Sequestration of soil nitrogen as tannin–protein complexes may improve the competitive ability of sheep laurel (Kalmia angustifolia) relative to black spruce (Picea mariana)


Author for correspondence:
Robert L. Bradley
Tel: +1 819 8218000 (ext. 62080)
Fax: +1 819 8218049


  • • The role of litter tannins in controlling soil nitrogen (N) cycling may explain the competitive ability of Kalmia relative to black spruce (Picea mariana), although this has not been demonstrated experimentally.
  • • Here, the protein-precipitation capacities of purified tannins and leaf extracts from Kalmia and black spruce were compared. The resistance to degradation of tannin–protein precipitates from both species were compared by monitoring carbon (C) and N dynamics in humus amended with protein, purified tannins or protein–tannin precipitates. The purity of the precipitates was verified using solid-state 13C nuclear magnetic resonance (NMR) spectra. The ability of mycorrhizal fungi associated with both species to grow on media amended with tannin–protein complexes as the principal N source was also compared.
  • • The protein precipitation capacity of Kalmia tannins was superior to those of black spruce. Humus amended with protein increased both mineral and microbial N, whereas humus amended with tannin–protein precipitates increased dissolved organic N. Mycorrhizal fungi associated with Kalmia showed better growth than those associated with black spruce when N was provided as tannin–protein precipitates.
  • • These data suggest that Kalmia litter increases the amount of soil N sequestered as tannin–protein complexes, which may improve the competitive ability of Kalmia relative to black spruce by favouring N uptake by mycorrhizas associated with the former.


Most of the nitrogen (N) entering the soil system consists of high-molecular-weight organic substrates such as proteins. Because roots only absorb mineral N (inline image or inline image) or low-molecular-weight organic N molecules such as amino acids (Nasholm et al., 1998), proteins must first be decomposed and mineralized before plants can have access to soil N. The first stages of protein degradation in soil are catalyzed by exo-enzymes that are released into the soil solution by microorganisms. These enzymes are understood to be inefficient in the presence of condensed tannins (Joanisse et al., 2007), which are plant-produced polymers of three-ring flavanols that form stable cross-links with protein.

The role of plant-produced tannins in controlling soil N cycling has frequently been evoked and studied in boreal forest–ericaceous shrub systems (e.g. Bradley et al., 2000a,b). Lebel et al. (in press) showed that the presence of sheep laurel (Kalmia angustifolia L.), a common boreal ericaceous shrub, was related to low soil N-mineralization rates and stunted growth of regenerating black spruce (Picea mariana (Mill.) BSP) seedlings, and they suggested that this resulted from the binding of Kalmia litter tannins to soil proteins. Joanisse et al. (2007) found that Kalmia leaves contained five times more condensed tannins than black spruce needles, which further corroborates this premise that Kalmia leaf litter may interfere with soil N cycling. However, condensed tannins produced by different plants have different chain lengths and stereochemistries, and these structural differences could logically result in different functional properties (Kraus et al., 2003). With regard to Kalmia and black spruce, previous studies have shown definite structural differences in the condensed tannins that they each produce (Nierop et al., 2005; Joanisse et al., 2007), and this could affect their respective N-binding capacities.

The ability of tannins to bind protein should not be confused with the stability of these precipitates to resist degradation in soil. For example, Howard & Howard (1993) found that tannins produced by a range of tree and shrub species sequestered different amounts of protein, and the resulting precipitates differed substantially in their ability to resist degradation. If the former is true, then we should expect an increase in both suspended and dissolved organic N (SON and DON) in soil solution and a concomitant decrease in mineral N (Northup et al., 1995). On the other hand, if tannins and tannin–protein complexes readily decompose, then they effectively would act as energy-yielding substrates to soil microorganisms. In this case, what could be perceived as lower mineral N as a result of the sequestration of soil protein would actually be the result of greater microbial immobilization of N. It is also unknown whether different tannins actually differ in their ability to bind to soluble soil protein as opposed to solid-phase organic matter (Fierer et al., 2001). Thus, attempts to evaluate the stability of tannins and tannin–protein complexes in soil should include measurements of microbial respiration, soil mineral N, DON and microbial N content.

Many boreal forest plants growing in tannin-rich soil environments may present biological systems that have evolved to circumvent the need for exogenous N-mineralization pathways. For example, mycorrhizal symbionts associated with some boreal plants have been shown to produce enzymes that degrade polymeric organic matter and subsequently assimilate the resulting hydrolysates (Read & Perez-Moreno, 2003; Read et al., 2004). As the taxonomic diversity of mycorrhizal species is, however, very high – 5000–6000 boreal species according to Buscot et al. (2000) – we expect the functional diversity of mycorrhizal species colonizing different plants also to be high. Therefore, in order to understand the ability of Kalmia and black spruce to cope with N sequestration in tannin-rich soil environments, there is a need to evaluate how various species of their respective mycorrhizal symbionts can grow in the presence of tannins or tannin–protein precipitates.

