Phytase activity in lichens

Summary Phytase activity was investigated in 13 lichen species using a novel assay method. The work tested the hypothesis that phytase is a component of the suite of surface‐bound lichen enzymes that hydrolyse simple organic forms of phosphorus (P) and nitrogen (N) deposited onto the thallus surface. Hydrolysis of inositol hexaphosphate (InsP6, the substrate for phytase) and appearance of lower‐order inositol phosphates (InsP5–InsP1), the hydrolysis products, were measured by ion chromatography. Phytase activity in Evernia prunastri was compared among locations with contrasting rates of N deposition. Phytase activity was readily measurable in epiphytic lichens (e.g. 11.3 μmol InsP6 hydrolysed g−1 h−1 in Bryoria fuscescens) but low in two terricolous species tested (Cladonia portentosa and Peltigera membranacea). Phytase and phosphomonoesterase activities were positively correlated amongst species. In E. prunastri both enzyme activities were promoted by N enrichment and phytase activity was readily released into thallus washings. InsP6 was not detected in tree canopy throughfall but was present in pollen leachate. Capacity to hydrolyse InsP6 appears widespread amongst lichens potentially promoting P capture from atmospheric deposits and plant leachates, and P cycling in forest canopies. The enzyme assay used here might find wider application in studies on plant root–fungal–soil systems.


Introduction
Lichens typically grow in habitats with low availabilities of nitrogen (N) and phosphorus (P) (Crittenden, 1989;Crittenden et al., 1994). There is evidence to suggest that supply of these key nutrients is growth limiting for the more productive lichens such as mat-forming species of Cladonia and large foliose cyanolichens (Crittenden et al., 1994;Kyt€ oviita & Crittenden, 2007;McCune & Caldwell, 2009). The majority of lichens acquire these nutrients from atmospheric deposits by efficiently scavenging inorganic N and P (Pi) from solutes in precipitation intercepted either directly or after being modified by plant canopies and then delivered to thalli as canopy throughfall and stem flow (Crittenden, 1983(Crittenden, , 1989(Crittenden, , 1998. Nitrogen compounds are routinely monitored in both precipitation and air, and there is an abundance of data on rates of N deposition with which lichen N concentration ([N]) is frequently positively correlated (Bruteig, 1993;Hyv€ arinen & Crittenden, 1998b;Remke et al., 2009;Hogan et al., 2010a;Boltersdorf et al., 2014). By contrast, P is not routinely measured in atmospheric deposits, there is a dearth of information on P deposition rates and how lichens respond to variation in P income is largely unknown. Detailed P-focused investigations suggest that Pi concentration in rainfall is in the range 0.03-0.33 lmol l À1 (Grimshaw & Dolske, 2002;Neal et al., 2003;Yoshioka et al., 2009;He et al., 2011); Pi at these low concentrations is taken up rapidly by lichens via high affinity phosphate transport systems (Farrar, 1976;Hogan, 2009).
In addition to a capacity for rapid uptake of inorganic ions and small nitrogenous organic molecules such as amino acids (Dahlman et al., 2004), lichens also have surface phosphatase activities that promote the release of covalently bonded P in organophosphates (Po) (Lane & Puckett, 1979;Hogan et al., 2010a;Crittenden et al., 2015). Artificial chromogenic Po substrates hydrolysed by lichens include the monoester para-nitrophenyl phosphate (pNPP), and the diesters bis-(para) nitrophenol phosphate (bis-pNPP) and para-nitrophenyl phenylphosphonate (pNPPP, a DNA analogue) (Hogan, 2009); the hydrolysis of these compounds can be measured colorimetrically because it results in the coloured product, para-nitrophenol (pNP). Capacity for phosphatase activity in lichens varies among both species and habitats probably reflecting differences in the availability of, and demand for, P (Lane & Puckett, 1979;Hogan, 2009;Lewis, 2012). N enrichment, due to anthropogenic pollution, alters thallus N : P ratio driving an upregulation of phosphomonoesterase (PME) activity (Hogan, 2009;Hogan et al., 2010a,b). However, despite the apparent diversity of phosphatase activities as indicated by analogous substrates, little is known about the forms and quantities of Po available to lichens in their natural habitats (cf. Tibbett, 2002;Turner et al., 2002).
Most natural waters to which lichens are exposed contain traces of organic P. Yoshioka et al. (2009) and He et al. (2011) found the ratio of total P : soluble reactive phosphate-P was between 3 and 5 in both rainfall and atmospheric aerosol, raising the possibility of a significant organic P component in atmospheric deposits. Annual deposition of this putative Po fraction was in the range 11-50 and 38-160 mg m À2 yr À1 for wet and dry inputs, respectively (He et al., 2011). However, specific compounds were not identified. Rainfall that passes through plant canopies as throughfall and stemflow can be enriched in organic carbon including organophosphates (Zimmermann et al., 2007). Again specific compounds have not been identified but these probably include a range of organophosphate metabolites from plant leaves, epiphytes and phylloplane microorganisms. Also, lichens that grow on soil, or on surfaces exposed to seepage water draining from soil, will be exposed to traces of soil organic P (Martinsen et al., 2013); most lichens can potentially experience deposition of animal faeces or solutes derived from them (Tomassen et al., 2005).
