance of LMWOAs produced A. niger dissolved gypsum only when phosphorus was available by fungi has been extensively studied. In plants, the carbon in the substrate. Passive weathering pathways are thus not metabolism is affected by phosphorus deficiency resulting in restricted to Zygomycetes. an increased exudation of organic acids (Ryan et al., 2001). In Biological weathering may act mainly on short distances, the current study LMMOAs were mainly produced by the requiring microbial colonization of mineral surfaces (Banfield zygomycete isolates ANG3 and ANG55 that acidify their et al., 1999). Fungal modification of mycelial surfaces when growth media as a consequence of mycelial growth. This faced with P deficiency may be an important mechanism to suggests that LMWOAs are not produced in response to induce fluorapatite dissolution by ligand–surface interactions. phosphorus-limiting growth conditions but rather to control Alternatively, the modified fungal surface may bind dissolved

Fig. 8 SEM images of fluorapatite surfaces reacted in the presence of fungal isolate: ANG3 for 36 days (A,B), with ANG55 for 12 days (C) and for 36 days (D,E) and with ANG74 for 12 days (F,G): for 36 days (H) and no fungal control for 36 days (I). Scale bars correspond to 50 µm in A, 20 µm in D, 5 µm in B, E, H and 2 µm in C, F, G, 10 µm in I. ions in solution, altering the saturation state and thus promoting 1990). The range of CEC values in this study are comparable dissolution. Increased amount and modified composition of to those in Marschner et al. (1998). The isolates that did not extracellular polymeric substances associated with marine regulate their substrate acidity (i.e. ANG3 and ANG55) in diatoms have been demonstrated in response to P deficiency. response to P deficiency were instead found to increase their These changes have been suggested to ameliorate the effect of surface CEC under P-limited conditions (Fig. 3). The fact that P deficiency as well as salinity stress (Abdullahi et al., 2006). ANG55 was rust colored in all treatments except +P suggests The metal-binding capacity of fungal cell walls has been that regulation of iron-binding compounds (i.e. siderophores) suggested to serve as a cation reservoir as well as a screen to on the mycelia surface is an important mineral response protect the fungi from toxic levels of metals (Siegel et al., mechanism in this fungus.

Secondary mineral formation Fungal dissolution of ﬂuorapatite

Crystals of calcium oxalate are commonly associated with Normalizing for active biomass is a common experimental fungi under a range of growth conditions (Verrecchia et al., problem in microbial dissolution experiments (Kalinowski et al., 2006). In soil their presence is acknowledged as an indication 2000). To minimize this problem, the current study included that oxalic acid is an important agent in biological weather- fluorapatite dissolution experiments under both biotic conditions ing, resulting in increased phosphorus availability to plants with fungal mycelia present and under abiotic conditions using (Graustein & Cromack, 1977). It has also been suggested that liquid media conditioned by fungal exudates. The use of calcium oxalate crystals may protect hyphae from dehydration, media conditioned by fungal growth under P and/or Ca provide a physical barrier against grazing microfauna (Arocena availability for 12 days enabled us to separate fluorapatite et al., 2001, and references therein) and serve as a means to dissolution by fungal exudates from the total fluorapatite regulate external calcium concentrations (Connolly et al., 1999). dissolution by exudates, physical contact, and mycelial uptake. In the present study, calcium oxalate crystals were observed Without destructive sampling, we were unable to measure on apatite particles reacted for 36 days by both Mucorales incremental dissolution rates in the biotic weathering isolates. While ANG3 resulted in calcium oxalate coating of all experiment directly, and therefore to correlate the dissolu- apatite particles (Fig. 8A,B) after 12 days, ANG55 produced tion mechanism with specific growth stages. Nevertheless, scattered oxalate crystals on highly weathered apatite particles comparison of the dissolution activity of exudates (i.e. high (Fig. 8D,E). Over 36 days, the total weathering rate of dissolution rate in reactor A_3 and A_74; Fig. 7) to the total ANG55 was significantly higher compared to that of ANG3, rates for ANG3 and ANG74 over 36 days suggests that the suggesting that calcium oxalate precipitation may prevent dissolution mechanism that is effective initially may be less long-term dissolution of apatite. ANG74 grew well on apatite efficient in the long run. ANG55 had the highest overall rate as the only source