Previous studies on protein–tannin binding in soil have mainly used complexes formed with commercially available tannic acid (e.g. Bending & Read, 1996; Wu et al., 2005), which is a hydrolysable tannin consisting of gallic acid esters linked to sugar moieties. On the other hand, Kalmia and black spruce do not produce hydrolysable tannins (Nierop et al., 2005). While the protein-precipitation capacity of tannic acid can either be superior (Kraus et al., 2003) or inferior (Mutabaruka et al., 2007) to that of condensed tannins, tannic acid is thought to degrade more easily than condensed tannins (Kraus et al., 2004). For this reason, we performed a study comparing various functional aspects of Kalmia and black spruce tannins, and included tannic acid treatments for comparative purposes. Our first objective was to compare the protein-precipitation capacities of purified tannins and leaf extracts from Kalmia and black spruce. As the capacity of tannins to precipitate protein is not tantamount to the stability (i.e. resistance to degradation) of the precipitates, our second objective was to compare carbon (C) and N dynamics in forest floor material amended with tannins and protein–tannin complexes from Kalmia and black spruce. In so doing, we verified the purity of tannin–protein precipitates formed from Kalmia and black spruce leaf extracts by comparing their solid-state 13C nuclear magnetic resonance (NMR) spectra with those of purified tannins and purified protein. Our third objective was to compare the ability of different mycorrhizal species, frequently associated with Kalmia or black spruce, to grow on media amended with purified tannins as well as on media containing Kalmia or black spruce tannin–protein complexes as the principal N source. Given the established notion that Kalmia tannins contribute significantly towards reducing the cycling of soil N, we hypothesized that these would have a greater protein-binding capacity and a greater stability in soil than black spruce tannins. We further hypothesized that mycorrhizal fungi associated with black spruce would grow less on media containing tannin–protein complexes as the principal N source than mycorrhizal fungi associated with Kalmia.

Materials and Methods

Expt 1: protein-precipitation capacity of tannins

Two methods were used to compare the protein-precipitation capacities of Kalmia and black spruce condensed tannins, and that of hydrolysable tannic acid. Given that the emphasis of our study was on the structural and functional differences between the tannin types, we used commercially available bovine serum albumin (BSA) as the test protein. Bovine serum albumin is not a perfect model for soil protein, but it is not clear what would be a good model because soil protein chemistry is complex. Bovine serum albumin was used here because it has been widely used in the past (e.g. Bending & Read, 1996; Mutabaruka et al., 2007), allowing comparability among studies. We first used the radial diffusion assay, as described by Hagerman (1987). This assay measured the formation of insoluble protein–tannin complexes in agar plates containing 10 ml of 0.1% BSA solution. Condensed tannins from Kalmia leaves and black spruce needles had previously been purified according to the method described by Preston (1999). The spectra from solution 13C NMR showed that both were predominantly procyanidin units (two OH groups on the B ring) with C2–C3 cis stereochemistry; however, the average chain length was 6.0 for black spruce and 2.3 for Kalmia (Nierop et al., 2005; Joanisse et al., 2007). Three solution concentrations (12.5, 25 and 50 mg ml−1) of Kalmia and black spruce tannins, and generic tannic acid (Sigma-Aldrich), were prepared in 50% acetone : water. For each of the nine treatments (i.e. three tannin types at three concentrations), four replicate agar plates were used. Four BSA-agar cores (3 mm in diameter) were removed from each plate and 20-µl aliquots of tannin solution were pipetted into the resulting wells. Four control agar plates were cored and amended in a similar manner with solution consisting of 50% acetone : water only. All plates were incubated at 30°C for 96 h, and the diameter of the opaque precipitation ring around each well was measured along two perpendicular axes. The depth of the agar around each well was determined using callipers in order to estimate the weight of protein that was precipitated around each well.

Precipitation capacities of the three tannin types, and those of Kalmia leaf and spruce needle extracts, were also measured in solution. Three grams of each tannin type was dissolved in 500 ml of H2O and then filter-sterilized through glass filter papers of 0.80, 0.45 and 0.22 µm pore-size diameter. Fresh green Kalmia leaves and spruce needles were freeze-dried, finely ground using a ball mill and 50-g samples were transferred to brown glass bottles to which were added 500 ml of acetone : water (70:30, v/v). The bottles were shaken for 1 h and the solutions were centrifuged at 1000 g. Supernatants were filtered (Whatman No. 5 filter paper, Whatman Inc., Florham Park, NJ, USA) whereas pellets were resuspended in 500 ml of 70% acetone : water and the extraction procedure was repeated a second time. Solutions from both extractions were combined and the acetone was removed from solution by roto-evaporation at 30°C. The remaining aqueous Kalmia and black spruce extracts were filter sterilized and stored in sterilized bottles. The pH values of the five solutions were as follows: Kalmia tannins, 3.99; black spruce tannins, 3.61; tannic acid, 4.00; Kalmia extract, 4.02; and black spruce extract, 3.77. A solution containing 2.5 g l−1 of BSA in sodium acetate buffer (10 µm, pH 4.5) was prepared, filter sterilized and kept in a sterile bottle at 4°C. Each tannin solution and each foliar extract were transferred in 2, 5, 10 and 15 ml to duplicate sterile centrifuge tubes, and the volume in each tube was adjusted to 20 ml using the BSA solution. This resulted in tannin : BSA (w/w) ratios of 0.27, 0.80, 2.4 and 7.2, and leaf : BSA ratios (w/w) of 7.4, 22.2, 66.7 and 200.0. The solution mixture in all 40 tubes was mixed using a vortex mixer and kept at 4°C for 24 h. The tubes were then centrifuged (1800 g) and the supernatants discarded. The pellets were rinsed with H2O, vortex mixed, centrifuged and the supernatants were discarded once again. The precipitates were freeze dried, weighed and analyzed colorimetrically for total N following acid digestion using a Technicon Autoanalyser (Pulse Instrumentation, Saskatoon, Canada). The amount of BSA precipitated in each solution was calculated using a mass conversion factor of 0.16 total-N : BSA.