Phytic acid (inositol hexakisphosphate; InsP 6 ) is a major source of Po in the environment (Turner et al., 2002;Turner, 2006). InsP 6 is a myo-inositol ring esterified with six symmetrically distributed phosphate groups (Shears & Turner, 2007). Lower-order inositol phosphates (InsP 5 -InsP 1 ), distinguished by the number of substituted phosphate groups on the inositol ring, also occur widely in nature. InsP 6 is a common constituent of eukaryotic cells where it is involved in signal transduction and ATP regeneration, amongst other processes (Raboy, 2003). It is also accumulated by plants (Frank, 2013), and to a lesser extent by microorganisms (Cosgrove & Irving, 1980;Turner et al., 2002), as a P storage compound. InsP 6 accounts for up to 90% of total organic P in seeds (Frank, 2013) but also occurs in leaves, roots, flowers and fruits, constituting up to 6% of organic P in leaves of herbaceous plants (Alkarawi & Zotz, 2014a,b), up to 0.22% of total P in coniferous leaves (Frank, 2013) and up to 70% of organic P in pollen . Data on InsP 6 concentrations in microorganisms are extremely scant, reported values include 76 nmol g À1 in Neurospora crassa (Lakin-Thomas, 1993) and 700 lM in the cytoplasm of Dictyostelium discoideum (Martin et al., 1987). Concentrations in lichens and bryophytes have not been documented, but Winkler & Zotz (2009) present evidence of InsP 6 accumulation in the epiphytic bromeliad Aechmea fasciata. InsP 6 becomes strongly complexed in soils during decomposition forming up to 60% of total soil Po reserves (Cosgrove & Irving, 1980;Turner et al., 2002Turner et al., , 2007Turner, 2006).
Although InsP 6 is likely to prove an important component of the terrestrial P cycle, there remains a dearth of information regarding the rates of release of Pi during its degradation (Turner et al., 2002). Phytases (myo-inositol hexakisphosphate phosphohydrolases) are probably the principal agents involved in hydrolysis of the ester linkages. This class of enzyme initiates a stepwise dephosphorylation of InsP 6 , releasing Pi and a series of partially dephosphorylated myo-inositols. Phytases have been most frequently reported in fungi and bacteria (Mukhametzyanova et al., 2012) but are also produced by plants (Hayes et al., 1999), most notably in germinating seeds (Gibson & Ullah, 1988;Hegeman & Grabau, 2001). Studies of phytase activity have been impeded by the lack of a sensitive assay. Inositol phosphates do not absorb visible or ultraviolet light and therefore cannot be detected spectrophotometrically. A new chromophoric substrate analogue of InsP 6 , 5-O-[6-(benzoylamino)hexyl]-D-myo-inositol-1,2,3,4,6pentakisphosphate (T-IP5), is under development; this permits the measurement of phosphate ester bond cleavage by phytase (Berry & Berry, 2005;Berry et al., 2007) but the authors consider that further refinement of the method is required before it can be used to assay biologically active samples. To date, measurement of phytase activity has largely relied on determining the rate of Pi release during InsP 6 hydrolysis (Antibus et al., 1992;Yadav & Tarafdar, 2003). However, this method suffers from the drawback that living material, such as lichens or mycorrhizal roots, frequently has a high Pi uptake rate (Farrar, 1976;Hyv€ arinen & Crittenden, 1998a;Smith & Read, 2008;Hogan, 2009) which potentially will lead to underestimation of Pi release promoted by phytase.
The aim of the present study was to test for the presence of phytase activity in lichens employing a method that did not rely on the measurement of Pi release from samples. Accordingly, we use a high performance ion chromatographic method to measure the rates of consumption of InsP 6 and the production of lowerorder inositol phosphates. We develop a phytase assay procedure suitable for lichens using the common epiphyte Evernia prunastri and then compare rates of activity among 13 lichens including 11 epiphytic and two terricolous species. We also test the hypothesis that phytase activity responds positively to N enrichment as has been demonstrated for PME activity in lichens (Hogan et al., 2010a;Crittenden et al., 2015). The results are discussed in relation to possible sources of InsP 6 for lichens in their natural habitats.

Lichen collection and pretreatment
Evernia prunastri was collected from the Derwent Valley, Derbyshire, UK, and from other sites subject to different rates of wet inorganic N (NO 3 À + NH 4 + ) deposition (Table 1). Modelled mean N deposition values pertaining to the collection sites for the period 2009-2011 were provided by R.I. Smith (Centre for Ecology and Hydrology, Edinburgh, UK) and were abstracted from 5 9 5 km gridded datasets. A further 12 species were collected for an interspecies comparative study (Table 1).
