Oecologia

, Volume 155, Issue 3, pp 583–592

Differential effects of sugar maple, red oak, and hemlock tannins on carbon and nitrogen cycling in temperate forest soils

Authors

    • Department of BiologyBoston University
  • Adrien C. Finzi
    • Department of BiologyBoston University
Ecosystem Ecology - Original Paper

DOI: 10.1007/s00442-007-0940-7

Cite this article as:
Talbot, J.M. & Finzi, A.C. Oecologia (2008) 155: 583. doi:10.1007/s00442-007-0940-7
  • 350 Views

Abstract

Tannins are abundant secondary chemicals in leaf litter that are hypothesized to slow the rate of soil-N cycling by binding protein into recalcitrant polyphenol–protein complexes (PPCs). We studied the effects of tannins purified from sugar maple, red oak, and eastern hemlock leaf litter on microbial activity and N cycling in soils from northern hardwood–conifer forests of the northeastern US. To create ecologically relevant conditions, we applied tannins to soil at a concentration (up to 2 mg g−1 soil) typical of mineral soil horizons. Sugar maple tannins increased microbial respiration significantly more than red oak or hemlock tannins. The addition of sugar maple tannins also decreased gross N mineralization by 130% and, depending upon the rate of application, decreased net rates of N mineralization by 50–290%. At low concentrations, the decrease in mineralization appeared to be driven by greater microbial-N immobilization, while at higher concentrations the decrease in mineralization was consistent with the formation of recalcitrant PPCs. Low concentrations of red oak and hemlock tannins stimulated microbial respiration only slightly, and did not significantly affect fluxes of inorganic N in the soil. When applied to soils containing elevated levels of protein, red oak and hemlock tannins decreased N mineralization without affecting rates of microbial respiration, suggesting that PPC formation decreased substrate availability for microbial immobilization. Our results indicate that tannins from all three species form recalcitrant PPCs, but that the degree of PPC formation and its attendant effect on soil-N cycling depends on tannin concentration and the pool size of available protein in the soil.

Keywords

TanninNitrogen cycleCarbon cyclePolyphenol–protein complex

Introduction

Interspecific differences in litter chemistry play a major role in regulating soil-N cycling (Aber et al. 1990; Hobbie 1992; Swift et al. 1979). Many studies have linked variations in leaf litter C-to-N ratios and lignin-to-N ratios to the rate of leaf litter decomposition (Hobbie 1992; Melillo et al. 1982; Wedin and Tilman 1990), and soil C-to-N ratios with the rate of mineralization and nitrification (Finzi et al. 2001; Ollinger et al. 2002). Indeed, differences in litter chemistry among species are often sufficiently large that there are discernable effects of tree species on microbial activity and rates of N cycling in soils in mixed-species stands (Finzi et al. 1998).

More recently, the effects of litter chemistry on microbial activity and N cycling have been expanded to include the role of polyphenolic compounds (Bowman et al. 2004; Constantinides and Fownes 1994; Kalburtji et al. 1999; Schweitzer et al. 2004; Steltzer and Bowman 1998, 2005). Polyphenols are a large group of secondary metabolites that vary in molecular structure and chemical properties and they are thought to regulate microbial activity in soils (Hattenschwiler and Vitousek 2000). Tannins are a particularly important class of polyphenols that are hypothesized to influence the soil-N cycle through their ability to bind protein (Kraus et al. 2003a). When tannins bind protein they form polyphenol–protein complexes (PPCs) that are resistant to decomposition (Basaraba and Starkey 1966; Benoit et al. 1968). If such binding occurs in soils, then the formation of PPCs could slow rates of proteolysis and N mineralization by decreasing the availability of organic N substrates to the microbial community (Northup et al. 1995; Yu et al. 2002).

Experimental support for PPC formation in soils amended with tannins has been mixed. In some cases, the addition of purified tannins to soil has resulted in decreased microbial activity and slower rates of inorganic-N production, implying that tannins bound soil protein into recalcitrant PPCs (Bradley et al. 2000; Fierer et al. 2001; Schimel et al. 1996). Moreover, some studies have found a positive correlation between soil polyphenol concentrations and the accumulation of organic N in soils, which is consistent with PPC formation (Berthrong and Finzi 2006; Northup et al. 1995). By contrast, other studies have found that the addition of purified tannins or litter rich in polyphenols stimulates microbial activity and microbial-N immobilization, implying that some polyphenols are labile substrates for microbial metabolism (Fierer et al. 2001; Kraus et al. 2004).