Expt 2: stability of protein–tannin complexes in soil

Forest floor humus material was collected from a 10-yr-old black spruce cutover located near the Town of Senneterre, Québec, Canada (c. 48°N, 76°W). In this area the mean annual temperature is 0.5°C and the mean annual precipitation is 972 mm. Soils of the region are mainly Humo-Ferric Podzols, and the drainage at this particular site is classified as ‘mesic’ (Blouin & Berger, 2001). The site was characterized by an average 10-cm-deep organic F–H layer (Soil Classification Working Group, 1998). The groundcover consisted of a feathermoss (Pleurozium schreberi (Brid.) Mitt) mat, and the shrub layer was dominated by Kalmia angustifolia. Samples (c. 500 g) of the organic F–H layer material were collected every 5 m along two 50-m transects and bulked to create a single sample. The material was sieved through a 5-mm mesh, transported on ice to the Laboratoire d’écologie des sols – Université de Sherbrooke and stored at 4°C until the experiment began, 2 wk later. Five subsamples were finely ground using a Retsch model MM200 ball mill (Retsch GmbH & Co., Haan, Germany); these were acid digested and analyzed colorimetrically for total N and phosphorus (P) using the indophenol blue and the vanado-molybdo-phosphoric acid methods, respectively. Forest floor pH was measured electrometrically from aqueous suspensions (soil : water, 1:10).

Exactly 10 g (3.75 g dry weight (DW) equivalent) of fresh humus material was transferred into 108 plastic jars (500 ml) and left for 5 d to condition at room temperature (20°C). Each jar was assigned to one of nine treatments, which were replicated 12 times. These comprised eight amendments (BSA, Kalmia and black spruce tannins, Kalmia and black spruce tannin–BSA precipitates, Kalmia and black spruce extract–BSA precipitates, and an unamended control). Tannin–protein complexes were prepared, as described above, by mixing 10 ml of each tannin solution or leaf extract (described above) with 30 ml of BSA solution. This resulted in tannin : BSA and leaf : BSA ratios of 0.80 and 22.2, respectively, thereby precipitating c. 95% of the BSA from all solutions. In order to ensure a uniform distribution in the humus, substrates were applied with talc in 500-mg mixtures. We used 30-mg tannin–protein complexes, 20 mg of purified tannins and BSA, and 500 mg of talc for the control. The C and N concentration of each amendment was determined using a Vario Macro CN Analyzer (Elementar GmbH, Hanau, Germany), which had not been available for the work previously described. The CN analyzer is more precise than the colorimetric approach, but this was not considered a problem as we were not comparing N concentrations of the humus with those of the amendments. In order to monitor CO2 evolution, a glass vial (50 ml) was fixed with adhesive tape onto the interior face of each jar to which we added 20 ml of 0.5 m KOH solution. The jars were immediately sealed with airtight lids. Empty jars were also prepared to correct for ambient CO2 concentrations.

Three jars of each treatment were destructively sampled 15 min after amendments were applied (t = 0), as well as after 2, 7 and 28 d. Upon opening each jar, 20 ml of 1 m BaCl2 was added to vials containing KOH solution, and these were titrated with 0.1 m HCl using a 0.05% thymolphthalein-ethanol indicator. HCl volumes were converted to mg of CO2-C using the stoichiometric relationship: (1 meq NaOH used to trap CO2) = (6 mg CO2-C).

Soil mineral N concentration in each jar was determined by extracting 4.5 g (fresh weight (FW)) of soil subsamples with 40 ml of 0.5 m K2SO4 solution. After shaking for 1 h, extracts were filtered (Whatman No. 5) and analyzed colorimetrically for inline image-N (using nitroprusside–hypochlorite–salycilate) and inline image-N (using Cd reduction–sulphanilamide) concentrations (Mulvaney, 1996). inline image-N concentrations were consistently below the detection limit of our instrument (0.06 µg ml−1). The remaining extract was subsequently filtered through a 0.45-µm syringe filter and analysed for total dissolved N (TDN) by persulphate oxidation (Cabrera & Beare, 1993). Briefly, 10 ml of persulphate solution was added to duplicate 5-ml subsamples of the filtered extracts, the mixtures were autoclaved at 121°C for 45 min and then analysed for inline image-N. The DON concentration was calculated by subtracting inline image-N from TDN concentrations.

The microbial biomass N (Nmic) in each jar was determined using the chloroform fumigation extraction technique (Brookes et al., 1985). A 4.5-g (FW) subsample of soil from each jar was transferred to a 125-ml Mason jar and placed inside a desiccator. A glass dish containing 50 ml of distilled CHCl3 and boiling chips was placed on the bottom of the desiccator, and humus subsamples were fumigated under vacuum for 24 h. Fumigated samples were then extracted in 0.5 m K2SO4 and the TDN concentration was measured (as described above). To calculate Nmic, the TDN concentrations of the nonfumigated samples (described above) were subtracted from those of the fumigated samples.