At each collection site, 10 replicate lichen samples were collected from locations > 10 m apart. Samples were returned to the laboratory in polythene bags, air dried on a laboratory bench for 24 h at 18°C, re-sealed in polythene bags and stored at À20°C until required. Before assays, lichens were gradually rehydrated overnight in water-saturated air (over water in a desiccator) at 10°C, then fully re-saturated by spraying with deionized water, cleaned of extraneous debris and then blotted to remove surface moisture. In order to reduce variability among replicates, assays were confined to either the terminal 10-mm branch tips of fruticose lichens or the marginal 10 mm of foliose thalli; these were cut off using a scalpel and used immediately in assays (see enzyme assays below). Powder-free latex gloves were used at all times when handling lichens in the field and laboratory.

Thallus surface pH
Thallus surface pH was measured following Hogan et al. (2010a). In brief, a flat tip pH electrode (Gelplas double junction flat tip electrode; VWR International Ltd, Lutterworth, UK) was held in contact with the apical or marginal 15-mm region of the lichen thallus, saturated with excess 0.025 M KCl, noting the pH value after 1 min.

Determination of phytase activity
Assays were initiated by adding c. 40 mg (unless otherwise stated) wet mass of lichen thallus to 3 ml of InsP 6 (dodecasodium salt hydrate; Sigma-Aldrich) solution made up in citric acidtrisodium citrate buffer containing major ions representative of UK precipitation (20 lmol l À1 MgSO 4 .7H 2 O, 8 lmol l À1 CaCl 2. 2H 2 O, 150 lmol l À1 NaCl, 15 lmol l À1 NH 4 NO 3 , 5 lmol l À1 KNO 3 ; Hayman et al., 2007). The citric acid-trisodium citrate buffer was prepared following Dawson et al. (1986) and then diluted 10-fold. Samples were incubated in a shaking water bath in the dark at temperatures and for time periods ranging from 5 to 30°C and 0.25 to 24 h, respectively. Reactions were terminated by transferring 2.5 ml of the assay medium to 250 ll of 0.5 M HCl in an autosampler tube. Thalli were then blotted to remove surface solution and dried at 80°C for 24 h before weighing.
Quantities of InsP 6 and its hydrolysis products in the acidified subsamples were determined by high performance ion chromatography (HPIC) following Chen & Li (2003) using a Dionex ICS5000 ion chromatograph (Dionex (UK) Ltd, Camberley, Surrey, UK). In brief, inositol phosphates were separated on a CarboPac TM PA-100 analytical column (250 9 4 mm) at 30°C using 0.5 M HCl as mobile phase in a linear gradient elution program followed by post-column reaction with a solution of 1 g l À1 Fe(NO 3 ) 3 in 0.33 M HClO 4 using a Dionex knitted coil (750 ll). InsP 6, Ins(1,2,4,5,6)P 5 and Ins(1,5,6)P 3 peaks were identified from retention times of standards and InsP 4, InsP 2 and InsP 1 peaks were identified from retention times given by Chen & Li (2003), although specific isomers of InsP 2 and InsP 1 compounds were difficult to resolve. The system was calibrated for InsP 6 , the response being linear up to 0.1 lmol on column. CHROMELEON 7.2 software (Dionex) was used for instrument control and data acquisition and processing. Run time required for InsP 6 , the slowest eluting analyte, was 50.1 min with an equilibration time of 10 min needed between each run. Rates of activity were expressed as lmol InsP 6 hydrolysed g À1 dry mass h À1 . 'No-lichen' and 'no-substrate' control samples were included in each batch of analyses.
Preliminary experiments examined phytase response to pH (2.5-6.7), and to substrate concentration (1-10 mM); rates from the latter experiment were fitted to the Michaelis-Menten equation using ORIGINPRO 8. This range of substrate concentrations was selected to help characterize the enzyme system and has no ecological relevance (cf. Whitton et al., 2005) because in nature InsP 6 is likely to be available to lichens in trace quantities only.

New Phytologist
The assay medium pH value selected for each species was either the putative optimum value determined in this study, the optimum value for PME activity determined by Lewis (2012) or the surface pH value (Table 1). The potential inhibitory effect on phytase activity of pNPP, the substrate for PME assays (see Determination of phosphomonoesterase activity below), was investigated in E. prunastri by including both pNPP and InsP 6 in phytase assays in equimolar concentrations.
