By inhibiting soil enzymes, tannins play an important role in soil carbon (C) and nitrogen (N) mineralization. The role of tannin chemistry in this inhibitory process, in conjunction with enzyme classes and isoforms, is less well understood.
Here, we compared the inhibition efficiencies of mixed tannins (MTs, mostly limited to angiosperms) and condensed tannins (CTs, produced mostly by gymnosperms) against the potential activity of β-glucosidase (BG), N-acetyl-glucosaminidase (NAG), and peroxidase in two soils that differed in their vegetation histories.
Compared with CTs, MTs exhibited 50% more inhibition of almond (Prunus dulcis) BG activity and greater inhibition of the potential NAG activity in the gymnosperm-acclimatized soils. CTs exhibited lower BG inhibition in the angiosperm-acclimated soils, whereas both types of tannins exhibited higher peroxidase inhibition in the angiosperm soils than in gymnosperm soils. At all of the tested tannin concentrations, irrespective of the tannin type and site history, the potential peroxidase activity was inhibited two-fold more than the hydrolase activity and was positively associated with the redox-buffering efficiency of tannins.
Our finding that the inhibitory activities and mechanisms of MTs and CTs are dependent on the vegetative history and enzyme class is novel and furthers our understanding of the role of tannins and soil isoenzymes in decomposition.
Litter decomposition sustains ecosystem productivity through the recycling of nutrients stored in senesced plant tissues (Allison & Vitousek, 2005; Prescott, 2010). The process of decomposition is primarily mediated by microbial extracellular enzymes that facilitate the depolymerization and mineralization of the various components of plant litter (Allison et al., 2007; Sinsabaugh et al., 2008). Hence, the properties of plant biopolymers that directly or indirectly affect the catalytic efficiency of extracellular enzymes partially regulate the rate of this decomposition process. Tannins, which are functionally characterized as water-soluble polyphenolic compounds that readily complex with proteins (Hagerman et al., 1998a; Kraus, 2003b), are one of the prominent classes of plant macromolecules that can control the rate of decomposition. Tannin input to soils is reported from various ecosystems to range from 20 to 200 tannin m−2 yr−1 (Kraus et al., 2003b; Tharayil et al., 2012). Tannins interfere with decomposition biologically, through their toxic effects on microbial metabolism (Schimel et al., 1995), and chemically, by inactivating microbial enzymes (Joanisse et al., 2007). In addition, tannins leach from litter into the soil, where they undergo condensation and polymerization reactions with protein substrates, thereby protecting nitrogenous compounds from subsequent decomposition (Northup et al., 1995, 1998; Bradley et al., 2000; Hattenschwiler & Vitousek, 2000; Kraus et al., 2003a). The effect of condensed tannins on protein precipitation is well understood (Kraus et al., 2003a; Norris et al., 2011). However, the enzyme inhibition capacities of different tannin groups, the differential susceptibilities of various isoenzymes to tannin inhibition, and the possible contribution of the redox quenching capacity of tannins to enzyme inhibition remain relatively unexplored. This lack of knowledge could limit the ability to predict the tannin-induced inhibition of soil processes in various ecosystems.
Based on their monomeric composition, tannins can be broadly divided into two classes: condensed tannins (CTs) and hydrolysable tannins (HTs). Condensed tannins are comprised of catechin and epicatechin units linked by C–C interflavon bonds at C4–C8 or C4–C6 positions, and the hydroxylation pattern of the B-ring allows them to be further subdivided into procyandins (PCs; dihydroxylated B-ring) and prodelphinidins (PDs; trihydroxylated B-ring; Schofield et al., 2001; Kraus et al., 2003a). HTs are composed of gallic acid units linked via an ester linkage to a glucose moiety, and they can be further classified based on the absence or presence of C–C linked galloyl groups as gallotannins and ellagitannins, respectively (Mueller-Harvey, 2001). In general, gymnosperms produce CTs but lack the ability to produce HTs, whereas most angiosperms produce a mixture of HTs and CTs, collectively referred to as mixed tannins (MTs). The overall protein complexation capacity of tannins depends on the composition of condensed vs hydrolysable tannins, the hydroxylation pattern of the B-ring, the extent of polymerization, the types of cross-linkages between monomeric units, the substitution pattern of the A-ring, and the cis vs trans conformation at C2–C3 (Kraus et al., 2003a; Maie et al., 2003; Nierop et al., 2006). The chemical reactivity of tannins is traditionally defined by their protein complexation affinities (Kraus et al., 2003b). Condensed tannins that interact with proteins through hydrogen–hydrogen bonding (between amine and hydroxyl groups) are thought to exhibit greater protein complexation efficiency on a per weight basis than hydrolysable tannins that associate with proteins through hydrophobic interactions (Hagerman et al., 1998a). Thus, CTs are anticipated to play a greater ecological role in inhibiting microbial enzymes (Kraus et al., 2004). However, the protein complexation is a poor predictor of enzyme inhibition (Juntheikki & Julkunen-Tiito, 2000) and hence, the influence of tannins on soil carbon (C) and nitrogen (N) dynamics could be less associated with their protein precipitation characteristics (Norris et al., 2011). Considering the higher overall reactivity the tannins might attain as a result of the association of HT subunits, the MTs produced by angiosperms could exhibit efficient enzyme inhibition capacities that may be similar or more ecologically relevant than CTs. As a consequence of their pro-oxidant activities at higher pH, ellagitannins are thought be better herbivore defense compounds than condensed tannins (Barbehenn et al., 2006; Salminen & Karonen, 2011).