The chemical nature of leaf extract precipitates was verified by comparing their solid-state 13C NMR spectra with cross-polarization and magic-angle spinning (CPMAS NMR), to those of BSA, purified tannins and tannin–BSA precipitates, using a Bruker MSL 300 spectrometer (Bruker Instruments, Karlsruhe, Germany) operating at 75.47 MHz (Lorenz et al., 2000; Preston et al., 2000). Dry samples were spun at 4.7 kHz in 7-mm diameter zirconium oxide rotors. Spectra were acquired with a 1-ms contact time, a 2-s recycle time and 4000–8000 scans, and were processed using 30-Hz line-broadening and baseline corrections in Win-NMR 6.0 (Bruker Instruments). Peak heights of pure tannins and BSA were adjusted to peak heights of leaf extract precipitates. In order to verify the purity of tannin–protein precipitates formed from Kalmia and black spruce leaf extracts, the corrected BSA spectrum was plotted against the spectra of both tannin types and both tannin–BSA precipitates, and the latter were plotted against the spectra of leaf extract precipitates.

Expt 3: mycorrhizal fungus growth in the presence of tannins and tannin–protein complexes

Four mycorrhizal fungi commonly found in black spruce forests, or in roots of ericaceous shrubs, were purchased from the University of Alberta Microfungus Collection (Edmonton, Canada). Two species, Hebeloma crustuliniforme (UAMH 5460) and Cenococcum geophilum (UAMH 6145), are ectomycorrhizal (ECM) endophytes common to black spruce (Byrd et al., 2000; Kranabetter et al., 2005; Robertson et al., 2006). Both species can grow on organic N sources (Abuzinadah & Read, 1986; Zhu et al., 1990), phenolic compounds and/or leaf extracts (Mallik & Zhu, 1995; Mallik & Zhu, 1998). The third species was Rhizoscyphus ericae (UAMH 8680, formally named Hymenocyphus ericae), an ericoid mycorrhiza (ERM) common among ericaceous shrubs (Massicotte et al., 2005) that can grow on organic N sources and phenolic compounds (Bending & Read, 1996, 1997; Read & Perez-Moreno, 2003; Read et al., 2004). The fourth species was the dark septa mycorrhizal (DSM) endophyte Phialocephala fortinii (UAMH 7137), which is commonly associated with roots of both ericaceous shrubs (Massicotte et al., 2005) and black spruce (Yamasaki et al., 1998; Massicotte et al., 2005; Olsrud et al., 2007), but may actually be pathogenic to conifers (Richard & Fortin, 1974; Wilcox & Wang, 1987). P. fortinii can also grow on a wide range of organic compounds (Caldwell et al., 2000), including Kalmia leaf extracts (Titus et al., 1995).

The experiment was designed to contrast the growth of the four mycorrhizal species on inorganic (inline image) vs organic N (BSA) substrates, with and without various tannins. The inline image treatments consisted of Modified Melin Norkrans (MMN) nutrient medium amended with Kalmia tannins, black spruce tannins, tannic acid or a nonamended control. The four BSA treatments consisted of MMN nutrient medium where inline image was substituted with BSA or one of the three tannin–BSA precipitates prepared as described in Expt 2. The MMN nutrient medium was prepared without malt extract, sterilized in an autoclave and solidified with agar. For the three tannin treatments, we dissolved 450 mg of each tannin type in 75 ml of H2O, after which the solutions were filter sterilized (0.22 µm) and mixed with the autoclaved MMN media at a proportion of 1:20 (v/v). For the BSA treatment, we dissolved 1.0 g of BSA in 250 ml of H2O, filter sterilized (0.22 µm) the solution using a low-protein-binding filter paper and then mixed the solution with the autoclaved MMN medium at a proportion of 1:20 (v/v). For the three tannin–BSA treatments, we added 500 mg of each precipitate to a litre of autoclaved MMN medium. Glucose was added to all eight solutions to give a final concentration of 5.0 g l−1. Given that boreal forest soils are typically acidic, the pH of each solution was adjusted to 4.8, the lowest pH at which agar can solidify (Marx, 1969). We poured 25 ml of each agar growth medium into 20, 9-cm-diameter Petri dishes. The total amount (µg) of N added to each Petri dish was as follows: control, 761; Kalmia tannins, 761; black spruce tannins, 761; tannic acid, 761; BSA, 795; Kalmia extract–BSA, 1000; black spruce extract–BSA, 988; and tannic acid–BSA, 1217. Five Petri dishes of each growth medium were inoculated with a single agar plug (4.0 mm in diameter) taken from colonies of each mycorrhizal species growing on MMN agar medium. The 160 Petri dishes were sealed with parafilm and then incubated for 60 d in the dark at 20°C. The Petri dishes were then set in a thin layer of hot (50–60°C) water in order to melt the agar, after which the mycelia were collected on preweighed glass filter papers and rinsed twice with distilled water. Filter papers and mycelia were freeze-dried and weighed.