Given the abundance and chemical diversity of tannins in leaf litter, tannins may play an important role in dictating plant species effects on microbial activity and soil-N cycling. This is particularly true in eastern North American forests where foliar litterfall is a major input of material to the detrital pool (Raich and Nadelhofer 1989; Smith et al. 2002). Foliage can exceed 40% tannin content by weight (Baldwin and Schultz 1984), and tree species show large interspecific differences in their effects on N transformations that are only partly explained by the C-to-N ratio or lignin-to-N ratio of litter and soils (Finzi et al. 1998; Lovett et al. 2004). Thus, to determine how tannins from different species affect microbial activity and soil-C and N cycling in mixed-deciduous forests of the northeastern US, we conducted two tannin-addition experiments using purified tannins from three dominant temperate forest tree species; sugar maple (Acer saccharum), red oak (Quercus rubra) and eastern hemlock (Tsuga canadensis). In the first experiment, tannins were applied directly to soils in a series of soil incubation experiments. Given that sugar maple stands have higher rates of N mineralization and nitrification than red oak stands or hemlock stands (Finzi et al. 1998; Lovett et al. 2004), we hypothesized that the addition of tannins purified from sugar maple litter would stimulate microbial activity and rates of soil-N mineralization by serving as labile C substrates for microbial metabolism, whereas the application of tannins derived from red oak and hemlock litter would slow microbial activity and decrease rates of soil-N mineralization as a result of PPC formation. In the second tannin-addition experiment, we determined the potential protein-precipitating capacity of tannins from these tree species by adding a large quantity of labile protein to soils. We hypothesized that soils treated with protein would have significantly higher rates of microbial respiration and microbial-N immobilization than control soils or soils to which tannin alone was added. In soils treated with tannins plus protein, we reasoned that the formation of PPCs would slow microbial activity relative to the samples receiving only protein. This would occur since protein in no longer limiting, so we would expect tannins from all species to form PPCs due to sufficient substrate for reaction. Thus in the tannin plus protein samples, we hypothesized a significant decrease in microbial respiration and N immobilization relative to soils treated with protein alone.

Materials and methods

Study site descriptions

This research was conducted at two sites in northwestern Connecticut, USA, that differed in tree species composition and in soil parent material. Previous research showed that these sites also differed in microbial activity and soil-N cycling (Berthrong and Finzi 2006; Finzi and Berthrong 2005). The “Schist” site was located at the Great Mountain Forest in Norfolk, Connecticut (42°N, 73°15′W). Soil at the Schist site is derived from glacial till overlying mica schist/gneiss bedrock. The soils are inceptisols and classified as Typic Dystrochrepts (Hill et al. 1980). Soils are low in available base cations and have a pH of 3.75 and a C-to-N ratio of 18:1. The dominant tree species at the Schist site are Quercus rubra (32% of basal area), Fagus grandifolia (21% of basal area), Acer rubrum (15%), and Tsuga canadensis (13%). The “Dolomite” site was located in the Housatonic State Forest in North Canaan, Connecticut. Soil at the Dolomite site is derived from glacial till overlying dolomite [CaMg(CO3)2]. The soils are mesic inceptisols and classified as Aquic Eutrochrepts (Hill et al. 1980). Soils at the Dolomite site are finer textured than soils at the Schist site, have a pH of 5.70, and have a C-to-N ratio of 11:1. The Dolomite site is dominated by Fraxinus americana (42% of basal area), Acer saccharum (27%), Carya ovata (9%), and Populus grandidentata (9%). From 1930 to present, the mean annual temperature in the forests was 7°C and the mean annual precipitation was 1,330 mm (Russ, personal communication). Additional site details can be found in Finzi and Berthrong (2005).

Soil and litterfall collection

Litterfall was collected in four large litter traps (4 × 2 m) made from 1-mm-mesh screening. Sugar maple litter was collected from the Dolomite site and red oak and eastern hemlock litter was collected from the Schist site. Litterfall was collected in November 2004, transported to the laboratory, sorted by species, and air-dried.