In addition to the structural features of tannins, the formation of tannin–protein complexes could also be governed by the structural characteristics of the proteins involved. Proteins with a higher structural flexibility, such as gelatin, have been shown to interact better with tannins compared with compact protein molecules, including cytochrome, lysozyme and myoglobin (Hagerman & Butler, 1981). Additionally, proteins that are abundant in hydrophobic amino acids are more susceptible to complexation with tannins than those containing charged amino acids (Hagerman & Butler, 1981; Bacon & Rhodes, 2002). From a decomposition perspective, similar to tannins, soil enzymes are functionally defined based on the characteristics of the chemical reaction they catalyze (Caldwell, 2005). Hence, enzyme classes that perform similar catalytic functions could differ structurally, resulting in different isoforms of the enzymes (isoenzymes; Di Nardo et al., 2004; Khalil et al., 2011; Stone et al., 2012; German et al., 2012). The structural differences that accompany the existence of different isoforms include variations in protein size, the number of domains, amino acid qualities, the percentage of glycosylation, the isoelectric point and pH optimum (Caldwell, 2005; Baldrin & Valášková, 2008). For example, bacterial β-glucosidase (which catalyzes the cleavage of β-1,4 linkages in polysaccharides) ranges from 50 to 210 kDa in size, while fungal β-glucosidase, which performs the same function, exhibits an average size of c. 100 kDa (Barrera-Islas et al., 2007; Baldrin & Valášková, 2008; Canizares et al., 2011). The production of these enzyme isoforms could be determined by the identity of the microbes involved, which, in turn, is a function of the abiotic and biotic conditions, including the vegetation history (Strickland et al., 2009b; Stone et al., 2012), to which the microbes are acclimated. Enzyme isoforms that perform the same catalytic function could exhibit different tannin-binding affinities as a consequence of differences in their protein structures, which in part are governed by the litter quality to which the microbial community is acclimatized. Our current understanding of how vegetation history affects the susceptibility of isoenzymes to tannin complexation is limited, and this lack of knowledge could hinder our ability to predict soil processes during plant invasions (Tharayil et al., 2012). Additionally, this differential inhibition of isoenzymes could cause shifts in the microbial community in ecosystems where the resident isoenzymes are more susceptible to novel tannins and thus fail to sustain the native microbial communities.
In addition to binding directly to enzymes, tannins can reduce the catalytic efficiency of enzymes by altering the chemical environment in which they operate. Functionally, soil enzymes are broadly categorized as hydrolases and oxidoreductases (Allison et al., 2007). The hydrolytic enzymes are substrate-specific because their identity is strictly based on the chemical bonds they cleave in their substrates (Caldwell, 2005). By contrast, oxidoreductases, including lignin peroxidases, manganese peroxidases and laccases, are less substrate-specific and undertake oxidation based on the redox potential of the substrate (Wang, 2009). Lignin peroxidases (LiPs) undergo two-electron oxidation in the presence of H2O2 as the electron acceptor, forming the oxidized intermediaries LiP-I and LiP-II (Wang, 2009). In turn, these intermediates oxidize aromatic substrates with a lower redox potential (< 1.4 V at pH 3) to form unstable phenoxy radicals that propagate ring cleavage in lignins. As a consequence of the inherent antioxidant properties of tannins (Hagerman et al., 1998b; Wei et al., 2010), they could form resonance-stabilized phenoxy-radicals upon donation of an electron (Bors & Michel, 2002), and could readily prevent the formation of a reactive enzyme intermediary or quench an oxidized enzyme (Hemeda & Klein, 2006), resulting in the preferential protection of other substrates from oxidation by peroxidase.
The objectives of this study were to compare the inhibition efficiencies of tannins produced by angiosperms (MTs) and gymnosperms (pure CTs): against the different soil enzyme functional classes and the isoforms within the same functional class that were acclimated to different vegetation histories. We hypothesized that, contrary to the traditional understanding, MTs produced by angiosperms would exhibit a similar or greater enzyme inhibition capacity than CTs produced by gymnosperms; the enzymes in soils acclimatized to gymnosperms would be more inhibited by MTs, and the inhibitory effect of CTs in soils acclimatized to angiosperms would be on a par with that of MTs; and, because of their dual mode of interaction (protein complexation and redox buffering), tannins would exhibit greater inhibition of oxidoreductases than hydrolases.