Statistical analyses

In Expt 1, simple linear regressions were performed to test the relationships between tannin concentration and the amount of BSA that was precipitated in the radial diffusion assay, and analysis of covariance was used to compare the slopes of these regression lines. Two-way ANOVAs were used to test the effects of tannin type, tannin concentration (or ratios to BSA), and the interaction between tannin type × tannin concentration on the N content of precipitates and on the amount of BSA precipitated from solution. When significant interactions were found, one-way ANOVAs within each tannin type were performed. Ryan-Einot-Gabriel-Welsch F (REGW-F) post-hoc tests were used to detect significantly different means. Data in Expt 2 were analyzed in the same way as in Expt 1, with incubation time and humus amendments as the experimental factors. For data in Expt 3, one-way ANOVAs were used to test the effects of the eight growth media on the mycelial mass of each mycorrhizal species. We then performed one-way ANOVAs to compare the mycelial mass of each mycorrhizal species within each N source (i.e. inline image and BSA). One-way ANOVAs and single degree of freedom orthogonal contrasts were also performed to compare relative mycelial mass in the tannin–protein media relative to the mycelial mass in the inline image and BSA media between mycorrhizal species. Statistical analyses were performed using spss 11.01 (SPSS Inc., Chicago, IL, USA) software. Before each analysis, we verified that the data conformed to assumptions of normality and homogeneity of variance, and used loge transformations when necessary to meet these assumptions. The level of significance for all tests was set to P ≤ 0.05.


Expt 1: protein-precipitation capacity of tannins

Two-way ANOVA revealed significant (P < 0.001) effects of tannin type, tannin concentration and their interaction term, on the amount of BSA precipitated in the radial diffusion assay. At all three tannin concentrations, tannic acid precipitated significantly more BSA than Kalmia and black spruce tannins (Fig. 1). For all three tannin types, the amount of precipitate increased linearly over the range of concentrations that were used and analysis of covariance revealed a significantly (F2,31 = 78.6, P < 0.001) higher regression slope for tannic acid than for Kalmia and black spruce tannins (Fig. 1). Regression slopes for Kalmia and black spruce tannins did not differ significantly.

Figure 1.

Effect of tannin concentrations on the amount of bovine serum albumin (BSA) precipitated during radial diffusion assays. All regressions are significant at P < 0.0001. Each point represents the mean value of four Petri dishes; vertical lines = 1 SD. Points depicting Kalmia and spruce (Picea mariana) tannins have been offset along the x-axis for clarity.

Two-way ANOVAs revealed significant (P < 0.001) effects of tannin type, tannin : BSA ratios and their interaction term, on the N concentration of precipitates and on the amount of BSA precipitated (Fig. 2a,b). For each tannin type, the N concentration of precipitates decreased with an increasing ratio of tannin : BSA (Fig. 2a). One-way ANOVAs and post-hoc REGW-F tests revealed significantly (P = 0.014) higher N concentrations for Kalmia and black spruce tannins than for tannic acid at the tannin : BSA ratio of 0.8, and a significantly (P < 0.022) higher N concentration for Kalmia tannins than for black spruce tannins and tannic acid at tannin : BSA ratios of 2.4 and 7.2 (Fig. 2a). On the other hand, there was significantly (P = 0.001) more BSA precipitated with tannic acid than with Kalmia or black spruce tannins at the lowest (0.27) tannin : BSA ratio (Fig. 2b). At higher tannin : BSA ratios, most of the BSA had precipitated.

Figure 2.

Effects of increasing ratios of tannins to bovine serum albumin (BSA) (a, b), and leaf extract to BSA (c, d), on the percentage of total BSA precipitated from solution and on the nitrogen concentration of these precipitates. (a, b) Kalmia tannins, triangles; spruce tannins, circles; tannic acid, squares. (c, d) Kalmia extract, triangles; spruce extract, circles. Each point represents the mean of duplicate solutions; vertical lines = 1 standard deviation. Overlapping points have been offset along the x-axis for clarity.

For leaf extracts, two-way ANOVA revealed significant effects (P < 0.001) of leaf : BSA ratios on the N concentration of precipitates (Fig. 2c). These decreased with an increasing ratio of leaf : BSA. Two-way ANOVA revealed significant effects of leaf type (P = 0.018) and leaf : BSA ratio (P < 0.001) on the amount of BSA precipitated from solution (Fig. 2d). More specifically, Kalmia extracts precipitated more BSA from solution than black spruce extracts at the lowest (7.40) leaf : BSA ratio (Fig. 2d). At higher leaf : BSA ratios, the proportion of BSA that was precipitated varied between 91 and 100%.

Expt 2: stability of tannin–protein complexes in soil

The initial characteristics of the humus used for this experiment were as follows: total N, 13.73 mg g−1; total P, 0.56 mg g−1; organic matter, 89%, pH = 3.6. The C and N contents of each amendment are given in Table 1. NMR spectra of Kalmia extracts–BSA, BSA and Kalmia tannins are shown in Fig. 3(a) and those of black spruce extracts–BSA, BSA and black spruce tannins are shown in Fig. 3(b). Tannin and BSA spectra were consistent with previous spectra and chemical-shift data (Preston et al., 2000). All peaks of the plant extract–BSA spectra were found in the combined spectra of BSA and purified tannins, indicating that leaf extract precipitates are primarily composed of pure tannins and BSA. This was confirmed by adding the spectra of BSA to those of purified tannins, which resulted in spectra that were nearly identical to those of leaf extract precipitates (Fig. 3c,d).