Soils were collected from four replicate 10 × 10-m plots at each site. Ten replicate 5-cm diameter cores were collected from the top 15 cm of mineral soil in each plot in June and August 2005 using a soil bulk density sampler (80 cores total: 2 sites × 4 replicate plots per site × 10 replicates). Soils were stored on ice, transported to the laboratory, and processed within 72 h of collection. Soils were sieved through a 2-mm-mesh sieve and the ten soil samples from within a replicate plot were placed into a single polyethylene bag to give a composite soil sample. Soils were homogenized by hand and stored at 4°C for an additional 72 h before the beginning of the experiment.

Tannin purification and characterization

Litter was ground in a Wiley mill (0.5 mm). Tannins were extracted from the ground litter and purified following the procedure of Schimel et al. (1996). The purified tannin fraction consisted of a mixture of structurally and chemically distinct tannins specific to each tree species. These purified fractions were resuspended in water for application to soils.

Total C and N of litter and tannin fractions from each species were determined by dry combustion on an Elantech NC2500 elemental analyzer (CE Elantech, Lakewood, N.J.). Tannin fractions were also analyzed for the two main types of tannins synthesized by plants: condensed and hydrolysable tannins. Tannin characterization was performed using standard colorimetric methods of analysis. In brief, condensed tannins were measured by the acid butanol assay (Porter et al. 1986) and hydrolysable tannins were measured by the potassium iodate method (Schultz and Baldwin 1982). Results are the means of triplicate measurements of single fractions from each species.

Protein precipitation analysis

We measured the capacity of tannins from each species to precipitate protein using the radial diffusion assay (Hagerman 1987). Agar treated with bovine serum albumin (BSA) was dispensed into standard 8.5-cm-diameter Petri dishes. Three 4-mm wells spaced 3.5 cm apart were punched in each Petri dish and 1 mg condensed tannins was pipetted into wells in 10-μl aliquots. The assay was performed with three replicates of purified tannin fractions from each species. Petri dishes were incubated at 30°C for 120 h, after which time the diameters of the tannin–protein precipitation rings were measured in two directions and the depth of agar in each dish was measured with callipers to the nearest 0.1 mm. The volume of agar gel for each ring was calculated and used to determine the amount of protein precipitated by tannin from each species in units of milligram protein per milligram tannin.

Concentration gradient study

To determine how tannin concentration affects microbial activity and N mineralization, we added tannins in six different concentrations to soils collected in June 2005. Since the decomposition of tannins in leaf litter is dependent on interactions with other plant compounds and soil organisms, tannin addition experiments likely do not reflect realistic effects of these compounds in field settings. Although tannin additions to soils in this experiment may lead to responses that are distinct from those observed in field settings, our approach allows us to isolate the potential effects of plant tannins on soil C and N cycling in the absence of other confounding factors. The six tannin concentrations we added to soils were 0, 0.03, 0.06, 0.13, 0.68, and 2 mg condensed tannin g−1 soil. The quantity of tannin added to the soil in this study was lower than that applied in similar, previous studies, which range from 2 to 50 mg tannin g−1 soil (Bradley et al. 2000; Fierer et al. 2001; Kraus et al. 2004; Schimel et al. 1996, 1998). However, the relatively low concentration of tannins used in this study is more typical of mineral soils (up to 3.5 mg g−1 soil; Table 1). Soils from each replicate field plot at the Dolomite and Schist sites were treated with tannins from sugar maple, red oak, and eastern hemlock litter, resulting in a total of 144 treated samples (2 sites × 4 plots × 3 tree species × 6 concentrations). Twenty grams of fresh, homogenized, field moist soil (35% water) was weighed into 125-ml Nalgene bottles and pre-equilibrated in the dark at room temperature for 7 days prior to experimental treatment to allow microbial respiration to stabilize. One subsample of soil from each replicate plot was dried at 100°C for 48 h to determine percent soil moisture.
Table 1

Concentrations of extractable tannins in mineral soils from various ecosystems. Adapted from Kraus et al. (2003). SEs are in parentheses (n = 4)

Tannin concentration (mg g−1 soil)

Soil type

References

0.005–0.093

0.006–0.029

0.009–0.040

Mature oak–blackjack soil (0–60 cm)