of P and Ca. However, decreasing pH at 10 × –3.3 µmol P m–2 s–1 but its exudates in conditioned resulted in increasing Ca concentrations. The observed loss of media (+A) had the lowest incremental rates detected (Fig. 7). mycelial vigor in the second half of experiment 1 may have resulted from stress induced by increasing external Ca Oxalic acid mediated fluorapatite dissolution H1X concentration. Free Ca in the solution had a negative effect on Fungal exudates (P_3 and A_3) with oxalic acid concentrations ANG74, resulting in lower biomass in treatment +CaP above 1 mM and at an initial pH of 2.8 and 3.1, respectively, compared to +P (Fig. 2). No oxalic acid was produced induce rapid release of phosphorus from fluorapatite (Fig. 5). by ANG74 and the isolate seemed to lack a mechanism At pH 3, Welch et al. (2002) found that 1 mM oxalate to control external Ca concentrations. Moderate oxalate significantly increased P release compared to inorganic production appears as the most favorable strategy for fungal conditions. At the end of that experiment, P concentrations growth on apatite as a source of P. Although these closed reached 880 µM P while pH increased almost one unit during experimental systems differ from the open soil system, these the reaction (Welch et al., 2002). Reactors P_3 and A_3 long-term effects may still be relevant in the field due to reached a maximum P concentration at 250 µM (Fig. 5) and microspatial heterogeneity and limited mobility of pH of the solution increased approximately 0.1 unit during compounds. the reaction. Similar to observations by Welch et al. (2002), The current study is the first report on hydroxyapatite the higher oxalate concentrations and smaller pH increase that precipitated in association with fungal mycelia. Biologically we observed suggest that consumption of protons by the induced precipitation of hydroxyapatite has been demonstrated dissolution reaction did not limit fluorapatite dissolution for the large sulfur bacteria Thiomargarita namibiensis, which towards the end of the experiment. Ca released from induce spontaneous precipitation by exudation of P to local fluorapatite is removed from the solution through the micro-environments in sediments (Schulz & Schulz, 2005). formation of calcium oxalate crystals in reactors with media We suggest that hydroxyapatite formation within mycelial conditioned by ANG3 or ANG55. Consequently cell-free spheres is similarly induced by high local P concentrations. media remained undersaturated with respect to fluorapatite Weather this phenomenon is relevant under natural conditions throughout the dissolution experiment, which will promote remains to be examined. It could potentially be a mechanism phosphorus release from apatite (Guidry & Mackenzie, of extracellular P immobilization in the direct vicinity of 2003). On the other hand, our results suggest that oxalic acid fungal hyphae. If sequestered within extracellular mucilage concentrations above 1 mM induce high initial dissolution associated with the hyphae, hydroxyapatite could provide a rates at pH values around 3 but that the total amount of future source of P controlled by the fungi. The precipitation dissolved P is limited by the formation of calcium oxalate of HAp during TCP dissolution in liquid media may crystals that eventually coat the fluorapatite particles however serve as a diagnostic trait suitable to detect bioge- (Fig. 8A,B). No P was released from fluorapatite in reactor ochemically active isolates with potential for active weathering Ca_3 even though the oxalate concentration was five times pathways. higher in this reactor than in P_55 (Table 3), probably due to

© 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd 278 A. ROSLING et al. the higher pH. The rate of phosphorus release from apatite is of different fungal groups changes with altered rainfall known to increase with decreasing pH, in the pH range 5–2 patterns and only a small subset of the biogeochemically active (Welch et al., 2002; Guidry & Mackenzie, 2003). We suggest isolates have been characterized for their ability to dissolve that fluorapatite dissolution in reactors P_3 and A_3 was fluorapatite. Given the clear differences in ability to stimulate dominated by the effect of oxalate concentrations and phosphate release that were demonstrated in this study, the conclude that oxalate-controlled dissolution rates are pH effect of changes in fungal populations on primary productivity independent for pH values below 4.5. Phosphorus solubil- might be difficult to predict at the ecosystem level. ization by fungi has been suggested to depend on the amount The results of this work have broader implications. The and composition of organic acids produced (Whitelaw, productivity of many