In order to determine whether enzyme activity could be removed from the thallus by washing, c. 150 mg dry mass of E. prunastri was shaken in 10 ml of assay medium at the optimum pH, and 15°C for 5 h. Phytase assays were then run on the following samples: (1) thalli removed from the wash and blotted dry with tissue to remove old assay medium (= washed thalli); (2) 2.9 ml of assay medium without thalli (= washings); (3) 2.9 ml of filtered (to pass 0.22 lm) assay medium without thalli (= filtered washings); and (4) fresh thalli not pre-treated by washing (= unwashed thalli). Washings were filtered to exclude lichen-associated bacteria and yeasts that might be a source of phytase (Mukhametzyanova et al., 2012). Assays on washings were initiated by adding 0.1 ml InsP 6 solution to 2.9 ml to yield a final substrate concentration of 1 mM and incubated for 1 h. Phytase activity in washings was expressed per unit mass of thallus from which the washings were derived.
Determination of phosphomonoesterase activity PME activity was determined using Bessey et al.'s (1946) pNPP colorimetric assay as described by Hogan et al. (2010a). Lichen samples were added to 2.9 ml assay medium of similar composition to that used above for phytase measurements. Assays were initiated by adding 0.1 ml pNPP, to yield a final concentration of either 0.5 or 10 mM and samples were incubated for 20 min as for phytase assays, after which the reaction was terminated by transferring 2.5 ml assay medium into 0.25 ml terminator solution (1.1 M NaOH, 27.5 mM EDTA, 0.55 M K 2 HPO 4 ) and the absorbance measured at 405 nm using a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Thalli were then dried and weighed as above. Effective enzyme activity was expressed as lmol pNPP hydrolysed g À1 dry mass h À1 using pNP to calibrate the assay. 'No-lichen' and 'no-substrate' control runs were included in each batch of analyses. Assay medium pH was adjusted within the range 2.5-6.7 depending on species being tested; in each case the value selected was either the optimum value for the species determined by Lewis (2012), the optimum value for phytase activity or, if unknown, the thallus surface pH value ( Table 1). Note that a 10 mM substrate concentration is considered saturating for PME activity in lichens (Hogan, 2009;Crittenden et al., 2015).

Assessment of canopy throughfall and pollen as sources on InsP 6
Canopy throughfall was sampled in triplicate in two stands of sycamore (Acer pseudoplatanus): during September in Nottingham City (Grid Reference SK 518375, 52°55 0 58 00 N, 1°13 0 48 00 W); and in October in the Derwent Valley, Derbyshire (Table 1) where trees support large populations of E. prunastri. Polyethylene funnels (diameter 22 cm) connected to 1-l bottles were positioned beneath trees with the funnel tops at 1 m above the ground. Funnels were fitted with stainless steel spikes on their circumferences to prevent bird perching and resultant contamination (Asman et al., 1982); they were deployed shortly before the onset of forecast rainfall events and left in the field for up to 24 h. Throughfall samples were passed in sequence through a cellulose filter paper and a nitrocellulose membrane filter (0.22 lm, Millipore), freeze-dried, and the residue stored at À20°C until required. Samples were then re-solubilized in 1 ml water and analysed by HPIC for the presence of InsP 6 .
Pollen (25 mg) was collected from oriental pink lilies (Lilium candidum), and shaken in 7 ml of assay medium at the optimum pH for 1 h. Filtrate (0.22 lm, Millipore) was analysed as above.
To determine whether enzyme activity in E. prunastri could hydrolyse InsP 6 present in lily pollen, lichen thalli were incubated in assay medium containing pollen for 5 h and then analysed as described above.

Data analysis
SPSS (SPSS (UK) Ltd, Woking, UK) was used to perform standard statistical analyses. All data were checked for normality of distribution (Kolmogorov-Smirnov test) and homogeneity of variances and where test assumptions were not met, either log-or square root-transformed. Relationships between rates of enzyme activity and environmental variables were subjected to either one-way ANOVA, ANCOVA, correlation or regression analysis.

Results
All raw data are available in Supporting Information Table S1.

Studies on Evernia prunastri
Hydrolysis of InsP 6 was readily detectable in Evernia prunastri. As incubation time increased from 0.5 to 24 h, InsP 6 decreased in abundance while a peak consistent with InsP 5 , the first hydrolysis product, appeared and rapidly increased. Peaks consistent with InsP 4 -InsP 1 were also produced and similarly increased over a 24 h incubation period. Hence, a reduction in InsP 6 abundance could be used as a measure of phytase activity (Fig. 1). The detection limit (= 59 standard deviation of repeat 'blank' samples; Wilson, 1974) for depletion of InsP 6 in samples containing 1 lmol ml À1 InsP 6 was 0.0018 lmol ml À1 , roughly equating to a rate of activity in E. prunastri of 0.86 lmol InsP 6 hydrolysed g À1 h À1 . 'No-substrate' controls did not produce substances that could be detected by HPIC.