Materials and Methods
Sites and soil description
Soils were collected from two forest ecosystems common to the southeastern USA: an Acer rubrum L.–Quercus alba L. forest (AQ soil) and a Pinus echinata Mill forest (Pinus soil). The stands were located in the Clemson University Experimental Forest, c. 16 km apart, and had been established for > 100 yr. The two sites had similar soil type (Typic Kanhapludult, described as a pacolet fine sandy loam soil), climatic conditions and soil pH (5.0). These two stands were chosen because they represented distinct differences in the tannin chemistries of the litter, with Acer–Quercus producing an HT + CT mixture (MTs) and Pinus solely producing CTs. At both sites, the understory vegetation was sparse, minimizing the impacts of other vegetation on soil enzyme dynamics. During April 2011, soil samples were collected from each stand at three randomly chosen sampling points that were at least 15 m apart. At each sampling point, the surface litter and organic matter layer were carefully removed, and the mineral soil was sampled to a depth of 5 cm. To increase the homogeneity of the sampling, at each sampling point, three cores were collected along the three corners of an equilateral triangle (20 cm). The cores were combined, mixed, placed on ice, and then immediately transferred to the laboratory. There, the samples were further homogenized, sieved through a 2-mm mesh and stored at −20°C for analysis.
MTs were purified from Acer rubrum, Quercus alba and Betula papyrifera (Marsh.), and CTs were isolated from Pinus banksiana (Lamb.), Abies balsamea L., Picea mariana Mill., Vaccinium boreale I.V. H & A, and Thuja plicata D. Donn. The CTs are those previously described by Norris et al. (2011); the B. papyrifera tannin, from foliage collected in Newfoundland, has not been previously reported and was prepared similarly to the other CTs. The A. rubrum and Q. alba tannins were extracted as described by Tharayil et al. (2011). Briefly, ground dry plant material was extracted with 70% acetone three times and once with 100% methanol. The extracts were pooled and evaporated under nitrogen. The aqueous phase was extracted three times with diethyl-ether, rotovaped and left under nitrogen to remove any residual solvent before loading it onto a Sephadex LH-20 column preconditioned in 50% ethanol. The column was washed with 50% ethanol to remove low-molecular-weight compounds, and the sorbed tannins were eluted with 75% acetone. The eluent was concentrated under vacuum and further evaporated under nitrogen to remove acetone, and the remaining aqueous solution was freeze-dried to recover the purified tannin. Tannins were characterized using solution-state 13C nuclear magnetic resonance as described previously (Norris et al., 2011; Tharayil et al., 2011; Supporting Information Table S1).
Condensed tannin assay
Purified tannin at a concentration of 1000 μg ml−1 was dissolved in 100% methanol for analysis. Proanthocyanidins were oxidatively cleaved by adding 3 ml of acid-butanol reagent (butanol and 12 N HCl (95 : 5) (v/v)) and 100 μl of Fe reagent (250 mg ferric ammonium sulfate to 10 ml of 12 N HCl) to 50 μl of a sample. The reaction mixture was incubated for 50 min at 95°C and cooled to room temperature before analyzing the cyanidin content at 550 nm. Multiple tannin concentrations were assayed and three analytical replicates were analyzed per concentration per species. Linear regressions with zero intercepts were fit to these data and the slope (1000 × AU (mg tannin)−1) were computed for comparison (Tharayil et al., 2011).
Hydrolysable tannin assay
The hydrolysable tannins were acid-hydrolyzed to gallic acid units, which were then methylated to methyl gallate (Hartzfeld et al., 2002). Briefly, 100 μl of a sample (1000 μg of tannin in 1 ml of 100% methanol) was incubated with 2 ml of methanol and 400 μl of concentrated H2SO4 at 85°C for 15 h. The extract was analyzed to quantify the methyl gallate content using a high-pressure liquid chromatography (HPLC) system equipped with a photodiode array detector, as per the conditions described by Tharayil et al. (2011).
Protein precipitation assay
The radial diffusion assay characterizes the protein precipitation capacity of tannins based on the dimensions of opaque rings (precipitates) that are formed upon incubating tannins in an agar medium containing a model protein. Because of its ecological similarity to soil enzymes, β-glucosidase from Prunus dulcis (almond; ABG) was used as the model protein to determine the protein precipitation capacity in our study. The assay was performed in triplicate according to Norris et al. (2011) using tannin concentrations optimized in preliminary studies. Briefly, plates were prepared with 0.1% ABG, and 14-μl aliquots of tannin were pipetted into each well and were incubated for 48 h at 25°C. The volumes of the opaque rings obtained were used to determine the amount of precipitated protein. The mass ratio of proteins to tannins (μg μg−1; P : T) was used as a measure of the protein precipitation capacity of the tannins.