Oak–pine soil (0–60 cm)

Tall grass prairie soil (0–60 cm)

Rice and Pancholy (1973)

0.5–2.4

0.6–3.5

1.5–2.3

Western red cedar (Thuja plicata) humus (0–20 cm)

Western hemlock (Tsuga heterophylla) humus (0–20 cm)

Sitka spruce (Picea sitchensis) humus (0–20 cm)

Preston (1999)

0.385

Black spruce (Picea mariana) humus

Bradley et al. (2000)

0.048 (0.003)a

0.133 (0.002)a

Mesic inceptisols at Dolomite site (0–15 cm)

Inceptisols at Schist site (0–15 cm)

Berthrong and Finzi (2006)

a Values are concentration of total phenolics extractable in aqueous methanol and measured by the Prussian blue method (Graham 1992)

Tannins were applied to each soil sample in a 2-ml aliquot, after which samples were mixed thoroughly to ensure even distribution of tannin solution. Two milliliters of nanopure water was added to control samples. Samples in the 125-ml Nalgene bottles were placed in 460-ml Mason jars fitted with rubber septa, sealed, and incubated in the laboratory. The treated soil samples were incubated in the laboratory for 14 days. The experiment was run for 14 days based on stabilization of daily microbial respiration rates in tannin-amended mineral soils after 2 weeks (data not shown). Five milliliters of nanopure water was added to the bottom of each Mason jar to maintain soil moisture over the 14-day incubation.

Microbial activity was monitored as CO2 flux from soils, which was measured every 2 days during the incubation period by sampling the headspace of each Mason jar with a 10-ml syringe and measuring CO2 concentration on an infrared gas analyzer (EGM-4; PP Systems, Amesbury, Mass.). The jars were aerated for 30 min every day to maintain headspace CO2 concentrations below 2%. Cumulative CO2 respired over the course of incubation was calculated as the sum of the daily CO2 fluxes in mg CO2 g−1 soil day−1.

Net N mineralization was measured in soils over the 14-day incubation. At the beginning and end of incubation, pools of NH4+ and NO3 were extracted from soils with a 2 M KCl solution. Concentrations of NH4+-N were assayed by the phenolate method on an autoanalyzer (Lachat Quickchem 8000; Zellweger Analytics, Milwaukee, Wis.) and concentrations of NO3-N were measured by the cadmium reduction method (Lachat Quickchem 8000). Net N mineralization was calculated as the difference in total inorganic N concentrations from the incubated and initial soil samples.

Tannin–protein interactions study

To determine whether tannins from sugar maple, red oak, and hemlock litter could form PPCs in soils that were not limited by protein substrate (see Berthrong and Finzi 2006 for substrate-limitation studies), we added protein and tannin to soils in factorial combination and measured microbial respiration and inorganic N fluxes. In August 2005, thirty-gram samples of fresh, homogenized soil were weighed into 125-ml Nalgene bottles, the moisture level was increased to that of soils collected in June, and the samples pre-equilibrated for 7 days prior to treatment. Based on the results of the concentration gradient study, we added tannins at a concentration of 2 mg condensed tannin g−1 soil. Tannins were applied to each soil sample in a 3-ml aliquot of a 20 mg tannin ml−1 solution. The sugar maple, red oak, and hemlock tannin treatments contained 1.01, 0.70, and 0.89 mg C g−1 soil, respectively, which increased soil C concentrations by 1.8–2.0% in Dolomite soils and 1.2–1.5% in Schist soils. BSA (A6003; Sigma-Aldrich) was used as a protein treatment, which was added to soils as a water solution in 3-ml aliquots at a concentration of 2 mg protein g−1 soil. The protein treatment contained 0.93 mg C g−1 soil and 0.29 mg N g−1 soil, increasing the total C and N concentration of the 30-mg soil sample by 2 and 30%, respectively at Dolomite, and 2 and 41% at the Schist site. In samples treated with tannin plus protein, the tannin and the protein solutions were added to soils separately at a concentration of 2 mg g−1 soil in 1.5-ml aliquots. Three milliliters of nanopure water was added to control soils. Samples were mixed and incubated in Mason jars in the laboratory for 14 days following the procedure in the concentration gradient study.