Although InsP 6 was progressively hydrolysed with time, rates during the first 0.25 h were often atypically high. This was clearly evident when the progress of InsP 6 hydrolysis in incubations of increasing duration was compared at different temperatures ( Fig. 2a-f). However, the rates of hydrolysis determined from the gradients of the regression lines for each temperature were highly www.newphytologist.com correlated with temperature ( Fig. 2g) suggesting that initial high rates often observed in short assays might result from nonenzymic loss of substrate (e.g. Fig. 2f). Therefore, to avoid overestimation of enzyme activity, rate of phytase activity in all subsequent assays was determined from the difference in the quantity of InsP 6 hydrolysed between 0.5 h and 5 h incubations. Phytase activity was strongly dependent on both pH (Fig. 2h) and substrate concentration (Fig. S1). Maximum rates were recorded at pH 2.5 and were 18 times higher than those at pH 6.7; this putative optimum is broadly consistent with the measured thallus surface pH of 3.4 AE 0.1 (n = 6). Phytase response to substrate concentration yielded apparent K m and V max values of 3.01 AE 0.5 mM and 13.08 AE 2.2 lmol InsP 6 hydrolysed g À1 dry mass h À1 , respectively. A non-enzyme-saturating substrate concentration of 1 mM was used in all subsequent routine assays to reduce costs and to obviate the need to routinely dilute samples before HPIC analysis.
Lichen washings contained appreciable phytase activity which was reduced little by filtering (Fig. 3a). Surprisingly, there were no significant differences between phytase activities in the filtered washings alone and in either fresh (i.e. not pre-washed) lichen or lichen prewashed for 5 h. Lichen washings also contained PME activity, but rates of pNPP hydrolysis recorded in filtered and unfiltered washings were only 17-23% of those in assays containing thalli.
Phytase and PME activities in E. prunastri strongly co-varied among sites with contrasting N deposition (Fig. 3b). However, neither enzyme activity was significantly related to wet inorganic N deposition (phytase activity, Fig. 3c; PME activity, data not shown, r 2 = 0.22, P = 0.19) or to its components (deposition of NO 3 À and NH 4 + singly, and their concentrations in rainfall, data not shown). When samples were exposed to equimolar concentrations (0.5 mM) of InsP 6 and pNPP, the rate of phytase activity was reduced by 48% (P = 0.02, as revealed by ANOVA) compared to InsP 6 -only treatments, whereas the rate of PME activity was reduced by 18% (P = 0.38) compared to pNPP-only treatments.
Collections of canopy throughfall from beneath sycamore trees were made during three rainfall events, each in triplicate. Throughfall volumes collected per funnel were in the range 220-490 ml. All nine collections were analysed by HPIC but this did not detect compounds that corresponded with either InsP 6 , InsP 5 or InsP 4 , which have retention times > 18 min. Several unknown compounds with retention times < 18 min were detected; these were particularly abundant in late October when tree leaves were at an advanced stage of senescence/abscission. InsP 6 was readily detected in lily pollen leachate. Approximately 25 mg of pollen produced 57-74 lmols InsP 6 in solution. When E. prunastri was incubated in filtered pollen leachate, InsP 6 decreased in abundance concomitant with production of InsP 5 , the first hydrolysis product (Fig. 4).

Interspecies comparisons of phytase and PME activities
Phytase activity was detected in the majority of lichen species tested with the highest rates in the epiphytic species Bryoria fuscescens, E. prunastri, Usnea subfloridana and Pseudevernia furfuracea (Fig. 5). In all cases where InsP 6 depletion was detected, it was associated with production of InsP 5 . Lower phytase activity was measured in most other lichens with no activity being recorded in the terricolous mat-forming lichen Cladonia portentosa. Phytase activity was one to two orders of magnitude lower than that of PME, but activities of the two enzymes among the 13 species tested were significantly correlated (r = 0.77; P < 0.01). For both enzyme activities the effect of species was significant at P < 0.001 (ANCOVA).