Table 1.  Carbon and nitrogen contents of the 500-mg talc–substrate mixtures added to 3.75- (dry weight (DW) equivalent) humus samples in Expt 2
  1. bdl, below detection limit; Control, talc only; BSA, bovine serum albumin; KT, Kalmia tannin; KE, Kalmia leaf extracts; BST, black spruce tannin; BSE, black spruce needle extracts; TA, tannic acid.

N (mg)bdl3.01bdl2.672.42bdl3.242.372.92
C (mg)bdl10.0511.1212.2517.6012.2015.6816.7516.02
Figure 3.

13C cross-polarization and magic-angle spinning nuclear magnetic resonance (CPMAS NMR) spectra of (a, b) Kalmia and black spruce (Picea mariana) leaf extract–bovine serum albumin (BSA) precipitates (KE – BSA and BSE – BSA, respectively), of BSA, and of purified Kalmia and black spruce condensed tannins (KT and BST, respectively). Lower panels (c, d) compare the spectra of leaf extract–BSA precipitates with the summed spectra of BSA and purified tannins. ppm, parts per million.

Two-way ANOVA revealed significant interactions between incubation time and humus amendments in controlling cumulative CO2-C. Subsequent one-way ANOVAs revealed that the effect of amendments on cumulative CO2-C was only significant after 28 d of incubation. At this time-point, post-hoc REGW-F analysis revealed significantly lower values in the black spruce tannins than in the BSA treatment (Fig. 4a).

Figure 4.

Effects of bovine serum albumin (BSA) (black bars), condensed tannins (cross-hatch bars) and tannin–BSA precipitates (grey bars) on (a) cumulative respiration, (b) mineral N and (c) dissolved organic N after 28 d of incubation, as well as on microbial N pooled across dates. Substrate abbreviations: control, talc only; BSA, bovine serum albumin; KT and BST, Kalmia and black spruce (Picea mariana) tannins, respectively; KE and BSE, Kalmia and black spruce leaf extracts, respectively; TA, tannic acid. Bars represent means of three (a, b, c) or 12 (d) experimental units; vertical lines = 1 SD. Different lower-case letters within each frame indicate significantly different means according to Ryan-Einot-Gabriel-Welsch F (REGW-F) tests. DON, dissolved organic N.

Two-way ANOVA revealed significant interactions between incubation time and humus amendments in controlling inline image-N concentrations. Subsequent one-way ANOVAs revealed a significant effect of amendments after 2, 7 and 28 d of incubation. On day 2, post-hoc REGW-F analyses revealed significantly (P < 0.01) higher inline image-N concentrations in the control and BSA than in the other treatments (data not shown). On day 7 (data not shown) and day 28 (Fig. 4b), BSA-amended humus had significantly (P < 0.01) higher extractable inline image-N concentrations than all other treatments. Assuming that this net increase in inline image-N over the control treatment was attributable to the mineralization of BSA, and that no losses of N through volatilization occurred, the results suggest that 32% of BSA–N was mineralized after 28 d of incubation.

Two-way ANOVA revealed significant interactions between incubation time and humus amendments in controlling DON. Subsequent one-way ANOVAs revealed a significant effect of amendments after 0, 2 and 28 d of incubation. On day 0 and day 2, post-hoc REGW-F analyses revealed significantly (P = 0.04 for both) lower DON concentrations in the control compared with other treatments (data not shown). On day 28, DON concentrations were significantly (P = 0.04) lower in the BSA-amended humus than in the other treatments (Fig. 4c).

Two-way ANOVA revealed a significant effect of humus amendments in controlling Nmic. Post-hoc REGW-F analysis of the data pooled across dates revealed significantly (P = 0.006) higher Nmic concentrations in the BSA-amended humus than in the other amended treatments (Fig. 4d).

Expt 3: mycorrhizal growth in the presence of tannin–protein complexes

One-way ANOVAs revealed significant effects of growth media on the mycelial mass of all mycorrhizal species except for P. fortinii (Fig. 5). For the other three species, the mycelial mass was lower following culture in growth media containing tannin–protein precipitates than in media containing inline image as the N source (Fig. 5a–c). Within the four media containing inline image as the N source, there was significantly higher mycelial mass in the Kalmia tannins than in the control treatments for both H. crustuliniforme (P = 0.04) and C. geophilum (P = 0.032) (Fig. 5a,b), and higher mycelial mass in the tannic acid than in the control treatment (P = 0.028) for R. ericae (Fig. 5c). Within the four treatments with BSA as the N source, the mycelial mass was significantly (P < 0.002) higher with BSA alone than with each tannin–BSA precipitate for H. crustuliniforme (Fig. 5a), C. geophilum (Fig. 5b) and R. ericae (Fig. 5c). Compared with the other three mycorrhizal species, the mycelial mass of C. geophilum (Fig. 5b) growing on media containing BSA alone was low relative to the four media containing inline image as the N source. Single degree of freedom orthogonal contrasts revealed that, relative to the control media, mycelial masses of the two ECM species (H. crustuliniforme and C. geophilum) growing on the three tannin–BSA media were significantly lower (P < 0.006) than those of R. ericae and P. fortinii.