We monitored microbial activity and changes in soil N pools over the 14-day incubation period. Microbial activity was monitored by measuring the daily CO2 flux from soils as described above. At the end of incubation, each soil sample was partitioned into three 10-g subsamples for extraction of inorganic N. NH4+-N and NO3-N were extracted from one subsample in 2 M KCl and analyzed as described above. Net mineralization was calculated as the difference in the concentration of inorganic N between the incubated and initial samples.

At the end of the 14-day incubation period, gross NH4+ production and consumption were measured by 15N pool (Hart et al. 1994). One soil subsample was used for t = 0 h 15NH4+ labeling and one set of samples for t = 24 h labeling with 15NH4+. A 0.5-ml aliquot of a 0.2 mM 15NH4+-N solution was added to each sample. One set of labeled soils was extracted immediately with 2 M KCl, and a second set was incubated in the laboratory and extracted after 24 h. NH4+-N was diffused from 10 ml of each extract onto an acidified cellulose disc using the procedure of Brooks et al. (1989). The atom%-15N excess of NH4+-N was analyzed on a mass spectrometer (Europa Integra, Cheshire, UK) at the University of California, Davis, California.

Data analysis

The four 10 × 10-m plots at each site were considered statistical replicates. For the concentration gradient study, microbial respiration and N cycling were analyzed by three-way ANOVA with site (two levels), species (three levels), and concentration (six levels) as main effects. For simplicity and clarity of data presentation we present the cumulative CO2 flux data and the N mineralization data separately for the Dolomite and Schist sites on a species-by-species basis.

The tannin–protein study was analyzed by four-way ANOVA with site (two levels), species (three levels), tannin (two levels: control, + tannin) and protein (two levels: control, + protein) as main effects. The effect of species, tannin and protein were all significant alone, and in combination with one another. However, the effect of site was only interactive with treatment for the CO2 flux data. As with the concentration gradient study, we present the CO2 flux data separately for the two sites on a species-by-species basis. Because there were no interactions with soil type, the N transformation data were averaged across sites and then presented on a species-by-species basis.

All data were assessed for normality and homogeneity of variance. In cases where the data did not conform to model assumptions, the data were log transformed prior to analysis. Post-hoc comparisons among means were analyzed by the LSD test.

Results

Tannin characterization and protein precipitation capacity

Tree species differed in both the quantity and chemistry of tannins in litterfall. Sugar maple litter contained higher concentrations of both condensed and hydrolysable tannins than red oak or hemlock litter (Table 2). Red oak litter contained higher concentrations of hydrolysable tannins and condensed tannins than hemlock litter (Table 2). The ratio of condensed:hydrolysable tannins increased in the order sugar maple < red oak < hemlock (Table 2). Tannins purified from sugar maple, red oak, and hemlock litter contained 50, 56, and 59% C, respectively (Table 2). Red oak and hemlock tannins contained 0.60 and 0.63% N, respectively, while N concentrations of sugar maple tannins were undetectable (<0.01%).
Table 2

Quantity and chemical characterization of tannins in leaf litter from sugar maple, red oak, and hemlock (SEs in parentheses). Values are the average of three replicates of a single extract (n = 3). DM Dry matter, ND not determined

Species

Quantity

Chemistry

Common name

Latin name

Extractable hydrolysable tanninsa (mg g−1 DM)

Extractable condensed tanninsb (mg g−1 DM)

Ratio of condensed: hydrolysable tannins

C (%)

N (%)

Protein precipitated (mg mg−1 tannin)

Sugar maple

Acer saccharum

73.46 (1.59)

24.00 (1.43)

3.07 (0.15)

50.4

ND

3.39 (0.79)

Red oak

Quercus rubra

53.94 (1.32)

5.63 (3.00)

12.49 (4.90)

56.1

0.63

0.34 (0.10)

Eastern hemlock

Tsuga canadensis

37.40 (1.94)

0.33 (0.042)

113.33 (20.55)

59.0

0.60

0.51 (0.11)

aMethyl gallate was used as a standard for the hydrolysable tannin assay

bPurified standards from sugar maple, red oak, and hemlock litter were used as standards for the condensed tannin assay

All tannins precipitated protein by diffusion. Sugar maple tannins precipitated 10 times more protein than red oak tannins and 7 times more protein than hemlock tannins (Table 2).