Almond β-glucosidase enzyme inhibition
The purpose of the ABG inhibition assay was to compare the inhibition capacities of MTs and CTs in an isolated system with ample potential enzyme activity and minimal interference from soil. The difference in the potential ABG activity with and without tannins was monitored based on the formation of saligen through the cleavage of β-1,4 linkages of the substrate salicin (Juntheikki & Julkunen-Tiito, 2000; Tharayil et al., 2011). In short, 3000 μg ml−1 of salicin and 8 μg ml−1 of β-glucosidase were incubated with 0, 5, 10, 20, 40, and 80 μg ml−1 tannins at 25°C for 20 min. The final volume of the reaction mixture across all tannin concentrations was kept the same, and the reaction time was optimized based on preliminary experiments. The reaction was stopped by the addition of 500 μl of 6 N H2SO4. The amount of saligen produced was quantified using an HPLC system equipped with a photo diode array detector (Shimadzu Corporation, Kyoto, Japan). Chromatographic separation was performed on a Gemini-C18 column (250 × 4.6 mm; Phenomenex, Torrance, CA, USA). The solvent conditions were as follows: 22.5% acetonitrile, run isocratically for 7 min at a flow rate of 0.8 ml min−1. Quantification was based on the peak area at 273 nm. The percentage of enzyme inhibition was calculated based on the product formation in the presence of tannins relative to the product formation in the absence of tannins:
The results were further nonlinearly regressed to a hyperbolic function similar to the Michaelis–Menton kinetics to acquire the half-saturation constant of inhibition (ki) and maximum inhibition (Imax; Sigmaplot v12.0.; Systat Software, Inc., San Jose, CA, USA).
Tannin antioxidant capacity
The radical quenching capacity of the tannins was quantified to assess the ability of tannins to hamper potential peroxidase activity. A common means of measuring the antiradical capacity is to monitor the depletion of 2,2-diphenyl-1-picrylhydrazyl (DPPH˙) spectrophotometrically when combined with antioxidant compounds. The scavenging capacity is defined as the ability to donate hydrogen or electrons to the radical, thereby causing the formation of a colorless compound (Chaillou & Nazereno, 2006). A working concentration of DPPH˙ (38 μg ml−1) was prepared in 100% methanol, and the tannins were prepared as 1000 μg ml−1 stocks in 100% methanol. The appropriate volume of the tannin stock was added to DPPH˙ to prepare working tannin concentrations of 2.5, 5 and 10 μg ml−1, and the volumes were adjusted accordingly with methanol. The experiment was optimized for 1.0 AU in the absence of tannins. The reaction mixture was shaken at room temperature for 15 min (26 rpm) and absorbance at 517 nm was measured immediately. The antioxidant capacity of the tannins was determined as follows:
Soil enzyme inhibition
The potential activities of peroxidase (PER), β-1,4-glucosidase (BG), and β-1,4-N-acetylglucosaminidase (NAG) in the AQ and Pinus soils were monitored in the presence and absence of CTs and MTs. To allow comparison of enzyme inhibition across tannins and sites, the amount of soil slurry used in the respective assays was optimized to achieve equivalent potential enzyme activities in the Pinus and Acer–Quercus soils in the absence of tannins (amount of slurry varied < 6% between the two soils for various enzymes). The tannin concentrations tested varied from 5 to 160 μg ml−1 which, based on the soil : buffer ratio, ranged from 2 to 64 μg g−1 soil and are ecologically relevant (Kraus et al., 2003b; Joanisse et al., 2007).
The potential peroxidase enzyme activity and its inhibition were determined colorimetrically using 3,3′5,5′-tetramethylbenzidine (TMB) as the substrate (Johnsen & Jacobsen, 2008) at tannin concentrations of 0, 5, 10, 20, 40 and 80 μg ml−1. The product of the oxidation of this substrate has a higher molar absorptivity (Josephy et al., 1981) than the traditionally used substrate L-Dopamine and, hence, was found to be more sensitive for tracking tannin inhibition. Soil slurries were prepared by homogenizing 0.5 g of soil in 150 ml of 50 mM acetate buffer (pH 5.0). The appropriate volume of soil slurry was added to each reaction vial to provide similar potential enzyme activities between the two sites, and the slurry was incubated with various tannin concentrations. The final volume of the assay mixture was adjusted to 400 μl using acetate buffer. The tannin and slurry mixture was incubated on a rotatory shaker (26 rpm) for 20 min at 25°C to provide ample time for tannin–enzyme complexation. The substrate, TMB (200 μl; 250 μg ml−1), was then added, and the reaction mixture was further incubated on the rotary shaker. This reaction was stopped after 20 min by the addition of 1 ml of 5% H2SO4. The sample absorbance was measured in a spectrophotometer at 450 nm.
Beta-glucosidase and NAG inhibition assays were performed using the same assay conditions as described above, with the following modifications. The potential activities of BG and NAG were quantified by measuring the 4-methylumbelliferone (MUB) formed after the cleavage of the MUB-linked substrates 4-methylumbelliferyl-β-d-glucopyranoside and 4-methylumbelliferyl-N-acetyl-β-d-glucosaminide, respectively. Based on preliminary experiments that showed lower tannin inhibition of these hydrolases, the tannin concentrations were increased to 0, 20, 40, 80 and 160 μg ml−1, and the substrate concentration was kept at 700 μg ml−1. After incubation, the reaction was terminated by adding 1 ml of 100% methanol to each reaction vial. Because the MUB fluorescence readings in the microplate reader varied greatly (especially in the presence of tannins), the MUB content was quantified using an HPLC system as described above, but with a fluorescence detector. The mobile phase consisted of 40% acetonitrile in 10 mM sodium acetate buffer (pH 5.0), which was run isocratically for 7 min. The MUB contents were quantified using Ex-Em wavelengths of 315 and 475 nm, respectively, and the MUB retention time was 6.5 min. The per cent inhibition of the enzymes was computed using Eqn (Eqn 1).