Figure 5.

Mycelial mass of (a) Hebeloma crustuliniforme, (b) Cenococcum geophilum, (c) Rhizoscyphus ericae and (d) Phialocephala fortinii grown for 60 d on various substrates: control, Modified Melin Norkrans (MMN) nutrient medium; KT, BST and TA, (Kalmia tannin, black spruce tannin and tannic acid, respectively) + inline image; BSA, bovine serum albumin; KE-BSA, BSE-BSA and TA-BSA, (Kalmia leaf extracts, black spruce needle extracts and tannic acid, respectively, + BSA). Different lower case letters within each frame indicate significant differences (established using the Ryan-Einot-Gabriel-Welsch F (REGW-F) test) between control and purified tannin media, all containing mineral N, whereas different upper case letters indicate significant differences between protein–tannin complexes and BSA media, all containing organic N. Vertical bars = 1 SD.


The results of the radial diffusion assay (Fig. 1) suggest that Kalmia tannins do not precipitate more protein than those produced by black spruce. While this assay is rapid and inexpensive, and has thus been used extensively in the past, caution is warranted in interpreting the results. First, protein-binding capacity is inferred from the volume of the opaque precipitation ring around each well in the agar plates. Therefore, the data follow a binary scale (opaque vs translucent) that assumes, perhaps falsely, the uniform precipitation of protein within the opaque rings for all tannin types. Second, the volume of the opaque ring can be related to the ability of tannins to diffuse through the agar medium as it is a measure of the protein-binding capacity. In this respect, Joanisse et al. (2007) found that Kalmia and black spruce tannins consisted of polymers with average chain lengths of 2.3 and 6.0 three-ring flavanol units, respectively, which would lead us to expect slower diffusion of black spruce tannins as a result of the higher molecular weight. However, the low apparent chain length of Kalmia root and shoot tannins (Nierop et al., 2005; Joanisse et al., 2007) may conceal other structural features, such as lateral branching or incorporation of p-coumaric acid (Nierop et al., 2005), which may also impede its diffusion. Based on the volume of the opaque rings, our results suggest that tannic acid precipitates significantly more protein than condensed litter tannins, which is consistent with other studies (Giner-chavez et al., 1997; Kraus et al., 2003). Again, caution is warranted in interpreting this result, as tannic acid may diffuse more easily through agar medium than condensed tannins. Thus, the caveats associated with the radial diffusion assay warranted the second part of this first experiment, which measured the actual weight and chemical characteristics of tannin–protein precipitates formed in solution, as they would naturally be formed in the soil.

All three tannin types were efficient in precipitating BSA from solution (Fig. 2a,b). Although BSA is not a soil protein, these results could nevertheless be ecologically relevant because the tannin : BSA ratios that we used corresponded to tannin : N ratios that were either smaller or within the same order of magnitude (i.e. 1.7–45.3) as the tannin : N ratios found in Kalmia leaves (32.5) and black spruce needles (23.6) (Joanisse et al., 2007). In fact, most of the BSA had precipitated from solution at a tannin : BSA ratio of 0.8, corresponding to a tannin : N ratio of 5.0. Given that the N concentration of precipitates decreased asymptotically with increasing tannin concentration, we posit that there is a ‘luxury amount’ of tannins that can be bound within precipitates as the free protein pool in solution becomes limited. We believe that this is plausible because condensed tannins are smaller than the protein they bind to, and hence many tannin molecules could bind to a single protein molecule. Given that these bonds are electrostatic, we can imagine a minimum number of tannin molecules necessary to precipitate a protein molecule and make it ‘unavailable’ to further degradation without necessarily saturating the ligand force field. Whether the extra tannins that are bound at elevated tannin : BSA ratios confer greater resistance to the precipitates remains to be determined. At the lowest tannin : BSA ratio (i.e. 0.27), we found that tannic acid precipitated more BSA in solution than Kalmia or spruce tannins. This confirms that the larger opaque rings measured around wells inoculated with tannic acid in the radial diffusion assay were not solely the result of hydrolyzable tannins diffusing more rapidly through agar medium than condensed tannins. The comparison with hydrolysable tannins is, however, of lesser importance given that Kalmia and black spruce both lack hydrolyzable tannins in their tissues (Nierop et al., 2005). The salient finding of this experiment is that the amount of protein precipitated per unit mass of tannins is greater for Kalmia than for black spruce, an observation that was not apparent with the radial diffusion assay. Thus, if the formation of tannin–protein precipitates in soil is a mechanism by which Kalmia can gain a competitive advantage over black spruce, then this strategy is not only ascribed to the higher tannin concentrations found in Kalmia litter (Joanisse et al., 2007), but also to the more effective protein-precipitating capacity of Kalmia tannins.