Discussion
We show for the first time that phytase activity is readily detectable in lichens. Rates of activity among the 13 species tested ranged from (0-) 0.3 to 11.3 lmol InsP 6 hydrolysed g À1 h À1 . These values are broadly comparable with those documented for mycorrhizal roots and mycorrhizal fungi in axenic culture which are generally in the range 1-20 lmol P released g À1 h À1 (Bartlett & Lewis, 1973;Straker & Mitchell, 1986;McElhinney & Mitchell, 1993;Antibus et al., 1997;Colpaert et al., 1997), although rates > 100 lmol P released g À1 h À1 have occasionally been reported (Antibus et al., 1992;Ahmad-Ramli et al., 2013). The highest activities in lichens were recorded in epiphytic species, whereas those in the two terricolous lichens tested (C. portentosa and Peltigera membranacea) were the lowest. This might seem surprising for two reasons. First, there are large reserves of InsP 6 in many soils (Turner et al., 2002(Turner et al., , 2014, including podsols (Vincent et al., 2012), and thus a capacity to hydrolyse InsP 6 could be advantageous in lichens growing in contact with soil. However, it has been argued that mat-forming species of Cladonia (= subgenus Cladina) are largely insulated from soil chemistry by a basal layer of dead thallus or necromass (Crittenden, 1991) and Ellis et al. (2003Ellis et al. ( , 2004 provided evidence to suggest that these terricolous species derive little N from soil and are largely dependent on atmospheric deposits for N supply. Cladina spp. typically occur in open habitats where rainfall is probably

Research
New Phytologist modified little by plant canopies. By contrast, stemflow and throughfall experienced by epiphytic lichens on trees are frequently enriched with organic compounds leached from the canopy above, especially during autumnal leaf senescence in deciduous forests (Le Mellec et al., 2010). Organic C in throughfall includes Po Zimmermann et al., 2007) of which  speculate that inositol phosphates might be components. In the present work, InsP 6 was not detected in canopy throughfall beneath sycamore trees supporting large populations of E. prunastri. Other compounds detected in throughfall could have included organophosphates and hence potential substrates for lichen PME, because the analytical system detects a range of Po compounds including glucose-6-phosphate, ATP, ADP and pNPP. Nonetheless, this analysis does not provide unambiguous evidence of Po in canopy throughfall because it also detects other anions such as sulphates and oxalate (Chen & Li, 2003). Lysing cells of microorganisms and pollen on the thallus surface are additional potential sources of InsP 6 . Pollen, in particular, contains high concentrations of InsP 6 (2-3% of pollen mass) which may comprise up to 75% of total pollen Po . Wind-pollinated trees that dominate northern temperate and boreal regions produce significant quantities of pollen annually; for example, Saito and co-workers have estimated pollen production on a dry mass basis in more than 10 forest types in Japan with values ranging from 55 to 96 kg ha À1 yr À1 in Pinus densiflora/P. rigida forests (Saito & Takeoka, 1985) to 70-356 kg ha À1 yr À1 in stands of Quercus serrata (Saito et al., 1991). Lee & Booth (2003) noted that pollen deposition in P. densiflora forests was only 30-50% of the estimated pollen production and proposed that this insert in (f) shows selected data plotted as mean apparent rate of InsP 6 hydrolysis (V, lmols g À1 h À1 ) determined in assays of increasing duration (T, h). (g) Effect of temperature on rates of InsP 6 hydrolysis (r 2 = 0.915; P < 0.01); plotted values are the gradients of regression lines in (a-f) and SE bars are less than the diameter of the plotted points. (h) Effect of pH on phytase activity at 15°C (n = 10); rate of activity was determined from the difference between the quantities of InsP 6 hydrolysed in 0.5-and 5-h incubation periods; data are fitted to a twoparameter exponential decay (r 2 = 0.977; P < 0.01).

Ó 2015 The Authors
New Phytologist Ó 2015 New Phytologist Trust New Phytologist (2015) 208: 544-554 www.newphytologist.com discrepancy might arise from pollen interception by trunks and leaves. Several authors have speculated that pollen deposition results in ecologically significant inputs of N and P to terrestrial and lake systems, with estimates of P inputs to pine forests in the range 1-7 mg m À2 yr À1 (Doskey & Ugoagwu, 1989;Lee et al., 1996;Cho et al., 2003;Lee & Booth, 2003), of which up to 3.5 mg P m À2 yr À1 could be in the form of InsP 6 , whereas Hutchison & Barron (1997) and Read pNPP hydrolysis (open bars) in thalli (+ assay medium) and in thallus washings alone; samples were incubated in assay medium for 5 h and activity in the supernatant (= washings, unfiltered (UF) or filtered (F)) compared with that in thalli (either with or without a 5-h washing pretreatment) (*, significantly different from thallus incubations; one-way ANOVA with post-hoc analysis). (b, c) Relationships between phytase activity and (b) phosphomonoesterase (PME) activity (r = 0.88, P < 0.01), and (c) nitrogen (N) deposition (r 2 = 0.16, P > 0.05); data points represent samples collected at sites subject to different rates of wet N deposition. In all cases assays were performed on the apical 10 mm of thalli in either 1 mM InsP 6 or 10 mM pNPP and at pH 2.5 and 15°C in the dark. Rate of phytase activity in thalli was determined from the difference between the quantities of InsP 6 hydrolysed in 0.5-and 5-h incubation periods. Plotted values are means AE 1 SEM (n = 6 (a) or 10 (b, c)).  open bars) activities in 13 lichen species. Assays were performed using either 1 mM InsP 6 or 10 mM para-nitrophenyl phosphate (pNPP), at 15°C in the dark and at the pH optimum pertaining to the species (Table 1). Rate of phytase activity was determined from the difference between the quantities of InsP 6 hydrolysed in 0.5 and 5 h-incubation periods. Plotted values are means (n = 10) + 1 SEM.  & Perez-Moreno (2003) have highlighted the potential importance of pollen as a nutrient source for saprotrophic and mycorrhizal fungi, respectively. The present results raise the possibility that some epiphytic lichens could utilize InsP 6 leached from pollen (Fig. 4) and thus might contribute to P cycling in forest canopies (cf. Read & Perez-Moreno, 2003). Our initial demonstration of InsP 6 release from lily pollen now requires confirmatory studies with pollen from ecologically relevant tree species together with analysis of canopy throughfall collected during the pollen season. Dwarf shrubs in the Ericaceae and Empetraceae that are characteristic of heathlands are predominantly insectpollinated and emit smaller quantities of pollen (Rodr ıguez-Rajo et al., 2005), whereas pollen of the Graminaceae tested by  contained negligible InsP 6 . Hence, in the heathland habitats of C. portentosa pollen deposition might be a less significant source of nutrients for lichens.