The protein precipitation capacity of the tannin was analyzed using a mixed-model analysis of variance with species as the fixed effect. For the remaining analyses, individual species were nested within the tannin type (MTs vs CTs) as a random effect. The half-saturation constant of inhibition (ki) and maximum inhibition (Imax) of ABG were analyzed using a mixed-model analysis of variance with the tannin type as the fixed effect with species nested within tannin type as a random effect, and the individual species were compared by one-way ANOVA. The nature of interaction of tannins with proteins changes with respect to the tannin : protein ratio (Siebert et al., 1996; Jöbstl et al., 2004; Poncet-Legrand et al., 2006; Pascal et al., 2007), and hence it proved to be less robust to compare the overall inhibition across all tannin concentrations. Hence, in the soil enzyme assays, ANOVA was performed at individual tannin concentrations, and contrasted how the response to that particular tannin concentration varied across the tannin types and soils. The inhibition of the soil enzymes of BG, NAG, and PER was analyzed with a mixed-model ANOVA with tannin type and site as fixed effects and species nested within tannin type as a random effect, followed by Tukey's HSD test. The antioxidant capacity was analyzed at each tannin concentration using a mixed-model ANOVA with tannin type as the fixed effect. The antioxidant capacity was further regressed with the structural data obtained through 13C NMR analysis of CTs (Method S1) and the peroxidase inhibition at 10 μg ml−1 in Pinus and AQ soils. All of the statistical analyses were conducted using sas 9.2 (SAS Institute, Cary, NC, USA), and an alpha value of 0.05 was employed for all hypothesis tests.
The acid-butanol assay measurements were based on the susceptibility of the CT polymer to undergo oxidative cleavage in an acidic medium in the presence of a catalyst (Fe). The susceptibility to oxidative cleavage differed according to the molecular identity of the CT involved, including the chain length, PC/PD content and cis vs trans ratio of the tannins (Method S1). Overall, there was an 11-fold difference in the value corresponding to the absorbance of procyanidins from the eight investigated species (Table 1). The response of the acid-butanol assay on a per weight basis was consistently higher for the tannins from the gymnosperms (pure CTs) compared with those from the angiosperms (MTs; F1,6 = 14.87; P =0.008). The tannins from the angiosperms (A. rubrum, Q. alba and B. papyrifera) differed in their HT content, as determined by the acid hydrolysis assay (F2,6 = 29.3; P =0.008). The A. rubrum and Q. alba tannins exhibited similar HT concentrations by mass, with a mean of 26%, while B. papryifera presented 15% HT by mass (P = 0.001; Table 1). Collectively, there was a three-fold difference in the protein precipitation capacities of individual tannins (Table 1). Tannins from V. boreale, (pure CT species) and Q. alba (MT species) showed higher protein precipitation capacities and precipitated an equivalent weight of protein per weight of tannin (P : T c. 1.0). Overall, however, the protein precipitation capacities of the CTs and MTs did not differ statistically (F1,6 = 1.32; P =0.2449).
Table 1. Chemistry characterizations of mixed (MT) and condensed (CT) tannins
Protein precipitation capacity, given as P : T mean (± SE); that is, mass of β-glucosidase precipitated per mass tannin. The same letters represent statistical similarity by one-way ANOVA with the individual species.
Condensed tannin reactivity. Similar letters represent statistical similarity by one-way ANOVA with the individual species.
0.809 (± 0)b
0.788 (± 0.07)b
1.01 (± 0.04)a
0.714 (± 0.03)bc
1.10 (± 0.01)a
0.369 (± 0.0)d
0.57 (± 0)c
1.03 (± 0)a
Almond β-glucosidase inhibition
Tannins from B. papryifera exhibited the highest maximum inhibition (Imax) of ABG (Imaxc. 80%; Table 2) whereas P. mariana exhibited the lowest maximum inhibition (Imaxc. 36%). The highest half-saturation constant of inhibition (ki) was observed for T. plicata, whereas Q. alba exhibited the lowest ki (c. 16.02). The maximum inhibition (Imax) of ABG by MTs was 54% greater than that of CTs (F1,6 = 20.53; P =0.004; Table 2). The half-saturation constant of MTs and CTs did not differ statistically (kic. 30; F1,6 = 0.466; P =0.520).
Table 2. Tannin inhibition kinetic parameters of almond beta-glucosidase
Half-saturation constant of inhibition by tannins. The same letters represent statistical similarity by one-way ANOVA with the individual species.
Maximum inhibition capacity of almond-beta glucosidase by tannins. The same letters represent statistical similarity by one-way ANOVA with the individual species.