soils and ecosystems worldwide is 2000). The comparably high dissolution rates in reactor P_55 phosphate limited, and phosphorus availability can influence could result from the combined activity of oxalate, malonate species composition and organism growth rates in a wide and malate (Table 3). However, these concentrations are fairly variety of ecological contexts (Elser et al., 2000, 2003). Changes low and other unknown compounds may be important to in climate that result in altered vegetation patterns and soil the high dissolution rate of reactor P_55 at pH 4. The presence carbon concentrations will also modify fungal community of other compounds, i.e. siderophores, that co-act to increase structure. Depending on which fungal types dominate under weathering rates (Cheah, 2003) may be significant in these the new conditions, the supply of phosphate to the ecosystem reactors. However, hydroxamate siderophores appear to be from mineral sources may increase or decrease, with potential insignificant for the documented weathering in the current significant consequences for soil and ecosystem productivity. study. ACKNOWLEDGEMENTS Fluorapatite dissolution mediated by other organic compounds This research was made possible through funding from the When grown under P-limiting conditions, ANG74 acidifies it Swedish Research Council (VR) and Lennander’s Foundation media (Fig. 4C) without producing significant concentrations at Uppsala University, Sweden. Some research expenses were of any measured LMWOAs (Table 3). Abiotic dissolution by covered by the National Center for Earth Surface Dynamics Ca_74 is not primarily proton driven since P is released from (NCED). We thank Peter Steel and the University of California fluorapatite (Fig. 5) without lowering pH (data not shown). Natural Reserve System for protection and stewardship of the The calculated saturation state of the solution (Fig. 6) points field site. Special thanks to Dr Paul D. Brooks for assistance toward oversaturated conditions that would contradict the with ICP analyses, Tim Teague for assistance with XRD observed release of P. We suggest then, that a fungally exuded experiments, Karlyn Cruz for field assistance and helpful compound forms complexes with either Ca or P in the solution. discussion, and Jonathan Giska, Lis Green and Dr Linda The ability of such fungi to induce fluorapatite dissolution in Kalnejais for their contributions. We thank Dr Sara Holmström response to P deficiency may be an important mechanism by for stimulating discussion on fungal mineral dissolution and which fungi can respond to local soil nutrient availability, siderophores, Dr. Christine Hawkes for valuable discussion of enabling them to overcome P-limiting conditions. Weathering field-sampling strategies and Dr Petra Fransson for statistical agents other than LMWOAs need to be considered when support. evaluating the potential importance of biotic weathering in the field (Sverdrup et al., 2002; van Hees et al., 2005). REFERENCES

Abdullahi AS, Underwood GJC, Gretz MR (2006) Extracellular Ecological significance matrix assembly in diatoms (Bacillariophyceae). V. Environmental effects on polysaccharide synthesis in the model diatom, Soil fungi were isolated from an experimental grassland in Phaeodactylum tricornutum. Journal of Phycology 42, 363–378. northern California where precipitation has been amended to Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, simulate future climate change since 2001 (Suttle et al., 2007). Lipman DF (1997) Gapped BLAST and PSI-BLAST: a new Although the grassland ecosystem is currently N limited, generation of protein database search programs. Nucleic Acids one of the strongest responses through 6 years of rainfall Research 25, 3389–3402. Arocena JM, Glowa KR, Massicotte HB (2001) Calcium-rich hypha manipulation has been the proliferation of plants associated encrustations on Piloderma. Mycorrhiza 10, 209–215. with N-fixing bacteria. As increasing soil nitrogen availability Arvieu J-C, Laprince F, Plassard C (2003) Release of oxalate and mitigates local N limitation, P availability could become more protons by ectomycorrhizal fungi response to P-deficiency and limiting to plant production. In this case, the subsequent calcium carbonate in nutrient solution. Annals of Forest Science 60, response of the ecosystem may depend on the ability of plants 815–821. Banfield JF, Barker WW, Welch SA, Taunton A (1999) Biological and microbes to obtain P from available sources, and fungal impact on mineral dissolution: application of the lichen model to weathering of apatite could play an important role in mediating understanding mineral weathering in the rhizosphere. PNAS 96, this. At this time, we do not know how the relative abundance 3404–3411.