Second, Peltigera spp contain cyanobacteria and have high N 2fixation capacities, high P demands (Crittenden et al., 1994;McCune & Caldwell, 2009) and, according to predictions of Houlton et al. (2008), might be expected to have appreciable phosphatase activities. Furthermore, at the collection site P. membranacea grows amongst mosses on decaying tree stumps and boulders in deciduous woodland and so, like epiphytes, will also be exposed to throughfall but not stemflow.
The initial hydrolysis of InsP 6 yielding InsP 5 is probably catalysed by a specific phytase (Meek & Nicoletti, 1986;Shan et al., 1993). Our analytical results are consistent with Ins(1,2,4,5,6)P 5 being the first hydrolysis product and with the lichen enzyme(s) responsible being a 3-phytase, that is it first attacks the 3-position on the inositol ring. 3-phytases occur in other fungi (e.g. Aspergillus ficuum; Chen & Li, 2003) in contrast to 6-phytases (yielding Ins(1,2,3,4,5)P5) and 5-phytases (yielding Ins (1,2,3,4,6)P5) that occur in plants ( Barrientos et al., 1994;Chen & Li, 2003;Jog et al., 2005), whereas all three enzyme types have been reported from bacteria (Sajidan et al., 2004;Puhl et al., 2008;Pontoppidan et al., 2012). Subsequent hydrolysis of lower-order inositol phosphate esters could be at least partly attributable to nonspecific PMEs (Meek & Nicoletti, 1986); indeed, data presented by Kemme et al. (1999) suggest that dephosphorylation of InsP 3 and InspP 2 by A. niger phytase is much slower than that of InsP 6 -InsP 3 and that final hydrolysis of InsP 1 might be achieved exclusively by nonspecific PME enzymes (i.e. not by phytase). Peaks broadly consistent with InsP 2 -InsP 1 increased in area with time indicating that dephosphorylation was ultimately complete, yielding inositol which could not be detected by the HPIC method. The proposed requirement for a specific phytase to initiate dephosphorylation of InsP 6 , but not that of InsP 5 -P 1 , is consistent with the observed progress in E. prunastri of InsP 6 hydrolysis which remained linear, even when potentially competitive lower-order inositol phosphates had accumulated in the assay medium ( Fig. 2a-f). Nevertheless, pNPP, the analogous substrate for PME, did significantly inhibit phytase activity. This could result from either a direct interaction between pNPP and phytase or an inhibitory effect of Pi released from pNPP hydrolysis which was typically 50-100 times faster than InsP 6 hydrolysis.
Phytase activity was pH dependent and hydrolysis products (e.g. InsP 5 and InsP 4 ) accumulated in the bathing medium. The hydrolysis of pNPP has similar properties which Bartlett & Lewis (1973), in their studies on mycorrhizal beech roots, considered indicative of extracellular PME activity. There is compelling evidence that PME activity in lichens is principally, if not entirely, associated with the fungus. Using an enzyme labelled fluorescent phosphatase substrate (ELF 97), Hogan et al. (2010a) showed that PME in C. portentosa was only detectable in the fungal symbiont and was concentrated in the outermost regions of the thallus. More recent work using ELF97 with other lichen species has shown PME activity to be restricted to the cortical cells (G. Brown & P. D. Crittenden, unpublished data). Furthermore, Hogan (2009) demonstrated PME activity in axenic cultures of C. portentosa and that it was positively related to the N : P mass ratio in the batch culture medium. It seems reasonable to suggest that extracellular phytase activity might have a distribution in the lichen thallus similar to that of surfacebound PME. However, the two enzymes differed in the extent to which they could be washed from E. prunastri into the bathing medium. Approximately 20% of total lichen PME activity was attributable to enzyme(s) leached into the assay medium, whereas phytase activity in lichen washings alone was not significantly different from that in either non-pretreated thalli plus washings or in thalli after 5 h washing (Fig. 3a); these results were confirmed in repeat experiments. It is possible that phytase has a substantially higher activity in free solution than in the thallus free-space where it might be partially immobilized and/ or in a different conformational state, such that in all lichen assays activity measured is largely that of the enzyme in free solution. This could make ecological sense if particulate matter such as pollen and microbial debris washed down from the canopy above is a principal source of InsP 6 although, equally, investment of N in secreted proteins seems prima facie an unlikely strategy in a lichen-forming fungus adapted to nutrientpoor conditions.