Overall, the radical quenching capacity of tannins was dependent on the tannin type (Fig. 1), and MTs exhibited 63, 65 and 42% higher radical quenching capacities than CTs at 2.5, 5 and 10 μg ml−1, respectively (F1,6 = 18.71; P =0.002). The linear regression analysis relating the radical quenching capacity of CTs at 10 μg ml−1 to their % PC content (Table S1) revealed a negative association (Fig. 2a; R2 = 0.788; P =0.0442). Additionally, there was a positive correlation between the radical scavenging capacity of tannins and their corresponding peroxidase inhibition at a 10 μg ml−1 tannin concentration in the Pinus soils (Fig. 2b; R2 = 0.57; P <0.0001); however, the radical scavenging capacities of the tannins were not correlated with their inhibition of peroxidase in the AQ soils (Fig. 2c; P =0.85).
Soil peroxidase enzyme inhibition
Depending on the tannin concentration, the inhibition of soil peroxidase was affected by site or site × tannin type interaction. At 5 μg ml−1, the inhibition of peroxidase by tannins was affected by the site (Fig. 3; F1,38 = 58.34; P <0.001), with tannins exhibiting 30% greater inhibition of the potential peroxidase activity in the AQ soils than in the Pinus soils (P <0.001). At tannin concentrations of 10 and 20 μg ml−1, there was a site × tannin type interaction (F1,38 = 40.72; P <0.001) for soil peroxidase inhibition. At 10 μg ml−1, the MTs and CTs exhibited similar levels of peroxidase inhibition in the AQ soil, which were 16% and 52% higher than the inhibition levels found in Pinus soil due to MT and CT, respectively (P <0.001). Similarly, at 20 μg ml−1, the MTs and CTs showed statistically similar levels of peroxidase inhibition in AQ soil, which were higher than the inhibition levels of peroxidase in Pinus soils (22% and 34% higher due to MTs and CTs, respectively). At 40 μg ml−1, the inhibition of peroxidase was dependent on the site (F1,38 = 49.73; P <0.001), with the inhibition of peroxidase being 9% higher in the AQ soils than in the Pinus soils (P <0.001). At 80 μg ml−1, both tannin types exhibited 100% peroxidase inhibition in both of the soils.
Soil hydrolase inhibition
Generally, the inhibition of hydrolases (BG and NAG) by MTs was similar to or greater than the inhibition by CTs in both soils. At 20 μg ml−1, both MTs and CTs inhibited soil BG activity in both soils at c. 28% (Fig. 4; P >0.05). The inhibition of BG exhibited a significant site × tannin type interaction at 40 μg ml−1 (F1,38 = 6.05; P =0.018) and 80 μg ml−1 (F1,38 = 19.22; P <0.001). At 40 μg ml−1, the CTs showed lower BG inhibition in AQ soils, (P =0.015), whereas at 80 μg ml−1, the CTs presented higher BG inhibition in Pinus soils (P <0.001). At 160 μg ml−1, the inhibition of BG was dependent on the site (F1,38 = 6.74; P =0.013) and tannin type (F1,6 = 12.93; P =0.011), and the BG activity in the Pinus soils was marginally more inhibited by tannins than the BG in the AQ soils (F1,14 = 4.15; P =0.05). Furthermore, the MTs exhibited 33% greater inhibition of BG activity than the CTs (P =0.003).
The inhibition of NAG showed a significant site × tannin type interaction at tannin concentrations of 40 μg ml−1 (F1,38 = 4.86; P =0.033) and 160 μg ml−1 (F1,38 = 6.83; P =0.012) (Fig. 5). Compared with CTs, MTs exhibited 40% inhibition of the NAG present in the Pinus soils.
Comparison of enzyme inhibition by tannins produced by gymnosperms and angiosperms
In ecological studies, CTs are traditionally thought to possess a higher enzyme inhibition capacity than MTs, which is often corroborated through protein precipitation assays. However, the protein precipitation capacity may be less representative of the enzyme inhibition capacity of tannins (Julkunen-Tiitto & Meier, 1992), and presence of HT units could enhance the overall reactivity of MTs. In the ABG assay, inhibition by tannins resembled a traditional hyperbolic curve, indicating saturation of the sorption sites on ABG that are accessible to tannins. Although the half-saturation constants of inhibition were similar between the two tannin types, MTs exhibited a higher Imax than CTs, indicating that MTs were a more potent inhibitor of ABG. In the agar-based protein precipitation assay, the capacity of tannins to precipitate ABG was below unity (protein : tannin ratios of 0.92 for pure CTs and 0.65 for MTs; Table 1). By contrast, in the enzyme inhibition assays, a 1 : 1 protein : tannin ratio resulted in < 15% inhibition of ABG, indicating that precipitation of the enzyme could take place before inhibition of its catalytic activity, and that a higher concentration of tannin was required to cause enzyme inhibition. This result indicates that the sorption of tannins to enzymes is not preferentially directed towards their catalytic sites. The observed enzyme inhibition at a high tannin : protein ratio could partly be attributable to the conformational changes in protein folding due to the excessive cross-links between the sorbed tannins and protein and/or physical blocking of catalytic sites, resulting in noncompetitive inhibition of the enzyme.