Barker WW, Banfield JF (1996) Biological versus inorganically microorganisms. Journal of General Applied Microbiology 40, mediated weathering reactions: relationship between minerals and 255–260. extracellular microbial polymers in lithobiontic communities. Hirsch P, Eckhardt FEW, Palmer JRJ (1995) Fungi active in Chemical Geology 132, 55–69. weathering of rock and stone monuments. Canadian Journal of Burford EP, Fomina M, Gadd GM (2003) Fungal involvement in Botany 73, S1384–S1390. bioweathering and biotransformation of rocks and minerals. Kalinowski BE, Liermann LJ, Givens S, Brantley SL (2000) Rates of Mineralogical Magazine 67, 1127–1155. bacteria-promoted solubilization of Fe from minerals: a review of Callot G, Maurette M, Pottier L, Dubois A (1987) Biogenic etching problems and approaches. Chemical Geology 169, 357–370. of microfractures in amorphous and crystalline silicates. Nature Magnusson JK, Lasure LL (2004) Organic acid production by 328, 147–149. filamentous fungi. In: Advances in Fungal Biotechnology for Casarin V, Plassard C, Souche G, Arvieu JC (2003) Quantification of Industry, Agriculture, and Medicine (eds Lange J, Lange L). oxalate ions and protons released by ectomycorrhizal fungi in Kluwer Academic/Plenum Publishers, New York, pp. 307–340. rhizosphere soil. Agronomie 23, 461–469. Mahmood S, Finlay RD, Erland S, Wallander H (2001) Solubilisation Cheah S-F, Stephan M, Kraemer SM, Cervini-Silva J, Sposito G and colonisation of wood ash by ectomycorrhizal fungi isolated (2003) Steady-state dissolution kinetics of goethite in the from wood ash fertilised spruce forest. FEMS Microbiology Ecology presence of desferrioxamine B and oxalate ligands: implications 35, 151–161. for the microbial acquisition of iron. Chemical Geology 198, Marschner P, Jentschke G, Godbold DL (1998) Cation exchange 63–75. capacity and lead sorption in ectomycorrhizal fungi. Plant and Soil Connolly JH, Shortle WC, Jellison J (1999) Translocation and 205, 93–98. incorporation of strontium carbonate derived strontium into Marx DH (1969) The influence of ectotrophic fungi on the resistance calcium oxalate crystals by the wood decay fungus Resinicium of pine roots to pathogenic infection. Phytopathology 59, 153–163. bicolor. Canadian Journal of Botany 77, 179–197. Oshima Y (1997) The phosphatase system in Saccharomyces cervisiae. Dahlén J, Hagberg J, Karlsson S (2000) Analysis of low molecular Genes and Genetic Systems 72, 323–334. weight organic acids in water with capillary zone electrophoresis Parkinsson (1994) Filamentous fungi. In: Methods of Soil Analysis. employing indirect photometric detection. Fresenius Journal of Part 2. Microbiological and Biochemical Properties (eds Weaver RW, Analytical Chemistry 366, 488–493. Angle S, Bottomley P). Soil Science Society of America, Madison, Ekblad A, Wallander H, Carlsson R, Huss-Danell K (1995) Fungal Wisconsin, pp. 329–350. biomass in roots and extrametrical mycelium in relation to Raaska L, Viikari L, Mattila-Sandholm T (1993) Detection of macronutrients and plant biomass of ectomycorrhizal Pinus siderophores in growing cultures of Pseudomonas spp. Journal sylvestris and Alnus incana. New Phytologist 131, 443–451. of Industrial Microbiology and Biotechnology 11, 181–186. Elser JJ, Sterner RW, Gorokhova E, Fagan WF, Markow TA, Cotner Raghothama KG (1999) Phosphate acquisition. Annual Review JB, Harrison JF, Hobbie SE, Odell GM, Weider LJ (2000) Plant Physiology and Plant Molecular Biology 50, 665–693. Biological stoichiometry from genes to ecosystems. Ecology Letters Robson G (1999) Hyphal cell biology. In: Molecular Fungal Biology 3, 540–550. (ed. Schweizer M). Cambridge University Press, Cambridge. Elser JJ, Acharya K, Kyle M, Cotner J, Makino W, Markow T, Rosling A, Lindahl BD, Taylor AFS, Finlay R (2004) Mycelial growth Watts T, Hobbie S, Fagan W, Schade J, Hood J, Sterner RW and substrate acidification of ectomycorrhizal fungi in response to (2003) Growth rate-stoichiometry couplings in diverse biota. different minerals. FEMS Microbiology Ecology 47, 31–37. Ecology Letters 6, 936–943. Ryan PR, Delhaize E, Jones DL (2001) Function and mechanism of Finlay RD, Rosling A (2006) Integrated nutrient cycles in forest organic anion exudation from plant roots. Annual Review