There was evidence in E. prunastri that the rate of InsP 6 consumption during short incubations (≤ 0.25 h) was atypically high. This putative artefact might have resulted from InsP 6 precipitation or its adsorption onto cell wall surfaces. Note that using Beckett's (1995) estimates of the free-space volume in lichens from xeric habitats (a mean of 32% of thallus volume), we estimate that the diluting effect of InsP 6 diffusing into the thallus free-space was probably negligible (0.3%). Pretreating thalli in InsP 6 for 0.5 h before assays excluded this apparent artefact from rate measurements. It is unlikely that further nonenzymic loss of InsP 6 could have been significant beyond 0.5 h for the following reasons: (1) InsP 6 consumption proceeded at a constant rate thereafter; (2) response to temperature was linear between 5 and 30°C and the temperature quotient between 15 and 25°C (Q 10 = 1.5) was broadly comparable with those for phytase activity in other microorganisms over the same temperature interval (e.g. c. 1.7 in Erwinia carotovora (Huang et al., 2009) (4) activity in assays on E. prunastri thalli was similar to that in filtered thallus washings. Rates of phytase activity in lichens are generally only a fraction of the rates of PME activity; similar proportionalities are evident in mycorrhizal roots and mycorrhizal fungi in axenic culture (Straker & Mitchell, 1986;Antibus et al., 1992Antibus et al., , 1997McElhinney & Mitchell, 1993;Colpaert et al., 1997). However, apparent K m values for the two enzyme systems in lichens when calculated from whole thallus data are similar: the K m value of 3.0 mM for phytase activity in E. prunastri can be compared with those for pNPP hydrolysis of 2.0 mM in the Antarctic species Usnea sphacelata (Crittenden et al., 2015), 2.2 mM in C. portentosa (Hogan et al., 2010a) and 8.9 mM in Cladonia rangiferina (Lane & Puckett, 1979). Further, the putative pH optima for phytase and PME activities in E. prunastri are similar (2.7 and 3.0, respectively; Lewis, 2012). Lichen species with high phytase activity also have high PME activity but the reverse is not necessarily true. Covariation between the two enzyme activities was striking when examined in a single species (E. prunastri) collected from locations with differing N deposition rates (Fig. 3b). Despite a growing body of evidence that upregulation of PME capacity in response to N enrichment is a general occurrence amongst a wide range of plants and microorganisms (Hogan et al., 2010b), neither phytase nor PME activities in E. prunastri were coherently related to wet N deposition (Fig. 3c). PME activity in lichens growing in treeless habitats is highly coupled to N deposition; PME activity in C. portentosa differed by a factor of 2 between British heathland sites with the lowest and highest N deposition rates (Hogan et al., 2010a) and in Antarctic fellfields PME activity in Usnea sphacelata differed by a factor of 6 between sites close to (2 km) and remote from (15 km) a penguin rookery (Crittenden et al., 2015). The weaker link between phosphatase activities and N deposition in E. prunastri might be explained by potential chemical modifications to rainfall that take place between interception by the tree canopy and deposition onto the surface of epiphytes.
The method used here for assaying phytase activity, which utilizes Chen & Li's (2003) HPIC-based quantification of inositol phosphates, has distinct advantages over methods based on Pi release. It measures depletion of a specific naturally occurring substrate and the production of intermediate products of hydrolysis, thus confirming that substrate depletion is not an artefact. The method can be used with living material because the substrate and the measured hydrolysis products appear not to be taken up by the test samples. Although it is to some extent reassuring that rates of phytase activity in lichens measured by InsP 6 depletion are similar to those documented for mycorrhizal fungi based on Pi release, it is perhaps surprising that rates in some lichen species are as high or higher than those in members of the Agaricomycotina, in which production of exoenzymes is a wellunderstood strategy for exploitation of organic substrates. It is possible that measurements of phytase activity in fungal-root systems based on the rate of Pi release from InsP 6 are underestimates due to the confounding effect of rapid active Pi uptake by physiologically active samples; a comparison of results produced by the two methods applied to the same test system is now needed to clarify this question. Disadvantages of the present method are that measurement of substrate depletion is generally less sensitive than measurement of product accumulation (cf. PME assay involving measurement of pNP production) and the analytical equipment is expensive to buy and maintain. Given the clear need for a robust phytase assay method, it is probable that the method used here will find application more widely in studies on plant root, fungal and soil systems.