The disparity between enzyme inhibition and protein complexation is reflected well by the observation that, even though the protein precipitation assay was not able to differentiate between the reactivity of CTs and MTs, in accordance with our hypothesis, MTs exhibited a higher Imax of ABG inhibition compared with CTs. This higher inhibition capacity of MTs could have been attributable to their structural flexibility (Fletcher et al., 1976), which would enable MTs to cross-link with the protein at multiple binding sites. The tannins from A. rubrum presented one of the highest ABG inhibition capacities of all of the examined tannins (Imax of 77), but exhibited the lowest protein precipitation capacity. This result further bolsters the argument that the protein precipitation and enzyme inhibition, though operationally similar (caused by the binding of tannins to proteins), could yield different functional results (enzyme inhibition and/or protein precipitation) depending on the composition and concentration of tannins. Additionally, the relatively high ABG inhibition capacity of MTs could be partly attributed to the differential affinity of HT and CT units for binding sites of the same enzyme. In MTs, the affinity of HT and CT subunits for ABG complexation differs among species (Method S2; Fig. S1), with the HT units in A. rubrum tannins exhibiting a higher affinity for ABG complexation than those in Q. alba tannins, possibly as a result of differences in structural linkages within and between these subunits.
Linking the antioxidant potential of tannins to their soil peroxidase inhibition
The redox potential of tannins is lower than the redox potential of H2O2 (which facilitates the activity of lignin peroxidase) (Wang, 2009) and that of peroxyl radicals (Hagerman et al., 1998b). Therefore, antioxidants such as tannins could readily quench or prevent the formation of radical intermediates of the peroxidase enzyme. Also, by chelating the soil Fe and manganese (Mn), tannins could prevent the Fenton reaction that generates hydroxyl radicals (Lopes et al., 1999), thus restricting the nonenzymatic oxidation of substrates and Mn peroxidase enzymes respectively (Sinsabaugh, 2010). In accordance with our hypothesis, the MTs exhibited higher antioxidant capacities than the CTs, and this higher radical quenching capacity of MTs could be explained by three mechanisms. First, compared with the monomers of CTs (catechin), the monomers of HT (penta-O-galloyl-d-glucose) exhibit 15 hydroxyl groups and a lower redox potential (Hagerman et al., 1998b), which increases their propensity for radical quenching. Secondly, the phenolic-coupling reaction that follows the abstraction of an electron by peroxidases is more favored in gallotannins, as it is an intra-molecular process in this case (resulting in the formation of ellagitannins), compared with CTs, for which it is an intermolecular process (Bors & Michel, 2002). These coupling reactions could, in turn, result in the reproduction of additional oxidizable phenolic moieties (Hotta et al., 2001; Bors & Michel, 2002) in the polymeric products, conferring a higher radical-scavenging capacity on these condensed HTs. Thirdly, the hydroxyl group attached to the para position of the phenolic ring serves as the primary site of radical quenching, whereas the meta hydroxyl group is involved in the stabilization of the phenolic radical (Stojanović et al., 2001; Bors & Michel, 2002). Thus, gallic acid with a para hydroxyl group would be more reactive than procyanidins. This position-dependent antioxidant property of the phenolic ring is further evident from our observation that the antioxidant capacity of CTs was inversely related to the per cent PC content. This result is also in agreement with the observation that the tannins that are abundant in prodelphinidins are more reactive than the tannins that are abundant in procyanidins (Nierop et al., 2006).
The radical quenching capacity of tannins was a robust predictor of their inhibition of the peroxidases present in Pinus soil, but not in AQ soil. This result would suggest that, even though the peroxidases in the two soils are functionally similar, these enzymes could be dissimilar in the redox states of their intermediaries and, hence, their preference with respect to accepting electrons from tannins. Additionally, once these enzymes bind to tannins, the antioxidant capacity of the tannins is reduced by < 50% (Riedl & Hagerman, 2001). Thus, the lack of correlation between the tannin-quenching capacity and the potential peroxidase activity in the AQ soils may also be attributed to the structural differences in the peroxidases from the two sites, which would alter their tannin-binding affinities. Because we normalized the potential enzyme activities of the two sites before the assay, the differences in the inherent redox potentials of the two soils systems should have had less of an influence on the above results. Irrespective of the tannin type and site history, at all tannin concentrations, the peroxidase inhibition was two-fold greater compared with the hydrolase enzymes. This result supports our hypothesis that the antioxidant properties of tannins could prove to be selectively inhibitory towards peroxidase. Although the potential role of antioxidants in soil organic matter dynamics has been studied previously (Rimmer, 2005; Rimmer & Abbot, 2011), this is the first investigation to suggest an alternate mechanism of peroxidase inhibition in soil matrices. The complete inhibition of peroxidase observed in the present study at relatively low tannin concentrations (40 μg ml−1) could help to explain the higher accumulation of litter and slower C mineralization found in polyphenol-rich ecosystems (Tharayil et al., 2012). This result also could provide a partial explanation for the general observation that litter layers that contain a high abundance of phenolic substrates often exhibit lower peroxidase activity (Sinsabaugh, 2010).