Plant ecosystems-the role of ectomycorrhizal fungi. In: Fungi in Physiology and Plant Molecular Biology 52, 527–560. Biogeochemical Cycles (ed. Gadd GM). Cambridge University Press, Schulz HN, Schulz HD (2005) Large sulfur bacteria and the Cambridge, UK, pp. 28–50. formation of phosphorite. Science 307, 416–418. Gadd GM (1999) Fungal production of citric and oxalic acid: Siegel SM, Galun M, Siegel BZ (1990) Filamentous fungi as metal Importance in metal speciation, physiology and biogeochemical biosorbents: a review. Water, Air and Soil Pollution 53, 335–344. processes. In: Advances in Microbial Physiology, Vol 41 (ed. Poole Stumm W, Morgan JJ (1996) Aquatic Chemistry: Chemical RK). Academic Press, Sheffield, UK, pp. 48–92. Equilibria and Rates in Natural Waters. Wiley-Interscience, New Gallmetzer M, Burgstaller W (2002) Efflux of organic acids in York. Penicillium simplicissimum is an energy-spilling process, adjusting Suttle KB, Thomsen MA, Power ME (2007) Species interactions the catabolic carbon flow to the nutrient supply and the activity of reverse grassland responses to changing climate. Science 315, catabolic pathways. Microbiology 148, 1143–1149. 640–642. Gharieb MM (2000) Nutritional effects on oxalic acid production and Sverdrup H, Hagen-Thorn A, Holmquist J, Wallman P, Warfvinge P, solubilization of gypsum by Aspergillus niger. Mycological Research Walse C, Alveteg M (2002) Biogeochemical processes and 104, 550–556. mechanisms. In: Developing Principles and Models for Gibson BR, Mitchell DT (2004) Nutritional influences on the Sustainable Forestry in Sweden (eds Sverdrup H, Stjernquist I). solubilization of metal phosphate by ericoid mycorrhizal fungi. Kluwer Academic Publishers, Dordrecht, the Netherlands, Mycological Research 108, 947–954. pp. 91–196. Graustein WC, Cromack K (1977) Calcium oxalate: occurrence in Tasaki Y, Ohata K, Hara T, Joh T (2004a) Three genes specifically soils and effects on nutrient and geochemical cycles. Science 198, expressed during phosphate deficiency in Pholiota nameko strain N2 1252–1254. encode hydrophobins. Current Genetics 45, 19–27. Guidry MW, Mackenzie FT (2003) Experimental study of igneous Tasaki Y, Azwan A, Hara T, Joh T (2004b) Structure and expression and sedimentary apatite dissolution: control of pH, distance from of a phosphate deficiency-inducible ribonuclease gene in Pholiota equilibrium, and temperature on dissolution rates. Geochimica et nameko. Current Genetics 45, 28–36. Cosmochimica Acta 67, 2949–2963. Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W: Gupta R, Singal R, Shankar A, Kuhad RC, Saxena RK (1994) improving the sensitivity of progressive multiple sequence A modified plate assay for screening phosphate solubilizing alignment through sequence weighting, position-specific gap

penalties and weight matrix choice. Nucleic Acids Research 22, Welch SA, Taunton AE, Banﬁeld JF (2002) Effects of microorganisms 4673–4680. and microbial metabolites on apatite dissolution. Geomicrobiology van Hees PAW, Rosling A, Essén S, Godbold DL, Jones DL, Finlay Journal 19, 343–367. RD (2005) Oxalate and ferricosin exudation by the extrametrical White TJ, Bruns T, Lee S, Taylor J (1990) Ampliﬁcation and direct mycelium of an ectomycorrhizal fungus in symbiosis with Pinus sequencing of fungal ribosomal RNA genes for phylogenetics. In: sylvestris. New Phytologist 169, 367–378. PCR Protocols: A Guide to Methods and Applications (eds Innis MA, Verrecchia EP, Braissant O, Cailleau G (2006) The oxalate- Gelfrand DH, Snisky JJ, White TJ). Academic Press, San Diego, carbonate pathway in soil carbon storage: the role of fungi and California, pp. 315–322. oxalotrophic bacteria. In: Fungi in Biogeochemical Cycles (ed. Whitelaw MA (2000) Growth promotion of plant inoculation with Gadd GM). Cambridge University Press, Cambridge, UK, phosphate-solubilizing fungi. Advances in Agronomy 69, 99–151. pp. 28–50. Zarcinas BA, Cartwright B, Spouncer LR (1987) Nitric acid digestion Wallander H, Wickman T, Jacks G (1997) Apatite as a P source in and multi-element analysis of plant material by inductively coupled mycorrhizal and non-mycorrhizal Pinus sylvestris seedlings. Plant plasma spectrometry. Communication in Soil Science and Plant and Soil 196, 123–131. Analysis 18, 131–146.