Concentration gradient study

The addition of sugar maple tannins significantly increased microbial respiration at concentrations ≥ 0.07 mg g−1 soil at both sites (Fig. 1a, b). Sugar maple tannins increased respiration significantly (P < 0.001) more than red oak or hemlock tannins at concentrations of 0.68 and 2 mg g−1 soil (Fig. 1a, b). The increase in microbial respiration with the addition of sugar maple tannins was associated with a significant decline in the rate of net N mineralization at 2.0 mg g−1 soil (Fig. 1c, d). Red oak and hemlock tannins increased microbial respiration at 0.68 and 2.0 mg g−1 soil in Dolomite soils, but did not increase microbial respiration in Schist soils (Fig. 1a, b). Red oak and hemlock tannins did not affect rates of net N mineralization in either soil type (Fig. 1c, d).
https://static-content.springer.com/image/art%3A10.1007%2Fs00442-007-0940-7/MediaObjects/442_2007_940_Fig1_HTML.gif
Fig. 1

Cumulative CO2 flux in soils from the Dolomite (a) and Schist (b) sites, and net N mineralization in Dolomite (c) and Schist (d) soils treated with tannins purified from sugar maple, red oak, or hemlock litter at concentrations of 0, 0.03, 0.07, 0.13, 0.68, or 2 mg tannin g−1 soil. Values represent the average of n = 4 samples (four replicates). Asterisks indicate significant (P < 0.05) differences from the control (0 mg tannin g−1 soil)

Tannin–protein interactions study

The addition of protein to soils significantly (< 0.0001) increased microbial respiration (Fig. 2). Soils treated with protein had significantly higher rates of net N mineralization (< 0.0001), gross NH4+ production (< 0.05), and gross NH4+ consumption (P < 0.0001) relative to control soils (Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs00442-007-0940-7/MediaObjects/442_2007_940_Fig2_HTML.gif
Fig. 2

Cumulative CO2 flux in soils from the Dolomite (a) and Schist (b) sites and net N mineralization in Dolomite (c) and Schist (d) soils treated with tannin (light gray bars) from sugar maple, red oak, or hemlock litter, protein (dark gray bars), tannin plus protein (black bars), or a water control (white bars). Significantly different (P < 0.05) treatment effects are designated by different letters (LSD). Vertical bars indicate +SE, n = 4 (four replicates)

https://static-content.springer.com/image/art%3A10.1007%2Fs00442-007-0940-7/MediaObjects/442_2007_940_Fig3_HTML.gif
Fig. 3

Gross NH4+-N production (a) and gross NH4+-N consumption (b) in soils treated with tannin (light gray bars) from sugar maple, red oak, or hemlock litter, protein (dark gray bars), tannin plus protein (black bars), or a water control (white bars). Values represent the average between soils from the Dolomite and Schist sites during the 14-day incubation. Significantly different (P < 0.05) treatment effects are designated by different letters (LSD). Vertical bars indicate ±SE, n = 8 (2 sites × 4 replicates)

The addition of sugar maple tannins alone to soils significantly (< 0.0001) increased microbial respiration at both sites (Fig. 2a, b). These soils had significantly lower rates of net N mineralization (< 0.005) and gross NH4+ production (< 0.01) relative to control soils (Figs. 2c, d, 3a). Soils treated with sugar maple tannins also tended to have lower rates of gross NH4+ consumption, although the effects were not significant (Fig. 3b). When added to soils in combination with protein, sugar maple tannins significantly (< 0.0001) increased respiration beyond that observed when protein was added to soil alone. Soils treated with sugar maple tannins plus protein had significantly lower rates of net N mineralization (< 0.0001) and gross NH4+ consumption (< 0.01) relative to soils treated with protein alone (Figs. 2c, d, 3b). When added to soils in conjunction with protein, sugar maple tannins tended to decrease the rate of gross NH4+ production, but the effect was not statistically significant (Fig. 3a).

The addition of red oak tannins alone significantly (< 0.0001) increased respiration in soils from both sites (Fig. 2a, b) but did not significantly alter rates of N transformation relative to control soils (Figs. 2c, d, 3a, b). Microbial respiration in soils treated with red oak tannins plus protein did not differ significantly from the rate of microbial respiration in soils treated with protein alone (Fig. 2a, b). When added to soils in conjunction with protein, red oak tannins did not significantly alter rates of soil-N fluxes relative to soils treated with protein alone (Figs. 2c, d, 3a, b).