The results of 13C CPMAS NMR analyses (Fig. 3) leave little doubt as to the purity of tannin : protein precipitates formed from leaf extracts and BSA. It is not surprising therefore that the N concentrations of precipitates formed from leaf extracts, and the corresponding amounts of BSA precipitated from solution, responded to variations in leaf : BSA ratios (Fig. 2c,d) as they did to variations in tannin : BSA ratios (Fig. 2a,b). In making these comparisons, two details may appear as inconsistencies: (1) precipitates formed with purified tannins of each species had different N concentrations (Fig. 2a), but those formed with extracts of each species had similar N concentrations (Fig. 2c); and (2) purified tannins of Kalmia and black spruce precipitated similar amounts of BSA (Fig. 2b), but Kalmia leaf extracts precipitated more BSA at the lower leaf : BSA ratio than spruce needle extracts (Fig. 2d). This can be explained by the fact that tannin concentrations in Kalmia leaves are up to five times higher than in black spruce needles (Joanisse et al., 2007) and thus any given leaf : BSA ratio would yield a higher tannin : BSA ratio for Kalmia extracts, which could reduce the N concentration of precipitates (as per Fig. 2a) to levels similar to those formed from spruce needle extracts, and result in a higher percentage of BSA precipitated. We did not observe significantly more BSA precipitated with Kalmia extracts at higher leaf : BSA ratios because extracts from both species precipitated most of the BSA at these concentrations. Hence, the data shown in Fig. 2(a,b) are not altogether inconsistent with those shown in Fig. 2(c,d).

Taken collectively, the respirometry data provide evidence that neither condensed Kalmia and black spruce tannins, nor precipitates formed from purified tannins and leaf extracts, are energy-yielding substrates to soil microorganisms. Thus, the substantially lower mineral N concentrations found in tannin-amended and precipitate-amended units, compared with the BSA-amended treatment (Fig. 4b), is not the result of higher microbial immobilization in the former, but of higher N mineralization in the latter. This is further corroborated by observations that microbial N content did not increase in units amended with tannins or precipitates as they did in those amended with BSA (Fig. 4d).

As we expected, the addition of tannins and precipitates did translate into significant increases in DON early during the incubation period (data not shown), and DON concentrations in these treatments were still generally higher than the control treatment after 28 d (Fig. 4c). The DON concentrations after 0 and 2 d of incubation were also higher in BSA-amended humus as a result of nondegraded BSA being extracted as DON. The fact that DON in BSA-amended humus was lower than the control treatment after 28 d of incubation implies that the degradation and uptake of BSA resulted in a concomitant degradation of native DON. This phenomenon, referred to as co-metabolic mineralization (Horvath, 1972; Criddle, 1993), is one in which soil microbes metabolize certain substrates more efficiently in the presence of certain other substrates. Here, the BSA was a readily decomposable substrate that provided energy to soil microbes, and this may have stimulated the synthesis of catabolic enzymes in the same way as root-derived carbon in the rhizosphere increases the degradation of more recalcitrant native soil organic matter (Bradley & Fyles, 1995; Allison & Vitousek, 2005).

We followed the example of previous studies (e.g. Emmerton et al., 2001) and compared the growth of mycorrhizal fungi in pure culture. This approach may not reflect, however, their physiological capabilities when grown in symbiosis with host plants. For example, Yang et al. (2006) found that extracellular protease production by ericoid mycorrhizas were partially controlled by the host. To address this issue would require, however, a very challenging experiment. Notwithstanding this limitation, our third experiment is the first of its kind to have tested the growth of mycorrhizal species using, as the sole N source, precipitates formed from condensed tannins extracted from specific plants, rather than precipitates formed from commercially available hydrolysable tannins (e.g. Bending & Read, 1996; Wu et al., 2005). The addition of pure tannins and inline image as the sole N source did not impede the growth of any of the mycorrhizal species tested, and in some cases growth was actually improved by the presence of tannins. The fact that C. geophilum could not metabolize BSA suggests that some ECM species associated with black spruce do not produce protease enzymes, although the extent of this phenomenon and whether it also occurs with some ericoid mycorrhizal species remains unknown. The most meaningful observation in this experiment was that both ECM species associated with black spruce grew very inefficiently when N was provided in the form of tannin : protein precipitates, whereas both mycorrhizal species associated with Kalmia showed better growth on these three substrates. Plant growth depends on the availability of soil mineral N (inline image and/or inline image) or soil amino acids (Nasholm et al., 1998); however, these compounds are generally found in low concentrations in boreal forest floors (Hannam & Prescott, 2003; Grenon et al., 2004). Soil proteins, either from plant or microbial origin, constitute an important resource that is readily degraded by microbial proteolytic exo-enzymes to replenish these plant-available N pools. That Kalmia produces far more foliar tannins capable of precipitating higher amounts of protein on a weight to weight basis than black spruce tannins, suggests that the presence of Kalmia in boreal black spruce forests will decrease soil-available N. However, the establishment of a tannin-rich environment coupled with the development of ericoid mycorrhizas that ostensibly degrade tannin–protein precipitates, appear to be traits that have co-evolved to procure a competitive advantage to Kalmia during forest succession. Future research should attempt to confirm this by investigating N uptake by mycorrhizal and nonmycorrhizal roots of both plant species growing on various tannin–protein complexes.


We are grateful to D. Duschesne for technical assistance. Financial support was provided by a National Science and Engineering Research Council of Canada (NSERC) Strategic Project Grant, as well as a NSERC graduate student grant awarded to the senior author.