Effect of vegetation history on the susceptibility of native enzymes to inhibition by tannins
Because the substrate identity driven by plant community could be one of the many drivers of microbial community composition (Ayres et al., 2009; Strickland et al., 2009a,b), the microbial isoenzyme composition could also be acclimatized to the inhibitors co-occurring with the substrate. In parallel with the hypothesis presented above, the N-acetyl-glucosaminidases in the Pinus soils were more prone to inhibition by MTs than by CTs. The lower NAG inhibition by MTs detected in the AQ soils could be attributed to the acclimatization of the soil microbial community to MTs, resulting in the abundance of isoforms of NAG which was less susceptible to inhibition by tannins. The acclimation of microbial enzymes was further demonstrated by the finding that the BG activity in the AQ soils was less effectively inhibited by CTs because the microbes in the AQ soils were acclimatized to the CTs within MTs. Joanisse et al. (2007) similarly determined the soil enzyme inhibition by CTs isolated from an invasive plant to be greater than that of the native tannins, which bolsters our findings. Additionally, Bending & Read (1996b) observed that the proteases of the ericoid mycorrhizas maintained activity in the presence of tannic acid (purified HTs) while the enzymes of ectomycorrhizas were inhibited. Similarly, Ximines et al. (2011) found that beta-glucosidase of Asperelligus niger was five times more resistant to inhibition by tannic acid than that of Trichoderma reesei. Combined, these investigations support our finding that, even though enzymes are operationally similar, they could exhibit different reactivity towards tannins in the soil based on the isoforms. These observed responses could also have been amplified due to differences in the catalytic efficiencies of isoenzymes in the two soils (Gerday et al., 2000), or the enzyme state (prior interaction with inhibitors reducing available tannin binding sites before experimental analysis; Zimmerman & Ahn, 2011).
Irrespective of the tannin type involved, peroxidase activity in the AQ soils was consistently more inhibited than that of the peroxidases in the Pinus soils. Bending & Read (1996a) determined that the oxidoreductases of ericoid mycorrhizas were up-regulated in the presence of tannins while those of ectomycorrhizas were not stimulated, again bolstering the conclusion that the soil microbial composition could define oxidoreductase sensitivity to tannins. Within MTs, the differences in fine-level structure could impose differential inhibition of enzymatic activity. For example, Sinsabaugh et al. (2002) reported differential cumulative enzyme activity in decomposing A. rubrum and Q. alba litter, with A. rubrum litter requiring a higher peroxidase and cellobiohydrolase activity and lower phenol oxidase and BG activity than Q. alba litter per unit mass loss. This could be partially attributed to the differential sensitivity of enzymes to inhibitors in the litter.
Oxidative degradation pathways control the release of N from organic matter (Northup et al., 1995; Sinsabaugh, 2010). The formation of tannin–protein complexes, which are relatively resistant to degradation by enzymes (Adamczyk et al., 2009), limits the release of protein bound in the complexes. In the present study, the levels of peroxidase and NAG inhibition caused by tannins were inversely related to each other. This result is in agreement with the findings of Sinsabaugh & Follstad Shah (2011), who reported that soil oxidoreductase activity was inversely related to N-acetylglucosaminidase activity but positively associated with protease activity. The inverse relationship observed for these enzymes suggests that the capacity of NAG to avoid enzymatic inactivation under poor potential peroxidase activity conditions would help to supply organic N through an alternative N source, such as chitin, under an otherwise N-limited environment.
Ecologically, the molecular properties of tannins and their structural diversity enable these compounds to interfere with multiple decomposition pathways. The present study demonstrates that MTs exhibit enzyme inhibition capacities that are equal to or greater than those of CTs. Moreover, the susceptibility to and the mechanism of inhibition are functions of the enzyme class and the presence of different isoforms of enzymes within a class. From an ecological perspective, these responses reveal that a shift in the tannin chemistry in an ecosystem could result in temporary accumulation of plant litter and hindrance of nutrient cycling until enzyme acclimatization occurs. Similarly, because the tannin chemistry readily responds to the environment (Tharayil et al., 2011), the preferential degradation of litter in its parent ecosystem (i.e. the home-field advantage hypothesis; Ayres et al., 2009) may be explained in part by the soil isoenzyme response to a novel tannin (inhibitor) quality. Thus, our study provides an additional mechanism by which soil microorganisms could adapt to changing litter quality and inhibitors therein – through isoenzyme acclimation.
The authors thank two anonymous reviewers for their constructive comments which improved the quality of the manuscript. This research was supported by an NSF Graduate Research Fellowship to D.J.T. and USDA Grant (2009-35320-05042) and NSF Grant (DEB-1145993) to N.T. This is Technical Contribution No. 6074 of the Clemson University Experiment Station.