The addition of hemlock tannins alone significantly (< 0.0001) increased respiration in soils from both sites (Fig. 2a, b) but did not significantly alter the rates of N transformation relative to control soils (Figs. 2c, d, 3a, b). When added to soils in conjunction with protein, hemlock tannins did not significantly alter the rate of microbial respiration relative to soils treated with protein alone (Fig. 2a, b) or net or gross net N mineralization (< 0.0001) relative to soils containing protein alone (Figs. 2c, d, 3a, b).

Discussion

In this study we purified tannins from three tree species that are dominant in northern hardwood–conifer forests of the eastern US, and applied them to two different types of soil. To create a more ecologically relevant study, we applied the tannins to the soil at concentrations characteristic of those found in the soil (Table 1). The results of this study are consistent with the emerging view that tannins from different species vary in their susceptibility to microbial degradation (Kraus et al. 2004). The addition of tannins to soil significantly increased the rate of microbial respiration by 24–250% (Figs. 1a, 3a, b), with sugar maple tannins stimulating respiration significantly more than either red oak or hemlock tannins (Figs. 1a, b, 3a, b). Similar to our study, Kraus et al. (2004) found that tannins from four out of five species in a California pygmy forest significantly increased microbial respiration when applied to soils, while tannins from the fifth species had no effect on microbial respiration. By contrast, Schimel et al. (1996, 1998) and Fierer et al. (2001) found that some tannins caused respiration to decrease by 40%.

Sugar maple tannins significantly increased rates of C mineralization (Figs. 1a, b, 2a, b) and decreased the rate of net N mineralization at their highest concentration (Figs. 1c, d, 2c, d). These results imply that the addition of sugar maple tannins increased the availability of C to the microbial community, creating a biosynthetic demand for N during soil incubation, stimulating microbial-N immobilization (Zak et al. 2000), and hence decreasing the rate of net N mineralization. However, discrepancies between measurements of net mineralization and microbial respiration suggest that PPC formation also occurred with tannin addition. In particular, the rate of microbial respiration did not increase between 0.68 and 2.0 mg tannin g−1 soil, yet the rate of net N mineralization dropped substantially (Fig. 1c, d). Sugar maple tannins contain a mixture of both labile, hydrolysable tannins and recalcitrant condensed tannins, which are more likely to be associated with PPC formation. Thus the decline in net mineralization but no change in respiration between 0.68 and 2.0 mg tannin g−1 soil treatments implies that PPC formation played an increasingly large role in soil-N cycling at higher tannin concentrations. Additional evidence of PPC formation comes from the protein-amended soils. When sugar maple tannins were added to protein-enriched soils, respiration increased significantly (Fig. 2a, b), but the increase in respiration was not equal to the sum of the respiration rates measured in the samples to which only tannin or protein were added. Furthermore, gross and net rates of N mineralization were substantially lower in soils amended with sugar maple plus protein compared to soils with only protein. Together, these observations imply that the decline in mineralization rates was due to PPC formation.

Although red oak and hemlock tannins significantly increased rates of microbial respiration (Figs. 1a, 2a, b), the increase in respiration was significantly lower than that elicited by the addition of sugar maple tannins. In addition, the red oak and hemlock tannins had little effect on net mineralization rates at any concentration (Figs. 1c, d, 2a, b). Thus red oak and hemlock tannins appear to be significantly less labile, but at the same time they do not appear to form PPCs under ambient levels of protein in the soil. When soils were amended with protein plus red oak or hemlock tannins, however, rates of gross and net N mineralization and nitrification were consistently lower than that observed in the soils containing protein alone (Figs. 2c, d, 3a, b). In combination with the observation that the rate of microbial respiration was unchanged in the protein plus red oak or hemlock tannin-amended soils, it appears that PPC formation may be controlling soil-N cycling under high concentrations of soil protein.

The differential effect of tannins on the rate of C mineralization and N cycling appears to be due to interspecific differences in tannin chemistry. There is evidence that tannin stereochemistry and chain length influence protein-binding capacity and CO2 release from soils (Kraus et al. 2003b, 2004; Nierop et al. 2006a). Our results support the more general observation that hydrolysable tannins are more labile than condensed tannins (Bhat et al. 1998; Field and Lettinga 1992; Lewis and Starkey 1968; Nierop et al. 2006b). In this study, the high rate of C mineralization in soils amended with low-levels of sugar maple tannins is likely due to the high concentration of hydrolysable tannins in this species (Table 2). By contrast, the lower rates of C mineralization in soils amended with red oak and hemlock tannins are likely due to the dominance of condensed tannins in the leaf litter of these species. Future studies must determine whether differences in molecular structure account for the difference in tannin lability among species.

There were subtle effects of soil type on the response to tannin addition that are likely mediated by differences in soil microbial activity at each site. In general, the rate of soil respiration, and hence overall microbial activity, was significantly higher at the Dolomite site than at the Schist site (Fig. 1a, b). Greater microbial activity in Dolomite soils was associated with greater sensitivity of the microbial community to tannin additions; tannins from all three tree species significantly increased respiration at the Dolomite site, whereas only sugar maple tannins affected microbial respiration at the Schist site (Fig. 1a, b). This was the only way in which soil type interacted with tannin addition. Microbial activity may have been affected by soil pH, since both soil pH and rates of microbial respiration were highest in Dolomite soils. However, the effect of soil pH on soil N-cycling did not override the direction of tannin effects given that the rate of net N mineralization with tannin addition followed a similar pattern among species at both research sites (Fig. 1c, d).

Conclusion

This study demonstrates that tannins from sugar maple, red oak, and hemlock leaf litter have markedly different effects on C and N cycling in northern hardwood–conifer forest soils. At low concentrations, sugar maple tannins acted primarily as labile substrates for microbial metabolism that decreased N availability by stimulating microbial-N immobilization. By contrast, tannins from red oak and hemlock litter stimulated microbial respiration to only a modest degree and had little effect on soil-N cycling at ambient levels of soil protein. However, the addition of higher concentrations of sugar maple tannins or, in the case of red oak and hemlock, at higher concentrations of soil protein, appeared to stimulate PPC formation which may decrease N availability to soil microorganisms or plants.

Tannins from these tree species may play a role in species effects on microbial activity and N cycling. In general, the rate of net mineralization and nitrification in red oak and hemlock stands is slower than that observed in sugar maple stands (Finzi and Berthrong 2005; Finzi et al. 1998; Lovett et al. 2004). In this study, red oak and hemlock tannins appeared to slow the rate of N mineralization via PPC formation only in soils with high protein concentrations. Thus tannins may be an important component of these species effects on nutrient cycling during periods of large protein inputs to soils, such as may occur following root turnover or after leaf senescence in the fall (Chapin et al. 2002). Contrary to the commonly observed “fast” rates of N cycling in sugar maple stands (Finzi et al. 1998; Lovett et al. 2004), the addition of sugar maple tannins appeared to slow net rates of N mineralization as a result of stimulating microbial-N immobilization. If however, the delivery of labile tannins in sugar maple litter initially stimulates microbial activity and population growth (Hamilton and Frank 2001; Weintraub et al. 2007), this may create a “priming effect” on the release of N from soil organic matter (Hamer and Marschner 2005), which would tend to increase gross and net rates of N cycling through time. Such an effect was observed in a recent laboratory study (Nierop et al. 2006b) in which the addition of tannic acid, a hydrolysable tannin, to soils increased microbial respiration and N immobilization for 7 days, after which time net N mineralization increased more rapidly than in control soils. Longer-term assays may be required to better understand the relationship between the input of sugar maple tannins to the soil and rapid rates of soil-N cycling in sugar maple stands.

Acknowledgements

We would like to thank the Great Mountain Forest Corporation and Jody Bronson, for their support of the research conducted at the Great Mountain Forest, and the State of Connecticut, Department of Environmental Protection, Natural Area Preserves Program for granting us access to the forests on the Canaan Mountain. Thanks also to Eddie Brzostek, Vikki Rodgers, and Anne Gallet-Budynek for their help with field and lab work, as well as for their support and review of this paper. This research complies with the current laws of United States and was supported by a grant from the United States Department of Agriculture (2000-00782).

Copyright information

© Springer-Verlag 2008